Science Research Methodology

# Introduction to scientific conduct and misconduct This section introduces the fundamental principles of scientific conduct and the reasons behind scientific misconduct, emphasizing the importance of trust in science and the motivations driving fraudulent practices. ### 1.1 The nature of trust in science Trust is a cornerstone of the scientific enterprise. It underpins the reliability and acceptance of scientific findings, both within the scientific community and among the public. The integrity of scientific research is directly linked to the level of trust it can garner [3](#page=3). ### 1.2 Scientific misconduct Scientific misconduct refers to the fabrication, falsification, or plagiarism in proposing, performing, or reviewing research, or in reporting research results. Understanding why misconduct occurs involves examining the psychological factors that can lead individuals to engage in fraudulent practices [4](#page=4) [5](#page=5). #### 1.2.1 Motivations for scientific fraud The motivations behind scientific fraud are complex and can be attributed to various pressures and psychological states. These include intense personal or professional needs, such as the pressure to "publish or perish," which drives a relentless demand for research output. Other motivations stem from a rationalization of the misconduct, where individuals may justify their actions by downplaying their severity or believing they can conceal them. The perceived value of a hypothesis, even if the supporting data is imperfect, can also contribute to this rationalization [4](#page=4) [7](#page=7). #### 1.2.2 The fraud triangle The "Fraud Triangle" is a model used to explain the common factors present when fraud occurs. It comprises three elements [6](#page=6) [7](#page=7): * **Motive / Pressure:** This refers to the underlying reasons or pressures that drive an individual to commit fraud. In an academic context, this often relates to the "publish or perish" culture, where career advancement is heavily tied to publication output [7](#page=7). * **Opportunity:** This element describes the perceived chance to commit fraud and go undetected. In scientific research, data can be controlled and manipulated by the scientist, presenting an opportunity for falsification or fabrication. The ability to conceal such actions is also a key aspect of opportunity [7](#page=7). * **Rationalization:** This is the psychological justification or excuse that an individual uses to rationalize their fraudulent behavior. Common rationalizations include believing the data is "a bit off, but the hypothesis is great" or that "I only cheated a little bit". The belief that "as always in life," these actions are somehow permissible or can be overlooked also falls under this category [7](#page=7). > **Tip:** Understanding the fraud triangle is crucial for identifying potential risks of misconduct and for implementing preventative measures within research environments. #### 1.2.3 Perceptions of cheating Surveys and observations of the scientific community suggest that a certain percentage of scientists may engage in cheating. Perceptions of incentives, research misconduct, and scientific integrity vary among researchers, highlighting the importance of ongoing discussions and education on these topics [4](#page=4). > **Example:** A study by Roy & Edwards explored NSF Fellows' perceptions regarding incentives for research, the prevalence of misconduct, and the state of scientific integrity in STEM academia [4](#page=4). --- # Classic and neo-classic examples of scientific misconduct This section explores historical and contemporary cases of scientific misconduct, illustrating various fraudulent practices and their significant repercussions across different scientific disciplines [33](#page=33) [8](#page=8) [9](#page=9). ### 2.1 Classic examples of scientific misconduct #### 2.1.1 Gregor Mendel (1822 – 1884) The accusation against Gregor Mendel, a pioneer in genetics, centers on the statistical analysis of his experiments. While Mendel is credited with fundamental insights into inheritance, some critics suggest his reported results might have been too perfectly aligned with his theoretical expectations, potentially indicating insufficient or incorrect statistical application. However, other interpretations suggest that these discrepancies could be due to honest mistakes or biases rather than deliberate fraud [10](#page=10) [11](#page=11) [12](#page=12) [13](#page=13). #### 2.1.2 Piltdown man . The Piltdown man fossil discovery in 1912 was a significant paleontological finding intended to be the "missing link" between humans and apes. However, the fossils were later revealed to be a sophisticated hoax. The jawbone belonged to an orangutan, the teeth were filed down, the bones were artificially colored, and radiocarbon dating of the skull bone indicated an age far younger than claimed. The accompanying club found at the site also showed evidence of being shaped by a steel knife. This case is a clear example of outright fraud involving forged data, results, and findings. Key figures involved in this forgery include Robert Kenwood, Charles Dawson, and Arthur Smith Woodward [14](#page=14) [15](#page=15) [16](#page=16). #### 2.1.3 Robert A. Milikan (1868 – 1953) Robert A. Milikan, a Nobel Prize winner in Physics for his work on the charge of the electron and the photoelectric effect, faced accusations of misconduct related to his oil drop experiment. A paper published in 1913 reported an error of only 0.2%, significantly lower than previous calculations of 3%. Milikan claimed that his results represented all drops experimented upon during 60 consecutive days, without selection. However, a review of his notebooks revealed that he used only 58 out of 75-100 trial results, suggesting a biased representation of his data. His private notes indicated that he discarded results that deviated significantly from his expected values [17](#page=17) [18](#page=18) [19](#page=19). > **Tip:** Biased selection of data, even if not outright fabrication, can lead to misleading scientific conclusions and is considered a form of misconduct. #### 2.1.4 Tuskegee Syphilis Study (1932 – 1972) The Tuskegee Syphilis Study was a 40-year longitudinal medical study conducted by the U.S. Department of Health on the long-term effects of untreated syphilis in African American men in Alabama. Over 400 men and their families were involved in this study. The study is a stark example of severe human ethics violations because participants were denied appropriate medical treatment for 25 years, even after penicillin became available in 1947 [20](#page=20) [21](#page=21). > **Example:** This case highlights that scientific misconduct is not limited to data manipulation but also encompasses egregious breaches of ethical treatment of human subjects. #### 2.1.5 William Summerlin (1938 – recent) William Summerlin's work at Memorial Sloan Kettering in 1974 involved experiments on tissue rejection using transplantation of human corneas into rabbits and skin transplants. The misconduct associated with his research was fabrication [22](#page=22) [24](#page=24). #### 2.1.6 Philip Felig / Vijay Soman and Helena Rodbard (Breach of peer-review) This case illustrates a breach of the peer-review process. Helena Rodbard submitted a manuscript to the New England Journal of Medicine, which was subsequently rejected. Felig and Soman, after their own manuscript was rejected, then submitted their work, which was also rejected after being sent out for review. The implication is that they may have stolen or misused review information or engaged in other unethical practices related to the review process [25](#page=25) [26](#page=26) [27](#page=27). #### 2.1.7 Andrew Wakefield (1957 - recent) Andrew Wakefield's 1998 study published in The Lancet falsely linked the measles, mumps, and rubella (MMR) vaccine to autism and bowel disease. The study's findings, which associated behavioral symptoms with vaccinations, were heavily flawed. Key issues included [28](#page=28) [29](#page=29) [32](#page=32): * Wakefield was paid by anti-vaccine activists to conduct the study [32](#page=32). * He performed medical examinations without proper qualifications or ethical approval, including colonoscopies, biopsies, and lumbar punctures [32](#page=32). * He purchased blood samples from children at his son's birthday party for a nominal fee [32](#page=32). * Data was falsified and twisted: * Three of nine children reported with regressive autism were not diagnosed with autism [32](#page=32). * Five of the 12 children reported as "normal" had documented pre-existing developmental concerns [32](#page=32). * Children whose behavioral symptoms were reported to appear days after MMR vaccination actually showed signs months later [32](#page=32). * Normal colonic histopathology results were altered to "non-specific colitis" after a "research review" in nine cases [32](#page=32). This misconduct created a dangerous misconception in society, leading to decreased vaccination rates and increased outbreaks of preventable diseases [31](#page=31). ### 2.2 Neo-classic examples of scientific misconduct #### 2.2.1 Diederick Stapel (1966 – recent) Diederick Stapel's misconduct involved fabricating and adapting research data across numerous studies in social psychology, leading to the retraction of 58 articles. His research topics included potentially sensational or "sexy" science, such as the link between meat-eating and selfishness, and the effect of power on infidelity. Stapel admitted to fabricating data "several times, not for a short period, but over a longer period of time," driven by the pressure to "score" and publish. He expressed shame and regret for his actions and the negative impact on his field [33](#page=33) [34](#page=34). > **Tip:** The pressure to publish and achieve results can be a significant factor contributing to scientific misconduct. #### 2.2.2 Woo Suk Hwang (1953 - recent) Woo Suk Hwang's work in stem cell research was marred by significant misconduct and ethical violations. He claimed to have cloned human embryonic stem cells with high success rates, but the actual efficiencies were much lower. Further allegations included [35](#page=35) [36](#page=36): * Obtaining egg donations from graduate students and lab technicians [35](#page=35) [36](#page=36). * Forging data [35](#page=35) [36](#page=36). * Manipulating photographs [35](#page=35) [36](#page=36). * Paying lab members and coworkers to remain silent about his fraudulent activities [35](#page=35) [36](#page=36). #### 2.2.3 Joachim Boldt (1954 – recent) Joachim Boldt is described as a "record breaker" in scientific misconduct, with 194 research papers retracted. He was a German anesthesiologist whose misconduct primarily involved the failure to obtain ethical approval for human studies and the fabrication of data [48](#page=48). ### 2.3 Other malpractices and consequences #### 2.3.1 Peer-review mafia and fictitious reviewers A significant issue in scientific publishing involves "peer-review mafia" and the use of fictitious reviewers. This practice involves authors creating fake peer reviews to get their papers accepted. In one extreme case, Springer retracted 107 papers from the journal *Tumor Biology* due to discovered fake peer reviews, where either fake experts were invented or real researchers' emails were spoofed to provide glowing reviews. Some authors may have used third-party editing services that supplied these fraudulent reviews [41](#page=41) [43](#page=43). #### 2.3.2 Merck and Elsevier's fake journals Merck allegedly paid Elsevier, a prominent publisher, to create fake scientific journals to promote Merck's drugs. This issue surfaced in 2009 during a civil lawsuit involving a heart attack patient who had taken Merck's drug Vioxx. The articles published in these fabricated journals, such as *Bone & Joint Medicine*, supported Merck's products [44](#page=44). #### 2.3.3 Plagiarism in academic theses A report by the Education Policy Research and Application Center (BEPAM) of Istanbul’s Boğaziçi University indicated that a substantial portion of academic theses in Turkey exhibit high plagiarism rates. Examining 600 theses (470 master's, 130 doctoral) written between 2007 and 2016, the study found that 34 percent had high plagiarism rates. Plagiarism was more prevalent in private universities (46%) compared to public universities (31%). Institutions with English-language programs, such as Boğaziçi University, ODTÜ, and Bilkent University, appeared to have better plagiarism records [58](#page=58). #### 2.3.4 Paper retractions on the rise The number of scientific paper retractions is increasing. Data indicates a rise in retractions per year and the time taken for retractions to occur. Retractions are also broken down by country, impact factor, and journal, highlighting global trends and specific publication venues where retractions are more common. The primary reasons for retractions include misconduct such as data fabrication, falsification, plagiarism, and authorship issues [51](#page=51) [52](#page=52) [53](#page=53) [55](#page=55) [56](#page=56) [57](#page=57). > **Example:** The journal *Acta Crystallogr E* is cited as having a significant number of retractions [54](#page=54). #### 2.3.5 Exposure and consequences Scientific misconduct can lead to significant exposure and severe consequences for individuals involved. Cases like Dr. Karl-Theodor Maria Nikolaus Johann Jacob Philipp Franz Joseph Sylvester Buhl - Freiherr von und zu Guttenberg and Anette Schavan illustrate prominent figures facing repercussions for academic dishonesty. These consequences can range from public scrutiny to the revocation of academic degrees and damage to careers. Resources like Retraction Watch track and report on these issues, bringing transparency to retractions and misconduct [45](#page=45) [46](#page=46) [47](#page=47) [49](#page=49) [50](#page=50). --- # Definitions and types of scientific misconduct This section formally defines research misconduct and elaborates on its primary forms: fabrication, falsification, and plagiarism. ### 3.1 Defining research misconduct Research misconduct is formally defined as the fabrication, falsification, or plagiarism of research results during the stages of proposing, performing, or reviewing research [59](#page=59). ### 3.2 Types of scientific misconduct The definition of research misconduct encompasses three core activities: #### 3.2.1 Fabrication Fabrication involves the act of making up data or results and then recording or reporting them as if they were genuine [59](#page=59). > **Tip:** Fabrication represents the creation of entirely false research findings, which are then presented as authentic. #### 3.2.2 Falsification Falsification is defined as the manipulation of research materials, equipment, or processes. It also includes altering or omitting data and results in a way that causes the research record to be an inaccurate representation of the actual work [59](#page=59). > **Example:** A researcher might falsify results by selectively removing data points that do not support their hypothesis, or by altering experimental parameters to achieve a desired outcome. #### 3.2.3 Plagiarism Plagiarism is the appropriation of another individual's ideas, processes, results, or words without providing proper attribution or credit to the original source [59](#page=59). > **Tip:** Always ensure that any borrowed ideas, data, or text are properly cited to avoid plagiarism. This includes both direct quotes and paraphrased information. --- ## Common mistakes to avoid - Review all topics thoroughly before exams - Pay attention to formulas and key definitions - Practice with examples provided in each section - Don't memorize without understanding the underlying concepts

| Term | Definition | |------|------------| | Scientific Conduct | Refers to the principles, practices, and ethical standards that guide the honest and responsible conduct of scientific research. It encompasses integrity, objectivity, and transparency in all research activities. | | Scientific Misconduct | Involves the fabrication, falsification, or plagiarism in proposing, performing, or reviewing research results, or in reporting research. It is a serious violation of ethical principles in science. | | Fraud Triangle | A conceptual model that explains why individuals commit fraud, comprising three elements: motive or pressure, opportunity, and rationalization. These factors collectively contribute to an environment where fraudulent behavior can occur. | | Motive / Pressure | One of the components of the fraud triangle, representing the perceived intense needs or drives that push an individual towards committing fraud, such as the pressure to publish or achieve specific career goals. | | Opportunity | The second component of the fraud triangle, referring to the circumstances or conditions that allow an individual to commit fraud, such as the ease with which data can be manipulated or concealed due to lax oversight. | | Rationalization | The third component of the fraud triangle, where an individual justifies their fraudulent behavior to themselves, often by minimizing its severity or believing it is acceptable under certain circumstances. | | Fabrication | The act of making up results and then recording or reporting them as if they were obtained through legitimate research methods. This directly misrepresents the findings of a study. | | Falsification | The manipulation of research materials, equipment, or processes, or the changing or omitting of results. This is done in a way that causes the research to be inaccurately represented in the scientific record. | | Plagiarism | The appropriation of another person’s ideas, processes, results, or words without giving them proper credit. This is a form of intellectual theft and a violation of academic integrity. | | Peer Review | The evaluation of scientific work by other experts in the same field. It is a critical process for ensuring the quality, validity, and significance of research before it is published. | | Retraction | The formal withdrawal of a published scientific paper or article. Retractions are typically issued when significant flaws, errors, or misconduct are discovered in the published work. | | Tuskegee Syphilis Study | A notorious historical study conducted by the U.S. Public Health Service where the effects of untreated syphilis were observed in African American men without providing adequate medical treatment, even after penicillin became available. | | Piltdown Man | A famous archaeological hoax involving the discovery of supposed early human fossils in England. The fossils were later proven to be a fraudulent combination of human skull fragments and an orangutan jawbone. | | Oil Drop Experiment | A landmark experiment conducted by Robert Millikan to measure the elementary charge of the electron. This experiment has been subject to scrutiny regarding potential data manipulation. | | Peer-review mafia | A colloquial term referring to a group or network of individuals who manipulate the peer-review process to their advantage, often by submitting fake reviews or colluding to promote or reject specific papers. | | Fake peer reviews | Reviews submitted for scientific manuscripts that are not conducted by genuine, qualified experts. This can involve creating fictitious reviewers or impersonating real researchers to manipulate the publication process. | | Vaccination coverage rate | The percentage of a population that has received a specific vaccine. This metric is crucial for assessing the effectiveness of public health interventions and understanding disease spread. |

# Understanding scientific misconduct Scientific misconduct is a serious ethical breach within research, driven by psychological factors that create opportunities for dishonesty and rationalizations for such behavior. ## 1. Understanding scientific misconduct Scientific misconduct encompasses acts of fabrication, falsification, or plagiarism in the research process. It represents a departure from ethical research practices and can have profound implications for the integrity of scientific knowledge. Understanding why it happens involves examining the psychological underpinnings that lead individuals to engage in dishonest practices [59](#page=59) [5](#page=5). ### 1.1 The definition of scientific misconduct The Department of Health and Human Services defines research misconduct as the fabrication, falsification, or plagiarism in proposing, performing, or reviewing research results [59](#page=59). * **Fabrication** involves making up results and then recording or reporting them as if they were genuine [59](#page=59). * **Falsification** refers to the manipulation of research materials, equipment, or processes, or the alteration or omission of results, thereby misrepresenting the research in the record [59](#page=59). * **Plagiarism** is the appropriation of another person's ideas, processes, results, or words without providing proper attribution [59](#page=59). ### 1.2 The psychology of scientific misconduct The occurrence of scientific misconduct is not a new phenomenon, with historical instances influencing its perception over time. The psychological drivers behind such behavior are complex and can be understood through several models, most notably the Fraud Triangle [5](#page=5) [6](#page=6). ### 1.3 The fraud triangle model The Fraud Triangle, a widely recognized model in criminology, provides a framework for understanding the elements that contribute to fraudulent behavior, including scientific misconduct. This model posits that three conditions must be present for fraud to occur: motive or pressure, opportunity, and rationalization [6](#page=6). #### 1.3.1 Motive / Pressure Motive or pressure refers to the internal or external forces that compel an individual to commit fraud. In the context of scientific misconduct, these pressures can be significant [7](#page=7). * **Perceived intense needs:** Scientists may feel intense pressure due to a "publish or perish" environment, where career advancement and funding are heavily dependent on publication output. This can create a desperate need to produce results, even if they are not entirely genuine [7](#page=7). * **General life pressures:** As in any profession, scientists may also face personal financial difficulties or other life stressors that create a motive for dishonest behavior [7](#page=7). #### 1.3.2 Opportunity Opportunity arises when there is a perceived chance to commit fraud without being detected. The nature of scientific work can present such opportunities [7](#page=7). * **Control over data:** In many research settings, the scientist has significant control over the data. This control means that data can be easily changed, manipulated, or fabricated, creating an avenue for deception [7](#page=7). * **Concealment:** The complexity and specialized nature of scientific research can sometimes make it difficult for peers or supervisors to fully scrutinize and detect subtle forms of misconduct, providing an opportunity for concealment [7](#page=7). #### 1.3.3 Rationalization Rationalization is the process by which individuals justify their dishonest actions to themselves, making them seem acceptable or less severe. This psychological mechanism allows individuals to maintain a positive self-image despite engaging in unethical behavior [7](#page=7). * **Minimizing the harm:** Individuals may rationalize their actions by believing that "the data may be a bit off, but the hypothesis is great". This involves downplaying the significance of the fabricated or falsified data by focusing on the perceived value of the underlying scientific idea [7](#page=7). * **Self-deception:** Another common rationalization is "I only cheated a little bit". This involves minimizing the extent of the misconduct, making it seem like a minor transgression rather than a serious breach of integrity [7](#page=7). * **Belief in ability to conceal:** A crucial element of rationalization is the belief that "I can do/conceal it!". This conviction that one can get away with the misconduct without facing consequences further enables the dishonest act [7](#page=7). > **Tip:** The Fraud Triangle highlights that scientific misconduct is often not the act of inherently bad individuals but rather a confluence of pressure, opportunity, and the psychological ability to rationalize the behavior. Understanding these components is key to developing strategies for prevention and detection. > **Example:** A PhD student struggling to meet publication deadlines for their thesis (motive/pressure) might notice that a few data points in a crucial experiment are slightly aberrant but would support their hypothesis if slightly adjusted (opportunity). They might then rationalize this by thinking, "The overall trend is correct, and this small adjustment doesn't really change the conclusion, plus no one will notice" (rationalization). This chain of thought could lead to falsification of results. --- # Classic and contemporary examples of scientific misconduct This topic examines significant historical and recent cases of alleged and confirmed scientific misconduct, detailing various forms of ethical breaches and data manipulation across different disciplines [20](#page=20) [33](#page=33) [8](#page=8). ### 2.1 Classic examples of scientific misconduct Classic instances of scientific misconduct often involve foundational figures and pivotal discoveries, where accusations of data manipulation, statistical inaccuracies, or fabrication have surfaced upon closer scrutiny. #### 2.1.1 Gregor Mendel Gregor Mendel, a pioneering figure in genetics, faced accusations related to insufficient or incorrect statistics in his work on inheritance. While his experiments were carefully planned, some interpretations suggest that his reported results, particularly concerning the F2 and F3 generations (e.g., F2 = 1:3 leading to F3 = 1:2, instead of the expected F2 = 1:2:1) might have been presented in a way that too closely aligned with his hypotheses. The debate continues whether this was due to honest mistakes, bias, or a deliberate attempt to demonstrate his conclusions more clearly [10](#page=10) [11](#page=11) [12](#page=12) [13](#page=13) [8](#page=8). #### 2.1.2 Piltdown man The Piltdown man fossils, discovered in 1912, were presented as a crucial "missing link" between humans and ape-like ancestors. However, the evidence later revealed extensive fraud. The jawbone belonged to an orangutan, its teeth had been filed to appear human, and the bones were artificially aged. Carbon dating indicated the skull bone was between 520 and 720 years old, and a club found at the site showed signs of being shaped by a steel knife. This case is a clear example of forged data, results, and findings, attributed to individuals like Robert Kenwood, Charles Dawson, and Arthur Smith Woodward [14](#page=14) [15](#page=15) [16](#page=16). #### 2.1.3 Robert A. Milikan Robert A. Milikan, a Nobel laureate in Physics for his work on the oil drop experiment, faced scrutiny regarding the accuracy of his published results. His 1913 paper reported an error of only 0.2%, a significant improvement from previous calculations. However, a review of his notebooks revealed that he used only 58 out of 75-100 trial results. The published statement claimed that only one drop out of 58 deviated by as much as 0.5% and represented all experiments conducted over 60 days, including instances where the apparatus was dismantled and reassembled. This suggested a biased representation of the data, with Milikan selectively choosing results that supported his conclusions [17](#page=17) [18](#page=18) [19](#page=19). #### 2.1.4 Tuskegee Syphilis Study The Tuskegee Syphilis Study, conducted from 1932 to 1972, was a longitudinal medical study by the U.S. Department of Health investigating the long-term effects of untreated syphilis in African American men in Alabama. Over 400 men and their families were involved in this study, during which they were deliberately denied appropriate medical treatment, even after penicillin became available in 1947. This case represents a severe violation of human ethics [20](#page=20) [21](#page=21). #### 2.1.5 William Summerlin William Summerlin's work at Memorial Sloan Kettering in 1974 involved transplantation experiments, including human corneas into rabbits and skin transplants, aimed at studying tissue rejection. The misconduct in this case was identified as fabrication [22](#page=22) [24](#page=24). #### 2.1.6 Philip Felig and Vijay Soman (Breach of Peer Review) The case involving Philip Felig and Vijay Soman highlights a breach of the peer-review process. After a manuscript was submitted and rejected by the New England Journal of Medicine, it was sent out for review by Helena Rodbard, who also rejected it. The underlying misconduct was identified as stealing [25](#page=25) [26](#page=26) [27](#page=27). #### 2.1.7 Andrew Wakefield Andrew Wakefield's research, published in 1998, linked the measles, mumps, and rubella (MMR) vaccine to autism and bowel disease. This led to the creation of a dangerous misconception in society, contributing to a decline in vaccination coverage. Further investigation revealed several critical issues: Wakefield was paid by activists to conduct the study, performed medical examinations without proper qualification or ethical approval (including colonoscopies, biopsies, and lumbar punctures), and purchased blood samples at a cost of five dollars each. The data was also falsified or twisted; for instance, three children reported with regressive autism were not diagnosed with autism, five children had documented pre-existing developmental concerns, and behavioral symptoms were reported to appear within days of vaccination when they actually manifested months later. Additionally, normal colonic histopathology results were altered to "non-specific colitis" after a "research review" [29](#page=29) [30](#page=30) [31](#page=31) [32](#page=32). ### 2.2 Contemporary examples of scientific misconduct Contemporary cases often reflect the pressures of the modern academic environment, characterized by "publish or perish" mentalities, and involve a wider range of sophisticated methods of deception. #### 2.2.1 Diederick Stapel Diederick Stapel was involved in what has been termed "sexy science," producing articles with sensational titles such as "Meat eaters are more selfish than vegetarians" and "Power increases infidelity among men and women". Stapel confessed to fabricating and adapting research data, not just once but multiple times over an extended period. He acknowledged the shame and regret he felt, recognizing the negative impact on his field, social psychology, and the academic community. He attributed his actions to the pressure to "score" and publish, and a desire for too much, too fast, particularly in a system with few checks and balances [33](#page=33) [34](#page=34). #### 2.2.2 Woo Suk Hwang Woo Suk Hwang faced accusations of misconduct, ethical violations, and fraud related to his claims of cloning human embryonic stem cells at a high success rate, when actual efficiencies were much lower. Further issues included egg donations from graduate students and lab technicians, forged data, manipulated photographs, and attempts to pay lab members to remain silent [35](#page=35) [36](#page=36). #### 2.2.3 Linda Buck Linda Buck, a recipient of the Nobel Prize in Physiology or Medicine in 2004, is mentioned in the context of scientific misconduct and honesty, though the specific details of any misconduct are not elaborated upon in the provided text [37](#page=37). #### 2.2.4 Peer-review mafia and fictitious reviewers This category encompasses "other malpractices" within scientific publishing. One significant issue is the use of "peer-review mafia" and fictitious reviewers, where fake peer reviews are submitted for articles. This can involve creating fictional experts or suggesting real researchers with fake email addresses that lead back to individuals who provide glowing reviews. In one documented instance, Springer retracted 107 papers from the journal *Tumor Biology* after discovering fake peer reviews [41](#page=41) [43](#page=43). #### 2.2.5 Merck and Elsevier's collaboration Merck allegedly paid Elsevier, a reputable publisher, to create fake scientific journals to promote its drugs. This issue came to light in 2009 during a civil suit related to the drug Vioxx, with articles in *Bone & Joint Medicine* appearing to support Merck's products [44](#page=44). #### 2.2.6 Politicians and academics facing consequences Several individuals, including politicians and academics, have faced exposure and consequences for scientific misconduct. * **Dr. Karl-Theodor Maria Nikolaus Johann Jacob Philipp Franz Joseph Sylvester Buhl - Freiherr von und zu Guttenberg:** Mentioned in the context of scientific misconduct and exposure/consequences [45](#page=45) [46](#page=46). * **Anette Schavan:** Also mentioned in relation to scientific misconduct and exposure/consequences [47](#page=47). * **Joachim Boldt:** Described as a "record breaker" for faking 194 research papers, primarily due to failure to obtain ethical approval for human studies and fabrication of data [48](#page=48). ### 2.3 Paper retractions and their impact The phenomenon of paper retractions is on the rise, serving as a consequence of scientific misconduct [51](#page=51). #### 2.3.1 Trends in retractions Analysis of retractions shows trends by year, retraction rate, and time to retraction. Retraction rates can also be broken down by country and by the impact factor or journal in which the research was published. Specific journals, such as *Acta Crystallogr E*, are noted in this context [51](#page=51) [52](#page=52) [53](#page=53) [54](#page=54). #### 2.3.2 Causes of retraction A breakdown of retractions by misconduct type reveals various reasons for papers being withdrawn. These include issues such as fraud, plagiarism, and data fabrication [55](#page=55) [56](#page=56) [57](#page=57). ### 2.4 Early indicators of misconduct Evidence suggests that misconduct can begin early in an academic career. A report by the Education Policy Research and Application Center (BEPAM) of Istanbul’s Boğaziçi University indicated that a significant percentage of academic theses in Turkey have high plagiarism rates. Their examination of 600 theses revealed that plagiarism was prevalent in both public and private universities, as well as in master's and doctoral theses. Institutions with English-language programs appear to have a relatively better standing regarding plagiarism [58](#page=58). --- # The consequences and exposure of scientific misconduct This section examines how scientific misconduct is identified and exposed, detailing the increase in paper retractions and the repercussions faced by individuals involved in such practices [38](#page=38) [39](#page=39) [40](#page=40) [41](#page=41) [42](#page=42) [43](#page=43) [44](#page=44) [45](#page=45) [46](#page=46) [47](#page=47) [48](#page=48) [49](#page=49) [50](#page=50) [51](#page=51) [52](#page=52) [53](#page=53) [54](#page=54) [55](#page=55) [56](#page=56) [57](#page=57). ### 3.1 Identification and exposure of misconduct Scientific misconduct can be identified through various means, including investigations into suspicious peer reviews and allegations of data fabrication or falsification. The rise of online platforms and databases dedicated to tracking retractions has also significantly contributed to the exposure of misconduct [43](#page=43) [48](#page=48) [49](#page=49) [50](#page=50). #### 3.1.1 Fictitious reviewers and fake peer reviews A concerning form of misconduct involves the creation of "peer-review mafias" where authors or third-party services submit papers with fake peer reviews. This often entails fabricating expert reviewers or providing fake email addresses that lead back to individuals who will invariably give the paper a positive assessment. In one instance, Springer retracted 107 papers from the journal *Tumor Biology* due to evidence of fake peer reviews, where real researchers' names were used with faked emails [41](#page=41) [43](#page=43). #### 3.1.2 Publisher complicity and conflicts of interest Misconduct can also extend to publishers, as seen in a case where Merck allegedly paid Elsevier to create fake scientific journals to promote its drugs, such as Vioxx. This issue came to light in 2009 during a civil suit related to Vioxx [44](#page=44). ### 3.2 Consequences of scientific misconduct The consequences of scientific misconduct can be severe, ranging from career damage to criminal charges, and most notably, the retraction of published work. #### 3.2.1 Paper retractions on the rise There has been a significant increase in the number of journal articles being retracted, indicating a growing awareness and action against scientific misconduct. This trend is tracked by various metrics, including the number of retractions per year, the overall retraction rate, and the time it takes for a paper to be retracted after publication [51](#page=51). > **Tip:** Understanding the trends in paper retractions is crucial for grasping the scale of scientific misconduct and the efforts being made to maintain research integrity. #### 3.2.2 Retraction rates and contributing factors Retraction rates can vary significantly by country, and analyses break down these rates by various factors, including the impact factor of the journal and the specific type of misconduct. For example, the journal *Acta Crystallogr E* has been highlighted in this context. Studies by Bhatt and Grieneisen & Zhang provide comprehensive surveys and analyses of retracted articles, categorizing them by the nature of the misconduct [51](#page=51) [52](#page=52) [53](#page=53) [54](#page=54) [55](#page=55) [56](#page=56) [57](#page=57). #### 3.2.3 High-profile cases of misconduct Several prominent individuals have faced repercussions for scientific misconduct: * **Dr. Karl-Theodor Maria Nikolaus Johann Jacob Philipp Franz Joseph Sylvester Buhl - Freiherr von und zu Guttenberg:** A notable case of scientific misconduct [45](#page=45). * **Anette Schavan:** Another individual implicated in scientific misconduct [47](#page=47). * **Joachim Boldt:** Dubbed "the record breaker," Boldt was identified as having faked 194 research papers. His misconduct primarily involved the failure to obtain ethical approval for human studies and the fabrication of data [48](#page=48). > **Example:** Joachim Boldt's case illustrates the scale of data fabrication, with nearly 200 papers affected, underscoring the severe impact on the scientific record and patient safety. --- # Research integrity and early detection This section addresses the crucial need to uphold research integrity and highlights how scientific misconduct can emerge early in academic careers, as indicated by high rates of plagiarism in theses. ### 4.1 The importance of repairing research integrity Upholding research integrity is a fundamental aspect of the scientific endeavor, essential for maintaining public trust and ensuring the reliability of knowledge. Efforts to repair and strengthen research integrity are vital for the long-term health of the scientific community. The focus on this aspect is underscored by the reality that scientific misconduct is not an issue confined to senior researchers; it can, and often does, begin early in an academic career [60](#page=60). ### 4.2 Early onset of scientific misconduct Evidence suggests that scientific misconduct can manifest even at the initial stages of an academic journey, particularly during thesis writing. #### 4.2.1 Plagiarism rates in academic theses A report by the Education Policy Research and Application Center (BEPAM) of Istanbul’s Boğaziçi University revealed concerning plagiarism rates in academic theses. * **Scope of the study:** BEPAM analyzed a total of 600 theses, comprising 470 master’s theses and 130 doctoral theses, written between 2007 and 2016 [58](#page=58). * **University types:** Of the examined theses, 477 were from public universities and 123 were from private universities [58](#page=58). * **Language of theses:** The study included 89 theses written in English and 511 written in Turkish [58](#page=58). * **Plagiarism prevalence:** * Across all analyzed theses, 34 percent exhibited high plagiarism rates [58](#page=58). * The number of plagiarized research studies in public universities was 150 (31%), while in private universities it was 57 (46%) [58](#page=58). * Master's theses showed a plagiarism rate of 173 (36%), compared to 34 (26%) in doctoral theses [58](#page=58). * Theses written in English had a plagiarism rate of 25 (28%), while those in Turkish had a rate of 182 (35%) [58](#page=58). #### 4.2.2 Institutional differences in plagiarism Institutions that offer education in English, such as Boğaziçi University, Middle East Technical University (ODTÜ), and Bilkent University, appear to be in a comparatively better position regarding plagiarism and similarity issues in their theses. This suggests that the academic environment and potentially the rigor of English-language programs may play a role in mitigating such misconduct [58](#page=58). > **Tip:** The high rates of plagiarism in theses underscore the need for robust training in research ethics and academic integrity from the very beginning of graduate studies. Early intervention and education are key to fostering a culture of honesty in research. --- ## Common mistakes to avoid - Review all topics thoroughly before exams - Pay attention to formulas and key definitions - Practice with examples provided in each section - Don't memorize without understanding the underlying concepts

| Term | Definition | |------|------------| | Scientific misconduct | The definition of research misconduct encompasses fabrication, falsification, or plagiarism in the proposing, performing, or reviewing of research results. | | Fabrication | The act of making up results and then recording or reporting them as if they were genuine findings. | | Falsification | The manipulation of research materials, equipment, or processes, or the alteration or omission of results, such that the research is not accurately represented in the record. | | Plagiarism | The appropriation of another person's ideas, processes, results, or words without giving proper credit or attribution to the original source. | | Fraud Triangle | A model that explains the reasoning behind committing fraud, consisting of three elements: motive or pressure, opportunity, and rationalization. | | Motive / Pressure | This component of the Fraud Triangle refers to the internal or external pressures that drive an individual to commit fraud, such as perceived intense needs or the pressure to publish. | | Opportunity | In the context of the Fraud Triangle, this refers to the circumstances or conditions that allow an individual to commit fraud, often because the data or processes are controlled and can be easily manipulated or concealed. | | Rationalization | This element of the Fraud Triangle involves the mental process by which an individual justifies their fraudulent actions, often by downplaying their severity or personalizing the behavior. | | Peer-review | The evaluation of scientific work by others who are competent in the same field, typically to assess its validity, quality, and originality before publication. | | Retraction | The formal withdrawal of a published article from a journal, usually due to serious errors, misconduct, or ethical concerns identified after publication. | | Tuskegee Syphilis Study | A historical study conducted by the U.S. Public Health Service where untreated syphilis was deliberately left untreated in African American men for decades, even after penicillin became available as a treatment. | | Piltdown Man | A famous paleoanthropological hoax involving the discovery of fossil fragments that were presented as belonging to a previously unknown early human ancestor, later revealed to be a forgery. | | Oil drop experiment | A significant experiment conducted by Robert Millikan to measure the elementary electric charge. Concerns were raised about potential data manipulation in his published results. | | Fabrication of data | The act of inventing or creating false data for research purposes, leading to misleading or incorrect scientific conclusions. | | Breaching peer-review | This refers to acts that undermine the integrity of the peer-review process, such as submitting fake reviews, manipulating reviewer identities, or stealing ideas from submitted manuscripts. | | Fake peer reviews | A malicious practice where fabricated reviews are submitted for a manuscript, often created by the author or a third party, to deceive editors and hasten publication. | | Research integrity | Adherence to ethical principles and professional standards in conducting, reporting, and reviewing scientific research, ensuring honesty, accuracy, and responsibility. | | Acts of cheating | Behavior that involves dishonesty or deception, particularly in academic or scientific contexts, to gain an unfair advantage or achieve desired outcomes. | | Voynich Manuscript | An illustrated codex of unknown authorship, written in an unknown writing system. Its authenticity and meaning have been subjects of intense debate and speculation for centuries. | | Misconduct / ethical violations / fraud | A broad category encompassing dishonest or unethical actions in research, including fabricating data, misrepresenting findings, or engaging in fraudulent activities. | | Honesty in science | The principle of conducting scientific research with integrity, truthfulness, and transparency, ensuring that findings are reported accurately and without deception. | | Exposure of misconduct | The process by which instances of scientific misconduct are brought to light, often through investigations, whistleblowers, or vigilant scrutiny of published work. | | Paper retractions | The official withdrawal of scientific papers from publication due to identified flaws, errors, or misconduct, signaling a failure in the research or publication process. | | Plagiarism rates | The statistical measurement of the occurrence of plagiarism within a given population or body of work, often used to assess academic integrity. | | Neo-classic examples | Refers to more recent or contemporary cases that exemplify well-established patterns of scientific misconduct, similar to historical landmark cases. | | Forged data | Falsified or fabricated research data that has been manipulated to support a particular hypothesis or outcome, thereby misrepresenting the actual findings. | | Manipulated photographs | The alteration or doctored images used in scientific publications to present a misleading representation of experimental results or observations. |

# The role and definition of a mentor A mentor is an experienced and trusted advisor whose role extends beyond skill transfer to include professional and academic growth, and the socialization into a discipline, often viewed as essential for success. ## 1. The definition and role of a mentor ### 1.1 Defining a mentor A mentor is defined as an experienced and trusted advisor [3](#page=3). ### 1.2 Key roles and relationships * Graduate thesis advisors and postdoctoral advisors are considered "natural" mentors in scientific contexts, assuming ideal circumstances [3](#page=3). * Universities and other institutions are increasingly formalizing mentor-trainee relationships [3](#page=3). * Mentoring is often perceived as crucial for success, though many scientists achieve success without explicitly formalized mentoring relationships [3](#page=3). ### 1.3 Rationale for academic mentoring The fundamental purpose of academic mentoring is to foster the professional and/or academic development of individuals early in their careers. It aims to promote excellence in various aspects, including teaching and learning, research, and academic leadership [13](#page=13) [14](#page=14). ### 1.4 Characteristics of mentoring relationships * A mentoring relationship can be lifelong, adapting and changing over time, or consist of a series of relationships that align with evolving career and personal needs [14](#page=14). * While mentoring is a reciprocal relationship, it is inherently asymmetric due to the mentor's greater experience [14](#page=14). * Beyond imparting specific skills, mentoring involves teaching a trainee the norms and practices of a particular discipline, a process known as socialization [14](#page=14). * Individuals are likely to experience both being a mentee and a mentor at different points in their lives [14](#page=14). > **Tip:** While formal mentoring is gaining traction, remember that informal mentorships can also be highly effective. The key is the quality of guidance and support provided. > **Example:** A postdoctoral advisor guiding a graduate student through their first major research project, offering technical advice, career counseling, and insights into the unwritten rules of academic publishing, exemplifies a "natural" mentor. --- # Principles and expectations of good mentoring Good mentoring is essential for supporting the professional and academic growth of individuals early in their careers, fostering excellence in teaching, learning, research, and leadership. This relationship can be lifelong or a series of relationships adapting to changing needs. While inherently asymmetric due to the mentor's greater experience, mentoring involves teaching not only skills but also the norms and socialization of a discipline. It is crucial for both mentors and mentees to understand their roles and expectations for a beneficial experience [13](#page=13) [14](#page=14). ### 2.1 Core values of a mentor-trainee relationship The foundation of a strong mentor-trainee relationship rests on several core values: * **Focus on the mentee's benefit:** The relationship should be structured to maximize the mentee's advantage. This requires effective communication from both parties, with the mentee actively seeking guidance [16](#page=16). > **Tip:** Mentees should proactively communicate their needs and progress to their mentors. * **Mentor's focus on mentee success:** A good mentor prioritizes the mentee's development and success over advancing their own immediate goals. While mentors and mentees may have parallel goals (e.g., grant renewal for the mentor, graduation for the mentee), the mentor's primary objective should be the mentee's growth [16](#page=16). > **Example:** Both mentor and mentee desire the mentee's scientific success. The mentor seeks data for grant renewal and career advancement, while the mentee requires it for graduation, personal development, and their own career trajectory. A good mentor ensures the mentee's needs are met. * **Legacy and future scientists:** A mentor's success is significantly measured by the achievements of their mentees, who represent the next generation of scientists and contribute to the mentor's lasting impact [16](#page=16). ### 2.2 Mentee expectations Mentees typically have several key expectations from their mentors: * **Support for professional and academic growth:** Mentees expect guidance and opportunities to advance their careers and academic pursuits [13](#page=13). * **Development of skills and discipline norms:** Beyond technical skills, mentees anticipate learning the unwritten rules, culture, and expectations of their field [14](#page=14). * **A structured relationship for maximum benefit:** Mentees expect the mentoring dynamic to be organized in a way that directly contributes to their development [16](#page=16). * **Mentorship focused on their success:** Mentees anticipate that the mentor's primary motivation is their own advancement, rather than the mentor using the relationship for personal gain [16](#page=16). * **Clear communication channels:** Mentees expect to be able to communicate openly with their mentors about their progress, challenges, and goals [16](#page=16). ### 2.3 Mentor expectations While not explicitly detailed in the provided pages, implicitly, mentors may expect: * **Proactive engagement from mentees:** Mentors may expect mentees to take initiative in seeking advice and sharing their work [16](#page=16). * **Effort and commitment from mentees:** Mentors invest time and energy, and may expect a commensurate level of dedication from their mentees. * **Respect for the mentor's time and expertise:** A balanced relationship involves mutual respect for each other's commitments and knowledge. ### 2.4 Criteria and sources for choosing a mentor Selecting the right mentor is a critical step, and several factors should be considered: * **Alignment of goals and interests:** Seek mentors whose research interests, career trajectory, and values align with your own [15](#page=15). * **Mentoring style:** Consider mentors known for their supportive and developmental approach, rather than those who might replicate less effective past mentoring experiences [16](#page=16). * **Availability and communication:** Assess a potential mentor's availability and their willingness to engage in regular, productive communication. * **Track record of mentee success:** Look for mentors who have a history of successfully guiding trainees through their academic and professional development [16](#page=16). Sources of information for choosing a mentor can include: * **Personal networks:** Talking to current and former students of potential mentors [15](#page=15). * **Faculty profiles and lab websites:** Reviewing a potential mentor's publications, research projects, and past student achievements. * **Departmental advisors and senior faculty:** Seeking recommendations and insights from experienced individuals in the field. ### 2.5 The optimal mentor-trainee ratio The provided document does not specify an optimal mentor-trainee ratio suggesting that the quality and effectiveness of the mentoring relationship are more important than a fixed numerical ratio [9](#page=9). ### 2.6 Addressing issues and changing mentors Situations can arise where the mentor-trainee relationship is not functioning effectively [10](#page=10). * **When something goes wrong:** If challenges emerge, open and honest communication is the first step. Both parties should attempt to address the issues collaboratively [10](#page=10). * **Conditions for changing mentors:** A change in mentor may be warranted if: * The relationship is consistently detrimental to the mentee's growth or well-being [11](#page=11). * There is a significant and unresolvable misalignment of expectations or goals [11](#page=11). * The mentor is unable or unwilling to provide adequate support or guidance [11](#page=11). > **Tip:** Before making a decision to change mentors, consider discussing the issues with a trusted advisor, departmental head, or ombudsperson. --- # Case studies in scientific mentoring scenarios This section explores common ethical dilemmas and challenges encountered in mentor-trainee relationships through a series of detailed case studies [17](#page=17). ### 3.1 Case study 1: Communication challenges in mentoring This case highlights the importance of open and timely communication between a mentor and mentee, especially when personal circumstances may impact academic progress [17](#page=17). * **Scenario:** Sami Sessiz, a PhD student, struggles to meet project deadlines due to personal issues (wife's pregnancy, father's Alzheimer's) which he did not disclose to his mentor, Prof. Ak. After a year of modest progress, Sami informs Prof. Ak [17](#page=17). * **Mentor's Reaction:** Prof. Ak is upset by the delayed communication, viewing it as a compromise to research progress. He reassigns Sami to a less demanding project, transfers the original experiments to another student, and defers decisions about Sami's authorship until the project's end [17](#page=17). * **Key Issues for Discussion:** * Whether Sami could have handled the situation differently [17](#page=17). * If changing mentors is advisable for Sami [17](#page=17). * The appropriateness of Prof. Ak's actions [17](#page=17). * Potential compromises Sami could propose to continue his original project [17](#page=17). > **Tip:** Effective mentoring requires a foundation of trust and open communication. Mentees should feel comfortable sharing significant personal challenges that might affect their work, and mentors should respond with understanding and a willingness to find solutions. ### 3.2 Case study 2: Plagiarism and mentor accountability This case examines plagiarism within a research proposal and the extent to which a mentor is responsible for a trainee's misconduct [18](#page=18). * **Scenario:** Prof. Ak enlists his PhD student, Hikmet Hile, to help write a research proposal, seeing it as a valuable learning experience and a path to continued funding. Hikmet writes the background, literature review, and preliminary results sections. The proposal is submitted, but later cancelled due to plagiarized content: verbatim paragraphs from a review article and an illustration taken from another lab's website without attribution. The Ethics Committee is notified [18](#page=18). * **Prof. Ak's Response:** Prof. Ak plans to deflect blame onto Hikmet, arguing he could not be expected to detect such plagiarism and that Hikmet should be punished, not him [18](#page=18). * **Key Issues for Discussion:** * How a Dean should respond to Prof. Ak's arguments [18](#page=18). * The degree of mentor accountability for trainee misconduct [18](#page=18). * Advice for Prof. Ak on handling the situation with Hikmet [18](#page=18). > **Tip:** Mentors have a responsibility to ensure the integrity of their lab's work. This includes educating trainees about academic integrity and thoroughly reviewing all submitted work for potential misconduct, even if the trainee is experienced. ### 3.3 Case study 3: Trust and reproducibility in data validation This case explores the ethical considerations of a mentor asking multiple students to independently verify crucial data, and the potential impact on trust within the lab [19](#page=19). * **Scenario:** Prof. Ak's student, Zeki Akılı, presents exciting data suggesting a novel enzyme function with therapeutic potential. To confirm reproducibility, Prof. Ak asks another student, Gizem Gizli, to perform the same experiments, instructing her not to discuss it with anyone to ensure independent data. All data will be disclosed later [19](#page=19). * **Key Issues for Discussion:** * Justification of the advisor's actions [19](#page=19). * Reasons for or against these actions [19](#page=19). * Alternative methods for confirming reproducibility [19](#page=19). > **Tip:** While confirming reproducibility is vital, mentors should consider how such requests are framed to avoid undermining trust between lab members. Transparency about the process and purpose can be beneficial. ### 3.4 Case study 4: Introducing trainees to manuscript review This case discusses a mentor's approach to involving a student in the scientific peer-review process and the adequacy of the mentorship provided [20](#page=20). * **Scenario:** Prof. Ak invites his PhD student, Tolga Tembel, to review a manuscript for a scientific journal, seeing it as a development opportunity. Tolga critically evaluates the data, identifies a flaw in the statistical analysis of Figure 4, and spends a night reanalyzing the data, concluding the manuscript needs substantial revision. He submits his detailed report. The next day, Prof. Ak states he forwarded Tolga's comments to the journal with a rejection recommendation, expressing complete trust in Tolga's judgment without reading the report himself [20](#page=20). * **Key Issues for Discussion:** * Whether Prof. Ak's behavior demonstrates good mentorship [20](#page=20). * A better method for introducing Tolga to manuscript assessment [20](#page=20). > **Tip:** Mentoring involves more than delegating tasks. It includes guiding trainees through complex processes, providing constructive feedback on their work, and fostering critical thinking rather than simply accepting their conclusions without review. ### 3.5 Case study 5: Policies on romantic relationships in the lab This case examines a mentor's policy prohibiting romantic relationships among lab members and its ethical implications [21](#page=21). * **Scenario:** Prof. Ak has a rule for his graduate students: romantic relationships between lab members are forbidden, and if one develops, one partner must leave the lab. Sevgi Aşk, a prospective PhD student, objects, considering this an interference with private matters. Prof. Ak explains his policy stems from past experiences where such relationships caused tension, hostility, low productivity, and negatively impacted group morale [21](#page=21). * **Key Issues for Discussion:** * Mentorship responsibilities, ethics, and conflicts of interest in this scenario [21](#page=21). > **Tip:** While mentors aim to maintain a productive and harmonious lab environment, policies regarding personal relationships must be carefully balanced with respect for individual privacy and autonomy. ### 3.6 Case study 6: Romantic relationships and conflicts of interest This case presents a severe conflict of interest involving a mentor's romantic relationship with a student in his class, and an ethically questionable proposed solution [22](#page=22). * **Scenario:** Prof. Ahmet Ak is romantically involved with a graduate student, "Ateşböcegı," whom he is teaching in a cell biology course. To address the conflict of interest, Prof. Ak proposes giving his friend, Prof. Haydar Büyük, the answer key to the midterm exam to grade Ateşböcegı's exam, while Prof. Ak grades the other students' exams. Prof. Ak also intends to inform the Department Chair to avoid future assignments that could create academic or working relationships between him and his girlfriend [22](#page=22). * **Key Issues for Discussion:** * Comments, advice, or suggestions for Prof. Ak if one were Haydar Büyük [22](#page=22). > **Tip:** Any personal relationship between a mentor and mentee, especially when the mentor is in a position of direct academic or professional authority over the mentee, creates significant ethical concerns and requires careful management to avoid exploitation and ensure fairness. --- ## Common mistakes to avoid - Review all topics thoroughly before exams - Pay attention to formulas and key definitions - Practice with examples provided in each section - Don't memorize without understanding the underlying concepts

| Term | Definition | |------|------------| | Mentor | An experienced and trusted advisor who guides and supports an individual, typically in their academic or professional development. This relationship is crucial for the growth of individuals in the early stages of their careers. | | Mentor-trainee relationship | A dynamic where an experienced individual (mentor) provides guidance, support, and knowledge transfer to a less experienced individual (trainee) to foster their professional and academic growth. This relationship is inherently asymmetric due to the mentor's greater experience. | | Academic mentoring | The formal or informal process of supporting the professional and/or academic development of individuals at the beginning of their careers, aiming to promote excellence in teaching, learning, research, and academic leadership. It can be a lifelong connection or a series of relationships. | | Scientific Conduct | The principles and practices that guide ethical and responsible behavior in scientific research, encompassing honesty, integrity, and accountability in all aspects of the research process. | | Graduate thesis advisor | An academic faculty member who supervises a student's master's or doctoral thesis research, providing guidance on the research project, methodology, and writing. They often serve as a primary mentor during a student's postgraduate studies. | | Postdoctoral advisor | A senior researcher who supervises a postdoctoral fellow, providing guidance on research projects, career development, and skill enhancement after the fellow has obtained their doctorate. This role is crucial for advancing a researcher's independent career. | | Plagiarism | The act of presenting someone else's work, ideas, or words as one's own without proper attribution, which is a serious breach of academic and scientific integrity. | | Conflict of interest | A situation where an individual's personal interests, such as financial gain or relationships, could compromise their professional judgment or objectivity in their work, particularly in academic or research settings. | | Ethics Committee | A body established within an institution to review and approve research proposals involving human or animal subjects, ensuring that the research is conducted ethically and in compliance with relevant regulations. They also investigate allegations of misconduct. | | Novelty | The quality of being new or original, especially in the context of scientific research, referring to a discovery, method, or idea that has not been previously described or known. | | Significance | The importance or impact of scientific findings or research work, referring to its potential contribution to knowledge, practical applications, or advancement of a field. |

# Scientific conduct and data documentation This section outlines the fundamental principles of scientific conduct, emphasizing the critical role of data recording, documentation, and ownership in ensuring research integrity and reproducibility. ### 1.1 The necessity of data documentation Data documentation is presented as a crucial requirement in scientific research, serving as the foundation for understanding and verifying scientific work [18](#page=18) [19](#page=19). ### 1.2 Principles of data recording and documentation Proper recording and documentation of data are essential components of scientific conduct [20](#page=20) [21](#page=21). #### 1.2.1 The lab book: a historical and practical tool The lab book, particularly following a "ZEN approach," is highlighted as a vital tool for recording scientific activities. It is not merely a record of what was done, but a comprehensive account that facilitates scientific rigor and integrity [22](#page=22). ##### 1.2.1.1 Content of a useful lab book A useful lab book should explain: * What you did [23](#page=23). * Why you did it [23](#page=23). * How you did it [23](#page=23). * When you did it [23](#page=23). * Where materials are located [23](#page=23). * What happened, including what did not happen [23](#page=23). * Your interpretations of the results [23](#page=23). * Contributions of others involved [23](#page=23). * What the next steps are [23](#page=23). ##### 1.2.1.2 Characteristics of good lab books Good lab books possess several key characteristics: * **Legibility:** They are legible and ideally written in ballpoint pen ink [24](#page=24) [25](#page=25). * **Organization and Up-to-dateness:** They are well-organized and kept current [24](#page=24). * **Accuracy and Completeness:** They are accurate and complete, including page numbers [24](#page=24) [25](#page=25). * **Inclusion of Data:** Data, such as photos, should be affixed to the pages [24](#page=24) [25](#page=25). * **Reproducibility:** They must allow for the repetition of experiments [24](#page=24) [25](#page=25). * **Compliance:** They should be compliant with funding agency and institutional requirements [24](#page=24) [25](#page=25). * **Accessibility:** They must be accessible to authorized persons [24](#page=24) [25](#page=25). * **Storage and Backup:** They need to be stored properly and appropriately backed up [24](#page=24) [25](#page=25). * **Witnessing:** They should be properly witnessed when necessary [24](#page=24) [25](#page=25). * **Ownership Recognition:** They must be properly recognized as the property of the institution [24](#page=24) [25](#page=25). * **Record of Contributions:** They serve as the ultimate record of scientific contributions [24](#page=24) [25](#page=25). > **Tip:** Lab books are not intended to be "neat books" but rather accurate and thorough records of experimental work [25](#page=25). ##### 1.2.1.3 Other considerations for lab books Further considerations for lab books include: * **Patents:** Patent applications require detailed records extending over significant periods [25](#page=25). * **Writing Tools:** Pens are preferred over pencils [25](#page=25). * **Error Correction:** Mistakes should be crossed out rather than erased or covered with correction fluid [25](#page=25). * **Format:** The choice between bound, spiral, or electronic lab books should be considered [25](#page=25). #### 1.2.2 Types of lab books and their trade-offs Different types of lab books offer distinct advantages and disadvantages: * **Bound/Stitched Books:** * **Advantages:** No lost pages; legally more convincing; difficult to alter [26](#page=26). * **Disadvantages:** Not logically organized; requires cross-referencing [26](#page=26). * **Spiral Bound/Loose Leaf Books:** * **Advantages:** Organized by experiment; data stored together [26](#page=26). * **Disadvantages:** Pages may get lost; difficult to authenticate [26](#page=26). * **Electronic Books:** * **Advantages:** Easy to search, read, and store; can offer security features [26](#page=26). * **Disadvantages:** Risk of corrupted files; potential software compatibility issues [26](#page=26). #### 1.2.3 The level of detail required in documentation The level of detail in documenting experiments is paramount. Specific information required includes: * **Reagents:** Source, product/order number, lot number, expiration date, and storage conditions [28](#page=28). * **Solutions:** How they were made and stored [28](#page=28). * **Cells:** Type, source, passage number, and growth medium used [28](#page=28). * **Instruments:** Type, name, location, and serial number of the equipment [28](#page=28). * **Settings and Procedures:** Detailed recording of instrument settings and experimental procedures is also critical [28](#page=28). > **Tip:** "Details – details – details" is the mantra for effective lab book entries [27](#page=27) [28](#page=28). ### 1.3 Data ownership Lab books are explicitly recognized as the property of the institution where the research is conducted. This underscores the importance of adhering to institutional policies regarding data management and intellectual property [24](#page=24) [25](#page=25). --- # Lab book examples and best practices This section explores exemplary lab books from historical scientists, outlines what constitutes useful and good lab book content, and discusses the pros and cons of different lab book formats [22](#page=22) [3](#page=3). ### 2.1 Exemplary historical lab books Classical examples of scientific inquiry documented in lab books include those by renowned figures such as Leonardo da Vinci, Isaac Newton, Charles Darwin, Thomas Edison, Marie Curie, and Albert Einstein. These historical records, while varying in style and detail, underscore the fundamental importance of meticulous documentation in scientific advancement [10](#page=10) [12](#page=12) [14](#page=14) [16](#page=16) [4](#page=4) [6](#page=6) [8](#page=8). ### 2.2 What useful lab books explain Useful lab books serve as comprehensive records of scientific activity, explaining key aspects of research. They should detail [22](#page=22) [23](#page=23): * **What you did:** A clear account of the procedures undertaken [23](#page=23). * **Why you did it:** The rationale and objectives behind the experiment [23](#page=23). * **How you did it:** The methodology and specific techniques employed [23](#page=23). * **When you did it:** The timeline of the research activities [23](#page=23). * **Where materials are:** Location and identification of all reagents, samples, and equipment used [23](#page=23). * **What happened (and what did not):** Comprehensive recording of all outcomes, including unexpected or negative results [23](#page=23). * **Your interpretations:** Personal analyses and conclusions drawn from the data [23](#page=23). * **Contributions of others:** Acknowledgment of any assistance or input from colleagues [23](#page=23). * **What’s next:** Future directions, follow-up experiments, or unanswered questions [23](#page=23). > **Tip:** Including what *did not* happen is as crucial as recording what did, as negative results can also be scientifically significant [23](#page=23). ### 2.3 Characteristics of good lab books Beyond simply recording what was done, good lab books possess several key characteristics that enhance their utility and integrity [24](#page=24) [25](#page=25): * **Legibility:** Entries must be clear and readable, ideally written in permanent ink like ballpoint pen, not pencil [24](#page=24) [25](#page=25). * **Organization:** The book should be well-structured and kept up-to-date to facilitate easy retrieval of information [24](#page=24) [25](#page=25). * **Accuracy and Completeness:** All entries must be factually correct and thoroughly detailed, including page numbering [24](#page=24) [25](#page=25). * **Inclusion of Data:** Direct incorporation of raw data, such as photographs or printouts, affixed to pages [24](#page=24) [25](#page=25). * **Reproducibility:** The documentation should be sufficient to allow for the repetition of experiments by oneself or others [24](#page=24) [25](#page=25). * **Compliance:** Adherence to requirements set by funding agencies and institutional policies [24](#page=24) [25](#page=25). * **Accessibility:** Records should be available to authorized individuals [24](#page=24) [25](#page=25). * **Proper Storage and Backup:** Secure storage practices and appropriate backup procedures for digital records [24](#page=24) [25](#page=25). * **Witnessing:** Proper witnessing of entries when required, particularly for patent or intellectual property purposes [24](#page=24) [25](#page=25). * **Property Acknowledgment:** Recognition that the lab book is the property of the institution [24](#page=24) [25](#page=25). * **Ultimate Record:** The lab book represents the definitive account of scientific contributions [24](#page=24) [25](#page=25). > **Tip:** A lab book is a record of research, not necessarily a polished narrative; focus on clear, accurate data over aesthetic perfection. Mistakes should be crossed out, not erased or whitened out [25](#page=25). ### 2.4 Types of lab books: advantages and disadvantages Different formats of lab books offer distinct benefits and drawbacks: #### 2.4.1 Bound or stitched books * **Advantages:** * Minimizes the risk of lost pages [26](#page=26). * Considered more legally convincing due to tamper-resistance [26](#page=26). * Difficult for pages to be surreptitiously altered or removed [26](#page=26). * **Disadvantages:** * May not be logically organized by experiment, requiring significant cross-referencing [26](#page=26). * Difficult to integrate supplementary materials like printouts without disorganization [26](#page=26). #### 2.4.2 Spiral-bound or loose-leaf books * **Advantages:** * Can be organized by experiment, keeping related data together [26](#page=26). * Easier to integrate supplementary data by inserting pages [26](#page=26). * **Disadvantages:** * Pages are more susceptible to getting lost or detached [26](#page=26). * May be more difficult to authenticate as a complete and unaltered record [26](#page=26). #### 2.4.3 Electronic lab books * **Advantages:** * Facilitates easy searching and retrieval of information [26](#page=26). * Generally easy to read and view [26](#page=26). * Offers efficient storage capabilities [26](#page=26). * Can incorporate robust security features [26](#page=26). * **Disadvantages:** * Vulnerable to corrupted files and data loss [26](#page=26). * Requires attention to software compatibility issues over time [26](#page=26). * Authentication can be complex if not managed properly [26](#page=26). > **Tip:** Regardless of the format chosen, ensure your lab book allows for the repetition of experiments and is compliant with all relevant institutional and funding requirements. Patents, for instance, may require records that can be defended for up to 23 years after filing [24](#page=24) [25](#page=25). --- # Data ownership and ethical considerations This topic explores the multifaceted issues surrounding data ownership and ethical conduct in scientific research through detailed case studies. ### 3.1 Data ownership Data ownership refers to who has the ultimate right to control, use, and disseminate the data generated during research. In a laboratory setting, the principal investigator (PI) typically holds institutional responsibility for the research and its associated data, even if generated by students or staff [29](#page=29). #### 3.1.1 Case study 1: Ownership This case involves a laboratory using new gel electrophoresis technology. Students provided photographs of their results to a company representative in exchange for a dinner. The PI, Prof. Ak, was unaware of this arrangement and did not benefit from it. * **Implications for data ownership and record keeping:** The scenario highlights that data generated within a lab usually belongs to the institution or the PI, not the individual students who collected it. Sharing data or images without the PI's knowledge or consent can lead to ethical breaches and potential conflicts. Proper laboratory record-keeping should include clear protocols for data sharing and external requests. * **Potential actions for Prof. Ak:** Prof. Ak could address the breach of protocol with his students, reiterating the lab's policies on data sharing and external engagement. He might also consider establishing clearer guidelines for such interactions in the future. ### 3.2 Primary data preservation Primary data refers to the original, raw data collected during an experiment. Its accurate preservation and documentation are crucial for scientific integrity. #### 3.2.1 Case study 2: Primary data Zehra Zehir, a graduate student, stored stained and dried polyacrylamide gels in plastic bags taped to her lab book pages, considering them primary data. Her mentor, Prof. Ak, ordered her to stop, citing the neurotoxic nature of polyacrylamide and instructing her to photograph the gels instead, considering the photographs as primary data. Zehra disagreed, believing the gels were more accurate and manipulation-proof than photographs, and that sealing them posed no danger. * **Advice for Zehra:** * **Safety first:** Prof. Ak's concern about the neurotoxic nature of polyacrylamide is valid and aligns with proper laboratory safety protocols. The gels should be disposed of as toxic waste [31](#page=31). * **Appropriate data recording:** Photographs, when taken with appropriate resolution and documentation, can serve as effective primary data records. However, Zehra's concern about manipulation is also valid. * **Documentation of process:** The key is meticulous documentation. This includes photographing the gels with proper scale and labeling, keeping high-resolution digital images, and recording all relevant experimental details in the lab notebook. The lab notebook should clearly reference these images and their location. * **Professional communication:** Zehra should engage in a constructive dialogue with Prof. Ak, acknowledging his safety concerns while articulating her own well-founded concerns about data integrity and proposing robust documentation methods that satisfy both safety and scientific rigor. ### 3.3 Authorship disputes Authorship in scientific publications signifies intellectual contribution and responsibility. Disputes can arise over who qualifies for authorship and the order of authors. #### 3.3.1 Case study 3: Authorship Didem Dirençli, a PhD student, obtained exciting results on ion channel function. A manuscript was submitted. While under review, Prof. Ak received another manuscript suggesting a key finding in Didem's work was incorrect. He questioned Didem, whose lab book records were incomplete. Prof. Ak suggested further experiments, which Didem refused due to her imminent graduation. The manuscript was accepted. Prof. Ak then insisted Didem withdraw the paper, but she refused. Their argument led to Prof. Ak requesting his name and grant reference be removed, and the paper was published with Didem as sole author. Subsequently, Prof. Ak's lab repeated the experiments, found Didem's findings incorrect, and published this correction. * **Prof. Ak’s handling of the situation:** Prof. Ak's initial handling was problematic. His questioning of Didem without revealing his source could be seen as confrontational. His insistence on removing his name after submission and acceptance, while perhaps stemming from a desire for accuracy, created a difficult situation for Didem and potentially damaged her career progression. His subsequent publication of contradictory results, while scientifically necessary, further complicated the relationship and the perception of the original publication. * **Alternative actions:** * Prof. Ak could have immediately and transparently discussed the concerns raised by the reviewed manuscript with Didem, explaining the source of the concern and the potential implications. * He could have collaboratively planned a rapid set of validation experiments to be conducted before the paper's publication, if time permitted. * If Didem was unwilling or unable to perform further experiments, a more collaborative discussion about the paper's future might have been beneficial, exploring options like revision or a joint correction. * The removal of his name and grant was a drastic step that ultimately left Didem solely responsible for potentially flawed data. * **Scientific misconduct:** The scenario raises questions about potential scientific misconduct, specifically regarding: * **Data Integrity:** The "incomplete and sloppy" lab records by Didem could suggest a lack of rigor in data recording. * **Falsification/Fabrication:** While not explicitly stated, the later findings by Prof. Ak's lab suggest that Didem's original results might have been incorrect, raising questions about whether she knowingly or unknowingly presented flawed data. * **Authorship Issues:** While Didem was removed as an author, the initial submission with potential inaccuracies and the subsequent dispute over authorship can be ethically complex. The decision to remove his name as a co-author after acceptance and while the paper was under review, citing concerns about accuracy, might be scrutinized, as authors are responsible for the content of their publications. ### 3.4 Data selectivity and manipulation The ethical presentation of scientific data requires transparency and honesty. Selectively presenting or manipulating data to achieve desired outcomes is a form of scientific misconduct. #### 3.4.1 Case study 4: Selectivity Asst. Prof. Serhat Saklamak admitted to manipulating a DNA gel image in his manuscript draft to "underexpose" smaller, unexpected DNA fragments. He did this to protect his hypothesis from being "scooped" while he conducted further work. He planned to include a note in the legend about "minor signals of unexplained origin." Prof. Ak cautioned him that this was data falsification and suggested electronically cropping the image to exclude the unexpected fragments, arguing no explanation would then be necessary. * **Comment on Serhat’s actions and Prof. Ak’s alternative solution:** * **Serhat's actions:** Serhat's attempt to hide or obscure data is a clear instance of data manipulation and falsification, which undermines scientific integrity. His rationalization of protecting his work from competition is a common, but ethically unacceptable, justification for such actions. The proposed figure legend, while acknowledging the presence of signals, still attempts to downplay their significance and is not a substitute for transparent data presentation. * **Prof. Ak’s alternative solution:** Prof. Ak’s suggestion to "electronically crop" the image is also ethically problematic. While it avoids the direct manipulation of the original image (like underexposing), it still amounts to selective presentation of data by omitting relevant bands. This is akin to a form of data suppression and misrepresentation, as it presents a partial truth. It might be less egregious than direct alteration, but it still fails to be fully transparent. * **Advice if they sought opinion:** * **For Serhat:** You would advise Serhat that his actions constitute data falsification. The ethical obligation is to present all relevant data transparently, even if it complicates the narrative or requires further research. He should aim to include the unexpected fragments and discuss their implications and the planned future work in the manuscript. * **For Prof. Ak:** You would advise Prof. Ak that his suggested "cropping" is also an unethical practice. True scientific transparency means presenting all pertinent findings. Instead of suggesting methods to hide data, he should guide Serhat towards ethically sound practices, such as including all data and clearly explaining its significance and the planned follow-up studies. The focus should be on robust scientific reporting, not on avoiding potential competition through data manipulation. ### 3.5 Recording of errors and corrections Maintaining an accurate record of research activities includes documenting any errors, unexpected events, and subsequent corrections made. #### 3.5.1 Case study 5: Confusion PhD student Samra Şaşkin collected blood samples from 100 patients, documenting clinical histories and meticulously labeling tubes. After assaying five racks, she discovered that labels had fallen off two racks in the freezer due to using inappropriate tape for cold temperatures. She re-numbered the tubes by rack location and repeated assays, finding the new results matched her original measurements. She then re-labeled the tubes with the assumed patient IDs. * **Advice on actions and lab book recording:** * **Document the event comprehensively:** Samra must immediately and thoroughly record the incident in her lab book. This includes: * The date the error was discovered. * The specific racks affected (racks 1 and 2). * The cause of the error (inappropriate tape for –70°C storage). * The steps taken to rectify the situation (re-numbering by rack location, repeating assays). * The validation process (comparing new results with original measurements). * The conclusion that the re-assigned patient IDs were likely correct based on the matching data. * The corrective action taken regarding the labeling (using appropriate tape). * **Ethical considerations:** While Samra's verification step is commendable and suggests her assumption about the IDs is likely correct, the initial loss of labeling and subsequent re-assignment, even with validation, introduces a layer of uncertainty. * **Transparency is key:** The lab book entry should be unambiguous. It should clearly state that the original labels were lost and that the patient IDs were re-established based on repeated assays and comparison. It should also note that while the data appears to match, there is an inherent assumption in the re-assignment. * **Witnessing:** It is good practice for such critical events to be witnessed and signed off in the lab book, as Samra requested from you. This adds an independent verification of the recorded events. * **Data integrity:** The goal is to ensure that the final dataset is as accurate and traceable as possible. The documentation should reflect the process, including the error and its correction, to maintain a transparent and auditable record. > **Tip:** Always use lab-appropriate materials for sample storage, especially at extreme temperatures. Thoroughly test new materials before committing large batches of samples. Proper labeling and tracking are foundational to reliable scientific data. > > **Example:** In Samra's case, the discovery of mismatched labels after assaying only three out of five racks means she could have potentially lost the identities of samples from racks 3, 4, and 5 if she hadn't discovered the issue before assaying them. This underscores the importance of regular checks of sample integrity. --- ## Common mistakes to avoid - Review all topics thoroughly before exams - Pay attention to formulas and key definitions - Practice with examples provided in each section - Don't memorize without understanding the underlying concepts

| Term | Definition | |------|------------| | Scientific Conduct | The adherence to ethical principles and best practices in the planning, execution, and reporting of scientific research. It encompasses integrity, honesty, and responsibility in all research activities. | | Data Recording | The systematic process of capturing observations, measurements, and experimental results in a structured and organized manner. This is a critical step in scientific research to ensure data is preserved for future analysis and verification. | | Data Documentation | The act of creating comprehensive records and explanations that accompany raw data. This includes details about the methods used, experimental conditions, and any observations made during data collection, ensuring transparency and reproducibility. | | Data Ownership | The legal and ethical rights and responsibilities associated with the data generated through research. This can involve who has the right to use, distribute, and control the data, often determined by institutional policies, funding agreements, and intellectual property laws. | | Lab Book | A detailed, chronological record of scientific experiments and observations maintained by a researcher. It serves as an essential tool for documenting procedures, results, hypotheses, and interpretations, ensuring a complete history of the research process. | | Primary Data | The original raw data collected directly from experiments or observations. This data is considered the most direct evidence of findings and is crucial for validating conclusions. | | Scientific Misconduct | Any action that violates widely accepted standards in the scientific community for proposing, performing, or reporting research. This includes fabrication, falsification, and plagiarism. | | Authorship | The designation of individuals who have made significant intellectual contributions to a published scientific work. Authorship criteria typically involve substantial contributions to conception or design, data acquisition, analysis, or interpretation, and drafting or revising the work. | | Fabrication | Making up data or results and recording or reporting them. This is a severe form of scientific misconduct. | | Falsification | Manipulating research materials, equipment, or processes, or changing or omitting data or results such that the research is not accurately represented in the research record. | | Plagiarism | The appropriation of another person's ideas, processes, results, or words without giving appropriate credit. | | Gel Electrophoresis | A laboratory technique used to separate DNA, RNA, or protein molecules based on their size and electrical charge. It involves applying an electric field to a gel matrix, causing charged molecules to migrate. | | PCR (Polymerase Chain Reaction) | A laboratory technique used to amplify specific segments of DNA. It allows for the creation of millions of copies of a particular DNA sequence from a small sample. | | Figure Legend | A descriptive text that accompanies a figure, table, or other illustration in a scientific publication. It provides necessary context and explanation for the visual content. | | Witnessing | In the context of lab books, this is the process where an authorized individual signs and dates specific entries to attest to their authenticity and accuracy at a particular time. |

# Scientific reading strategies This section details how scientists typically read research papers, outlining both minimalist and extended approaches and explaining the focus within different sections of a paper [13](#page=13) [2](#page=2) [7](#page=7). ### 1.1 The minimalist approach to reading scientific papers The minimalist approach is the most common method employed by experienced scientists to quickly grasp the essence of a research paper. This strategy involves a selective reading of specific sections to efficiently gain understanding without necessarily reading the entire paper in detail [13](#page=13) [7](#page=7). #### 1.1.1 Key sections in the minimalist approach Scientists employing the minimalist approach typically focus on the following sections: * **Abstract:** The abstract provides a concise summary of the paper's content, outlining the problem addressed, the approach taken, the main results, and the interpretation of those results. It serves as a crucial first point of contact to understand what the paper is about [13](#page=13) [8](#page=8) [9](#page=9). * **Last paragraph of the introduction:** This section often summarizes the study's objectives and the authors' specific hypothesis or research question, giving the reader immediate insight into the paper's focus [10](#page=10) [13](#page=13). * **Figures and figure legends:** Figures are designed to visually represent the core data and findings of the study. The accompanying legends provide essential context and explanations for these visuals. These elements are often self-sufficient in conveying the main results [11](#page=11) [13](#page=13). * **Discussion:** This section presents the authors' interpretation of their data and findings. Reading the discussion allows readers to understand how the researchers perceive their own results and provides an opportunity for the reader to agree or disagree with the interpretation [12](#page=12) [13](#page=13). > **Tip:** The minimalist approach is efficient for scientists who need to stay abreast of a large volume of literature. It prioritizes key information to determine the relevance and impact of a paper quickly. ### 1.2 The extended approach to reading scientific papers While the minimalist approach is prevalent, scientists may sometimes opt for an extended reading strategy when specific circumstances warrant a more in-depth engagement with the paper [14](#page=14). #### 1.2.1 When to employ the extended approach An extended approach may be taken in the following situations: * **Reading the results section:** This is done when the authors' line of reasoning behind the experiments is difficult to understand or when the experimental setup is not immediately obvious from the figures and legends [14](#page=14). * **Reading the materials and methods section:** This is typically undertaken when the techniques described in the figure legends are confusing, incomplete, or particularly interesting. It is also a common practice if the reader intends to replicate or adapt the described techniques for their own research [14](#page=14). > **Tip:** The extended approach is crucial for deep understanding, reproducibility, and the adoption of new methodologies within the scientific community. ### 1.3 The social aspect of scientific publishing Scientific publishing involves a complex social system where the way papers are read and interpreted is influenced by conventions and norms within the scientific community. The minimalist approach itself is a reflection of these social dynamics, enabling efficient knowledge dissemination and critical evaluation [13](#page=13) [6](#page=6) [7](#page=7). --- # The scientific publishing process The scientific publishing process involves a structured workflow from authoring a manuscript to its final publication after peer review and editorial decisions [18](#page=18). ### 2.1 Overview of the publishing workflow The journey of a scientific paper begins with the author's writing process. This is followed by submission to a journal, where it undergoes initial format screening. If it passes this stage, it is assessed by a handling editor who makes an initial decision, which can lead to rejection or forwarding for peer review. Peer reviewers then evaluate the manuscript's quality and soundness. Based on reviewer feedback, the handling editor makes a decision, which could be acceptance, rejection, or a request for revisions. Revisions may involve multiple cycles of review and resubmission, potentially at the same journal or a different one. Once accepted, the paper moves into the production phase before final publication [18](#page=18) [19](#page=19) [20](#page=20) [21](#page=21) [22](#page=22) [23](#page=23) [24](#page=24) [25](#page=25) [29](#page=29) [61](#page=61). > **Tip:** Understanding this multi-stage process is crucial for authors to effectively navigate their work towards publication. ### 2.2 Step 1: Writing the manuscript The writing process is the foundational step in scientific publishing. A recommended order for writing can be following the structure of figures, tables, results, discussion, methods, introduction, conclusions, abstract, and title. This order allows for the clear presentation of data interpretation within the context of the research [18](#page=18) [26](#page=26) [27](#page=27) [28](#page=28). * **Figures and Tables:** Visual representations of data [27](#page=27). * **Results and Discussion:** Presentation and interpretation of findings [27](#page=27). * **Methods:** Detailed description of experimental procedures [27](#page=27). * **Introduction:** Background and rationale for the study [27](#page=27). * **Conclusions:** Summary of key findings and their implications [27](#page=27). * **Abstract:** A concise summary of the entire paper [27](#page=27). * **Title:** A clear and informative title [27](#page=27). A typical length for a submitted manuscript is between 25-30 pages, focusing on essential data. This includes a title page, a single-paragraph abstract (around 250 words), a 1.5-2 page introduction, 2-4 pages for methods, 10-20 pages for results and discussion, and a single-paragraph conclusion. The manuscript should include 6-9 figures, 1-3 tables, and 20-50 references. Letters or short communications have stricter length limitations, often around 3,000 words with 3-5 illustrations [50](#page=50). > **Tip:** Focus on presenting only essential data to maintain a concise and impactful manuscript. ### 2.3 Step 2: Deciding on the right journal Selecting the appropriate journal is a critical step, considering the vast number of scientific journals available (over 12,500). Journals vary in their scope, impact factor, and publication model (open access vs. subscription) [30](#page=30) [31](#page=31) [32](#page=32) [33](#page=33) [43](#page=43). #### 2.3.1 Journal impact and ranking Journal indices, such as the impact factor (IF), are used to gauge a journal's influence. The impact factor is calculated based on citation data within a specific period. For example, data from 2022 shows a large number of journals and their associated impact factors. Journals can be categorized by impact as high, medium, or low, and by scope as general or specific [32](#page=32) [34](#page=34) [35](#page=35) [36](#page=36). * **Impact Factor (IF):** A measure of the average number of citations received by articles published in that journal in a particular period [35](#page=35). * **Journal Rank:** Position of a journal within its subject category based on metrics like the impact factor [36](#page=36). #### 2.3.2 Publication models Journals operate under different publication models: * **Subscription-based:** Readers or institutions pay to access content [43](#page=43). * **Open Access:** Content is freely available to everyone, often funded by article processing charges (APCs) paid by authors or their institutions. Some open access journals are free for authors [33](#page=33) [43](#page=43). #### 2.3.3 Preprint publication Authors may also consider preprint publication, where their manuscript is made available online before formal peer review. This allows for wider dissemination and feedback [44](#page=44) [46](#page=46) [47](#page=47). #### 2.3.4 Costs associated with publication Publication in certain journals can incur significant costs, especially for open access models. These charges can range from hundreds to thousands of euros or dollars, depending on the journal's impact and publication model. For instance, a general, high-impact journal with print/online hybrid access might cost 8,500 euros, while a general, low-impact journal that is online only and open access might be free [64](#page=64). > **Example:** A specific, high-impact journal charging 7,600 euros for print/online hybrid publication vs. a general, low-impact journal with free online open access [64](#page=64). ### 2.4 Step 3: Reading the instructions Before submission, it is crucial to carefully read the journal's instructions for authors. These instructions detail requirements for manuscript length, formatting, article types, and submission procedures [49](#page=49) [50](#page=50). #### 2.4.1 Article types Journals typically publish several types of articles: * **Full articles / original articles:** Regular, substantial research papers [49](#page=49). * **Letters / rapid Communications / short communications:** Brief reports of significant and original advances [49](#page=49). * **Review papers / perspectives:** Summaries of recent developments on a topic, often invited and not presenting new data [49](#page=49). ### 2.5 Step 4: Editorial assessment and peer review After submission, the manuscript is first handled by an editor, who conducts a format screening. The editor assesses whether the paper meets the journal's scope, standards, and quality. If it passes this initial assessment, it is sent for peer review [18](#page=18) [20](#page=20) [21](#page=21) [22](#page=22). #### 2.5.1 Role of peer reviewers Peer reviewers, typically experts in the field, evaluate the manuscript based on several criteria [22](#page=22) [52](#page=52) [53](#page=53): * **Technical soundness:** Is the paper technically correct and well-executed [52](#page=52)? * **Convincing claims:** Are the claims supported by evidence and appropriately discussed in the context of existing literature [52](#page=52)? * **Clarity of writing:** Is the manuscript clearly written and accessible to the intended readership [52](#page=52) [53](#page=53)? * **Statistical analysis:** Is the statistical analysis sound [52](#page=52)? * **Originality and significance:** Does the paper present new and significant information [53](#page=53)? * **Methodology:** Is the research well-designed and are the methods appropriate [53](#page=53)? * **Results and conclusions:** Are results presented clearly and do conclusions align with the findings [53](#page=53)? * **Ethical concerns:** Are there any ethical issues related to the use of subjects [52](#page=52)? Reviewers provide recommendations, which can range from acceptance to rejection, with various levels of required revisions in between [24](#page=24) [63](#page=63). > **Tip:** Addressing reviewer comments thoroughly and respectfully is vital, even if you disagree. A well-written response letter explaining your revisions is crucial. #### 2.5.2 Historical examples of peer review Throughout scientific history, many impactful papers initially faced rejection or criticism from reviewers and editors. For instance, Albert Einstein and Nathan Rosen encountered issues with peer review for their work on gravitational waves. Similarly, groundbreaking papers on citric acid cycle, beta radiation, lasers, the endosymbiont hypothesis, cell cycle control, and PCR were initially rejected by journals. These examples highlight that the peer review process, while essential, can sometimes be flawed [54](#page=54) [55](#page=55) [56](#page=56) [59](#page=59) [60](#page=60). ### 2.6 Step 5: Revising the manuscript If a manuscript requires revisions, authors are expected to address the reviewers' comments and the editor's feedback. This often involves making minor or major changes and resubmitting the revised manuscript, sometimes to the same reviewer. This revision process can involve multiple cycles [19](#page=19) [25](#page=25). ### 2.7 Step 6: Editorial decision Following revisions and resubmission, the handling editor makes a final decision. This decision can be acceptance, rejection, or a request for further revisions. Editors aim to ensure the paper is technically sound, its claims are convincing and supported by data, and it is clearly written. An acceptance decision signifies that the paper meets the journal's standards for scope, quality, and technical correctness [23](#page=23) [24](#page=24) [61](#page=61) [62](#page=62) [63](#page=63). > **Example:** An editor might write that a manuscript is "beautifully written" and "so well-crafted that it may become a classic," leading to an acceptance without changes. Conversely, a paper might receive mixed reviews, with some praising it and others raising significant concerns requiring substantial revisions or even rejection [62](#page=62) [63](#page=63). ### 2.8 Step 7: Production and publication Once a manuscript is accepted, it proceeds to the production phase, which includes typesetting, proofreading, and final formatting before being published [23](#page=23) [29](#page=29) [61](#page=61). --- # Scientific journal metrics and evaluation This section delves into the various metrics used to assess the significance and impact of scientific journals, exploring their calculation, interpretation, and inherent limitations [66-81. ### 3.1 The impact factor The Impact Factor (IF) is a widely recognized metric for evaluating the relative importance of a scientific journal. It is calculated based on the number of citations received by articles published in a journal within a specific timeframe [70](#page=70) [71](#page=71). #### 3.1.1 Calculation of the Impact Factor The Impact Factor for a given year, say year X, is calculated as follows [71](#page=71): $$ IF_{year X} = \frac{\text{citations}_{year X}}{\text{publications}_{year X-1} + \text{publications}_{year X-2}} $$ This formula represents the average number of citations received in year X by articles published in the journal during the preceding two years (year X-1 and year X-2) [71](#page=71). **Example:** If a journal published 880 articles in year X-1 and 902 articles in year X-2, and these articles collectively received 74,090 citations in year X, the Impact Factor for year X would be calculated as [72](#page=72): $$ IF_{2017} = \frac{74,090}{880 + 902} = 41.577 $$ #### 3.1.2 Limitations of the Impact Factor Despite its widespread use, the Impact Factor has several significant limitations: * **Author self-citations:** Journals can artificially inflate their Impact Factor through self-citation practices [73](#page=73). * **Editor pressure:** Editors may pressure authors to cite articles from their own journal to boost the Impact Factor [73](#page=73). * **Disciplinary differences:** Citation rates vary significantly across different scientific disciplines, making direct comparisons problematic [73](#page=73). * **Exclusion of negative citations:** The Impact Factor does not account for citations that critically evaluate or refute a published work [73](#page=73). * **Influence of review articles:** Review articles, which tend to be heavily cited, can disproportionately influence a journal's Impact Factor [73](#page=73). * **Manipulation:** The metric is susceptible to various forms of manipulation, including citing oneself, padding introductions with citations, and forced citations [74](#page=74). > **Tip:** The "corrected" Impact Factor diagram illustrates various ways the metric can be influenced or distorted, highlighting issues like self-citations, forced citations, and questionable inclusion of articles [74](#page=74). #### 3.1.3 Criticism and alternatives The Impact Factor has faced considerable criticism from the scientific community for its perceived flaws and potential for misuse. This has led to discussions and the development of alternative metrics designed to offer a more nuanced evaluation of journal influence [75](#page=75). ### 3.2 The Eigenfactor score The Eigenfactor score is a metric that aims to assess the total importance of a scientific journal [80](#page=80). * **Consideration of citation origin:** It takes into account where citations come from, giving more weight to citations from journals that are themselves highly regarded [80](#page=80). * **Reflection of researcher access:** It reflects how frequently researchers access content from a particular journal [80](#page=80). * **Calculation period:** The Eigenfactor score is calculated based on citations received over a five-year period [80](#page=80). * **Disciplinary adjustment:** Eigenfactor scores are adjusted to account for citation differences across various disciplines [80](#page=80). Journals that generate a higher impact on the scientific field are awarded larger Eigenfactor scores [80](#page=80). ### 3.3 The Article Influence score The Article Influence (AI) score is a metric that scales the Eigenfactor score by the number of articles published by a journal. This allows for direct comparison of the AI score with the Impact Factor [80](#page=80). ### 3.4 The Immediacy Index The Immediacy Index measures the average number of times an article is cited in the same year it is published [81](#page=81). * **Indicator of rapid citation:** It indicates how quickly articles within a journal are being cited [81](#page=81). * **Discounting large journals:** Because it is a per-article average, the Immediacy Index tends to reduce the advantage typically held by larger journals over smaller ones [81](#page=81). * **Potential bias for frequently issued journals:** Journals that publish frequently may have an advantage, as articles published earlier in the year have a greater chance of being cited than those published later [81](#page=81). Journals that publish infrequently or have late publication schedules may exhibit lower Immediacy Indexes [81](#page=81). --- # Scientist evaluation metrics This section outlines metrics used to evaluate the scientific output and influence of individual researchers, with a primary focus on the h-index [82](#page=82). ### 4.1 The h-index The h-index is a bibliometric indicator designed to quantify both the productivity and citation impact of a researcher's publications. It is defined as the largest number $h$ such that the researcher has published $h$ papers that have each received at least $h$ citations. This metric aims to balance the number of publications with the impact of those publications as measured by citations [82](#page=82). #### 4.1.1 Definition and calculation The h-index calculation involves listing a researcher's publications and their corresponding citation counts. The index is then determined by finding the point where the number of papers is equal to or less than the number of citations each of those papers has received [82](#page=82). > **Tip:** To calculate the h-index manually, sort your publications by citation count in descending order. Then, find the highest number $h$ for which the $h$-th paper in the sorted list has at least $h$ citations. #### 4.1.2 Significance and application The h-index has become a widely adopted metric in academia for evaluating researchers, particularly for academic hiring, promotion, and grant applications. It provides a single, quantitative measure that is less susceptible to extreme values from a single highly cited paper or a large number of uncited papers compared to simpler metrics like total citations or total publications [82](#page=82) [86](#page=86) [87](#page=87) [88](#page=88). #### 4.1.3 Limitations and considerations Despite its popularity, the h-index has several limitations: * **Field dependency:** Citation practices vary significantly across different scientific disciplines. A high h-index in one field may not be comparable to the same h-index in another [84](#page=84) [85](#page=85). * **Career stage:** The h-index naturally increases over a researcher's career, making direct comparisons between early-career and established researchers difficult. * **Self-citation:** The index can be inflated through excessive self-citation. * **Bias towards certain publication types:** Review articles and highly cited seminal works can disproportionately influence the h-index. * **Lack of nuance:** It does not account for the quality of journals, the contribution of co-authors, or the specific impact of individual papers beyond their citation count. > **Tip:** When interpreting h-index values, always consider the researcher's field, career stage, and publication history. It is best used in conjunction with other qualitative and quantitative measures of scientific contribution. #### 4.1.4 Related metrics While the h-index is prominent, other metrics are also used to assess scientific output and influence, though they are not detailed in the provided text. These can include the i10-index (number of publications with at least 10 citations), total citation counts, and journal impact factors. The document focuses primarily on the conceptual basis of the h-index as proposed by Jorge E. Hirsch [82](#page=82). --- ## Common mistakes to avoid - Review all topics thoroughly before exams - Pay attention to formulas and key definitions - Practice with examples provided in each section - Don't memorize without understanding the underlying concepts

| Term | Definition | |------|------------| | Abstract | A concise summary of a research paper, typically including the problem addressed, the approach taken, the main results, and the interpretation of those results. | | Introduction (last paragraph) | The final section of the introduction, which usually sets the stage for the rest of the paper by outlining the specific research question, hypothesis, or the paper's main contribution. | | Figures and Figure Legends | Visual representations of data or concepts within a paper, accompanied by descriptive text (legends) that explains what the figure depicts, allowing it to be understood independently. | | Discussion | The section of a scientific paper where the authors interpret their findings, relate them to existing literature, discuss limitations, and suggest future research directions. | | Materials and Methods | The part of a scientific paper that details the experimental design, procedures, materials, and techniques used, enabling other researchers to replicate the study. | | Peer Review Process | A critical evaluation of a manuscript by experts in the same field to assess its validity, originality, significance, and quality before publication. | | Handling Editor | An editor at a scientific journal responsible for overseeing the review and publication process of submitted manuscripts, including assigning reviewers and making recommendations. | | Editorial Board | A group of scientists who provide expertise and guidance to a journal, often involved in setting editorial policies and making final publication decisions. | | Impact Factor (IF) | A journal metric that measures the average number of citations received by articles published in that journal over a specific period, indicating its relative importance within its field. The formula is generally: citations in year X divided by the sum of publications in year X-1 and X-2. | | Journal Rank | A classification of journals within a specific field based on their metrics, such as impact factor, allowing for comparison and categorization of journal prestige. | | Open Access | A publishing model where research articles are made freely available to the public online, often supported by author publication charges. | | Subscription Model | The traditional publishing model where access to journal content is granted to individuals or institutions upon payment of a subscription fee. | | Preprint Publication | The release of a research manuscript online before it has undergone formal peer review and publication in a journal. | | Article Types | Different formats of scientific publications, including full/original articles (detailed research), letters/rapid communications (brief, timely findings), and review papers/perspectives (summaries of existing research). | | Manuscript | The typed or written text of a research paper submitted for publication. | | h-index | A metric proposed by Jorge Hirsch to quantify an individual researcher's scientific output, defined as the number of papers with a citation count greater than or equal to that number. | | Eigenfactor Score | A metric that measures the total importance of a journal by considering the origin of citations and how frequently researchers access its content over a five-year period. | | Article Influence Score | A score that scales the Eigenfactor score by the number of articles published by a journal, making it directly comparable to the Impact Factor. | | Immediacy Index | The average number of times an article is cited in the same year it is published, indicating how quickly articles gain traction. |

# Understanding authorship in scientific conduct Authorship in scientific publications is a critical aspect of research integrity, carrying significant academic, social, and financial implications for individuals and the broader scientific community [13](#page=13). ### 1.1 The importance of authorship Authorship signifies intellectual contribution to a published study. It is a primary mechanism for assigning credit and recognition for research efforts. Beyond academic recognition, authorship plays a role in career progression, influencing graduation, promotion, and employment opportunities. It can also be linked to securing research funding and maintaining employment. Furthermore, authorship contributes to an individual's reputation and recognition within their field [10](#page=10) [11](#page=11) [13](#page=13) [9](#page=9). ### 1.2 Trends in scientific authorship The landscape of scientific authorship has evolved, showing a clear trend towards an increasing number of authors per publication. Data indicates a steady rise in the average number of authors from before 1975 through to recent years. This trend suggests a growing emphasis on collaborative research models. For instance, some studies now report an average of over 1.014 authors and involve as many as 71 different affiliations [12](#page=12) [7](#page=7). ### 1.3 Who should be an author? The definition of an author is generally understood as an individual who has made substantial intellectual contributions to a published study. This principle underscores the idea that authorship should be inclusive of all who meaningfully contribute to the research [13](#page=13) [8](#page=8). > **Tip:** Navigating the complexities of authorship can be challenging, often described as a "maze". It is essential to understand the criteria and expectations surrounding authorship to ensure fairness and integrity in scientific publishing [28](#page=28). --- # Criteria and contributions for authorship Authorship signifies substantial intellectual contributions to a published study and carries significant academic, social, and financial implications [13](#page=13). ### 2.1 Defining authorship An author is generally recognized as an individual who has made significant intellectual contributions to a published research study. These contributions are crucial for academic advancement, promotions, employment, research funding, and professional reputation [13](#page=13). ### 2.2 Core criteria for authorship Each author is expected to take public responsibility for the content of the published work. This responsibility encompasses several key components [14](#page=14): * **Conception or design, or analysis and interpretation of data, or both** [14](#page=14). * **Drafting the article or revising it for critically important intellectual content** [14](#page=14). * **Final approval of the version to be published** [14](#page=14). Participation solely in the collection of data does not, by itself, justify authorship [14](#page=14). The International Committee of Medical Journal Editors (ICMJE) recommends authorship be based on four criteria [15](#page=15): * Substantial contributions to the conception or design of the work; or the acquisition, analysis, or interpretation of data for the work; AND [15](#page=15). * Drafting the work or revising it critically for important intellectual content; AND [15](#page=15). * Final approval of the version to be published; AND [15](#page=15). * Agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved [15](#page=15). ### 2.3 Contributions that may justify authorship Several types of contributions are generally considered sufficient to warrant authorship, provided they involve substantial intellectual input [17](#page=17): * **Design:** The formulation of the study's conceptual framework and methodology [17](#page=17). * **Supervision:** Providing oversight and guidance for the research project [17](#page=17). * **Provision of Resources:** Supplying essential materials, reagents, animals, or novel tools needed for the research [17](#page=17). * **Provision of Material:** Similar to resources, this includes supplying specific items necessary for the study [17](#page=17). * **Data Collection:** Actively gathering data, particularly if novel methods are employed or a specific role is played, such as in statistics or imaging [17](#page=17). * **Analysis and/or Interpretation:** Performing analyses, including statistical analyses, of study data. Basic data analysis that requires minimal intellectual input may not suffice [17](#page=17) [18](#page=18). * **Literature Search:** Conducting comprehensive searches for relevant existing literature [17](#page=17). * **Writing:** Drafting the manuscript, which can warrant first authorship [17](#page=17) [18](#page=18) [24](#page=24). * **Critical Review:** Providing substantial feedback and commentary on the manuscript [17](#page=17) [18](#page=18). An idea alone may justify authorship only if it is highly original and unique. Supervision and other intellectual contributions can also justify authorship, assuming active involvement [18](#page=18). ### 2.4 Contributions that may not justify authorship Certain contributions, while important, typically do not warrant authorship on their own and are more appropriately acknowledged in the manuscript: * **Technical help:** Providing routine technical assistance without significant intellectual input [17](#page=17). * **Financial and material support:** Providing funding or general laboratory supplies that do not constitute unique resources for the specific project [17](#page=17) [18](#page=18). * **Administrative support:** Offering general organizational or managerial assistance [17](#page=17). * **Editorial contributions:** Assisting with the editing of the manuscript without critical intellectual input [17](#page=17). Merely mentoring or training another author does not automatically confer authorship unless a substantive contribution to the study itself is made. Honorary authorship, for individuals like lab chiefs or celebrities who have not contributed intellectually, is generally discouraged [18](#page=18). ### 2.5 Author order and responsibilities The order of authors on a publication should reflect their relative overall contributions to the manuscript. This order should be a joint decision among all co-authors, and authors should be prepared to explain the rationale behind the listed order [23](#page=23). #### 2.5.1 The first author The first author is typically the individual who contributed most significantly to the work, including writing the initial draft of the manuscript. This role usually involves developing the hypothesis, defining precise methods, participating substantially in data analysis and interpretation, and writing the paper [24](#page=24). #### 2.5.2 The lead/corresponding author The lead or corresponding author is responsible for ensuring that all co-authors review and approve the final version of the manuscript [25](#page=25). #### 2.5.3 Co-authors Co-authors should have made significant contributions to the planning and/or execution of the research, including the methods, procedures, and collection and analysis of data [25](#page=25). #### 2.5.4 The senior author The senior author typically formulated the original hypothesis or provided significant intellectual resources. This role also involves providing constructive criticism of the manuscript and accepting ultimate responsibility for the findings and the authorship. In some contexts, the "boss" who obtained funding may be listed last, even if they haven't read the paper [25](#page=25) [5](#page=5). ### 2.6 The CRediT system The Contributor Roles Taxonomy (CRediT) is a system designed to ensure visibility and diversity in research contributions, shifting the focus from pure authorship to contributorship. CRediT defines specific roles that individuals can fulfill in a research project, including [21](#page=21): * **Conceptualization:** Ideas; formulation or evolution of overarching research goals and aims [21](#page=21). * **Methodology:** Development or design of methodology; creation of models [21](#page=21). * **Software:** Programming, software development; designing computer programs; implementation of computer code and supporting algorithms; testing of existing code components [21](#page=21). * **Validation:** Verification of the overall replication/reproducibility of results/experiments and other research outputs [21](#page=21). * **Formal analysis:** Application of statistical, mathematical, computational, or other formal techniques to analyze or synthesize study data [21](#page=21). * **Investigation:** Conducting a research and investigation process, specifically performing experiments, or data/evidence collection [21](#page=21). * **Resources:** Provision of study materials, reagents, laboratory samples, animals, instrumentation, computing resources, or other analysis tools [21](#page=21). * **Data Curation:** Management activities to annotate, scrub, and maintain research data for initial and later reuse [21](#page=21). * **Writing - Draft:** Preparation, creation, and/or presentation of the published work, specifically writing the initial draft [21](#page=21). * **Review / Editing:** Preparation, creation, and/or presentation of the published work by those from the original research group, specifically critical review, commentary, or revision [21](#page=21). * **Visualization:** Preparation, creation, and/or presentation of the published work, specifically visualization/data presentation [21](#page=21). * **Supervision:** Oversight and leadership responsibility for research activity planning and execution, including mentorship [21](#page=21). * **Project administration:** Management and coordination responsibility for research activity planning and execution [21](#page=21). * **Funding acquisition:** Acquisition of financial support for the project leading to the publication [21](#page=21). An example of how CRediT roles can be assigned to authors for a specific publication is provided. For instance, "Conceptualization: JD, MM; Investigation: JD, FM, RF, JRL; Formal analysis: JD, FM, RF, KBS, JRL, MM; Visualization: JD, FM, MM; Resources: LR; Writing – Original draft: JD, JRL, MM; Review / Editing: JD, FM, RF, KBS, LR, JRL, MM; Funding acquisition: MM" [22](#page=22). --- # Ethical dilemmas and case studies in authorship This topic explores various scenarios and ethical challenges related to authorship, including guest and gift authorship, ghost authorship, and specific case studies that illustrate disputes over credit and responsibility. ### 3.1 Defining authorship and related ethical considerations Authorship in scientific publications signifies accountability and credit for research work. However, disputes and ethical dilemmas frequently arise concerning who qualifies for authorship and the appropriate order of names on a publication [27](#page=27). #### 3.1.1 Types of non-meritorious authorship * **Guest authors:** Individuals listed as authors who do not meet the accepted authorship criteria, often due to their seniority, reputation, or perceived influence [27](#page=27). * **Gift authors:** Individuals listed as authors without meeting authorship criteria, typically as a personal favor or in exchange for payment [27](#page=27). * **Ghost authors:** Individuals who meet authorship criteria but are not listed on the publication [27](#page=27). #### 3.1.2 Establishing authorship criteria Generally, authorship should be based on intellectual and conceptual contributions to the work being prepared for publication. Technical assistance, regardless of its complexity or extent, is typically not considered sufficient for authorship [35](#page=35). ### 3.2 Case studies illustrating authorship dilemmas The following case studies highlight common ethical challenges encountered in scientific authorship. #### 3.2.1 Case study 1: Help and technical contributions **Scenario:** Ayşe Küçük, a postdoc with expertise in gene cloning, trained a senior graduate student in these methods. The student, under Ayşe's supervision, generated a cDNA library and isolated a gene. The resulting manuscript listed the principal investigator and the student as authors, with Ayşe acknowledged but not credited as an author. Ayşe argued for authorship based on her training and supervision, while Prof. Ak maintained that her contribution was purely technical and did not warrant authorship, adhering to standards that prioritize intellectual contributions over technical assistance [35](#page=35). * **Analysis:** Ayşe has a case for authorship if her contribution involved significant intellectual input beyond mere technical instruction, such as designing the experiments or interpreting the results. If her role was solely to teach techniques and supervise their execution, Prof. Ak's stance aligns with common authorship guidelines [35](#page=35). #### 3.2.2 Case study 2: First authorship and CV representation **Scenario:** Dr. Birol Birinci, a PhD student, had several publications, two of which listed him as a co-first author, with the order decided by a coin toss. On his CV, Birol altered the order to place his name first for these papers to emphasize his contribution when applying for faculty positions. He worried his publication record appeared weak and lacked senior-authored papers [36](#page=36). * **Advice:** Misrepresenting author order on a CV is unethical and can damage credibility. Birol should accurately reflect the original author order and clearly state the nature of his co-first authorship. He could include a footnote or a statement explaining the circumstances of shared first authorship [36](#page=36). #### 3.2.3 Case study 3: Authorship credit and university affiliation **Scenario:** Prof. Fatih Küçük moved from A University to B University and found a completed manuscript in his former office, listing B University as his address. A published paper identical to this manuscript appeared in *Cell*, crediting B University and thanking a technician from A University in the acknowledgments. Prof. Ak suspected Fatih had published work conducted at A University under his new affiliation to appear more productive [37](#page=37). * **Advice for Department Chair:** The department chair should investigate the allegations of scientific misconduct thoroughly. This may involve reviewing Fatih's expense reports, lab notebooks, and communication logs from both universities. The department chair should consult institutional policies on research integrity and potentially involve a research integrity officer. Deliberately falsifying information regarding the location or timeline of research constitutes serious misconduct [37](#page=37). #### 3.2.4 Case study 4: Responsibility for image manipulation **Scenario:** A published paper co-authored by PhD student Mustafa, his mentor Prof. Ak, and another student faced allegations on a blog of manipulated Western blot images, with claims of "erased" lanes and "cut-and-pasted" bands. The authors admitted to editing the image for clarity but denied deceptive intent or altering the data [38](#page=38). * **Analysis of Strategies:** * **Defending on the blog:** While defending their position is an option, it lacks the formal rigor of a peer review process and could escalate the conflict. * **Notifying the editor:** Informing the journal editor is a crucial step, as it allows the journal to initiate its own review and potentially issue a correction or retraction. * **Doing nothing:** Ignoring public allegations can be interpreted as an admission of guilt or a lack of accountability. * **Turning over materials to integrity officer:** This is a responsible and proactive approach that allows for an impartial investigation into the allegations [38](#page=38). * **Advice for Prof. Ak:** Prof. Ak should immediately notify the journal editor and the institutional research integrity officer. They should transparently provide all raw data and image preparation files for scrutiny. Even if they believe no wrongdoing occurred, cooperating fully with an investigation is essential for maintaining scientific integrity [38](#page=38). #### 3.2.5 Case study 5: Neglect and submission delays **Scenario:** Metin Miskin, a recent PhD graduate, left academia for a career as a skydiving pilot. A manuscript based on his thesis remained a preliminary draft. His former advisor, Prof. Ak, revised the manuscript, but Metin did not provide comments or consent to submit. Subsequently, similar results were published by another lab. Prof. Ak and a postdoc prepared a new manuscript with Metin as first author, but Metin remained unresponsive to submission agreements. A friend indicated Metin blamed Prof. Ak for delays and was intentionally hindering the publication [39](#page=39). * **Publishing the manuscript:** Prof. Ak cannot ethically submit or publish the manuscript if Metin, as a co-author, has not provided consent to submit. This would violate authorship agreements and potentially result in the journal retracting the paper. * **Authorship:** Metin should remain the first author as the work is based on his thesis. The postdoc would be an additional coauthor. * **Acknowledgments:** If Metin's data or previous work is essential for the new manuscript, it might be appropriate to include an acknowledgment for his original contribution, especially if he is not listed as an author on the revised manuscript due to non-responsiveness. However, given his current stance and stated intentions, this is complex [39](#page=39). #### 3.2.6 Case study 6: Authorship based on a suggested experiment **Scenario:** Prof. Hayvan asked Mustafa Can to perform an experiment before his thesis defense. Although the experiment's direct results were not new, a positive control led to a significant discovery about a ligand metabolite. Mustafa and Prof. Ak wrote a manuscript, but Prof. Hayvan demanded authorship, arguing his insistence on the experiment was a "significant idea" qualifying for authorship and expressed dissatisfaction with not being acknowledged [40](#page=40). * **Mentor's Response:** As Mustafa's mentor, the response should be to carefully evaluate Prof. Hayvan's claim against established authorship criteria. While suggesting an experiment is a contribution, it does not automatically qualify for authorship unless it involves significant intellectual input beyond merely requesting an action. * **Analysis and Actions:** The situation requires a balanced assessment. If Prof. Hayvan's suggestion was a mere procedural request with no intellectual contribution to the discovery itself, authorship is not warranted. However, if his suggestion was a novel idea that directly led to the experimental design and subsequent discovery, a co-authorship or at least an acknowledgment might be considered. In this case, the discovery stemmed from a positive control within the requested experiment, suggesting the "idea" was more about execution and observation of unexpected results, rather than a conceptual breakthrough proposed by Hayvan. * **Why:** Authorship should reflect substantive intellectual contributions. While Prof. Hayvan initiated the experiment, the significant discovery emerged from an unexpected result and Mustafa's subsequent interpretation and rigorous data collection. The mentor should discuss the situation with both Mustafa and Prof. Hayvan, explaining the authorship criteria and proposing an appropriate course of action, which could range from acknowledgment to declining authorship based on the depth of intellectual contribution [40](#page=40). #### 3.2.7 Case study 7: Intellectual origin versus flawed execution **Scenario:** PhD student Murat Orman hypothesized a mechanism for a knockout mouse phenotype and tested it, producing reproducible data. However, upon leaving the lab, another student, Ayşe, could not replicate Murat's findings and discovered his data were flawed due to improper assay conduct. Ayşe's subsequent experiments provided an alternative explanation for the phenotype. Murat requested coauthorship based on his initial hypothesis, despite the flawed execution of his work [41](#page=41). * **Prof. Ak's query:** Prof. Ak seeks advice on whether Murat has a case for authorship. * **Advice:** Murat does not have a strong case for authorship. While his initial hypothesis was an intellectual contribution, the actual experimental work was flawed and did not lead to valid findings. Authorship is generally based on substantial contributions to the conception, design, data acquisition, analysis, and interpretation of the work being published. Murat's work did not meet these criteria due to its irreproducibility and fundamental errors in execution. Ayşe's diligent work in correcting the errors and conducting rigorous experiments forms the basis of the publication [41](#page=41). * **Why:** The principle is that authorship credits those who have contributed meaningfully and accurately to the research. Murat's flawed execution invalidated his experimental results, and his hypothesis, while potentially a starting point, did not lead to the established findings without significant correction and further research by others [41](#page=41). #### 3.2.8 Case study 8: Coercive citation and journal policies **Scenario:** Jane Doe, an assistant professor, was invited to serve on the peer review board of an online journal. The publisher encouraged her to submit her own papers and cite relevant publications from the journal to increase its impact factor. A board member confirmed this practice, calling it "coercive citation," and mentioned pressure from the editor-in-chief to cite previous journal articles [42](#page=42). * **Advice and Take:** Jane should decline the invitation to serve on this journal's peer review board. The practice of coercive citation is highly unethical and undermines the integrity of the peer review process and scientific literature. Such practices artificially inflate a journal's impact factor and can mislead researchers and funding agencies. Accepting this position would compromise her own ethical standing and scientific reputation. * **Why:** Scientific integrity demands honest representation of research. Coercive citation manipulates citation metrics for personal or institutional gain, which is a form of academic dishonesty. Jane's professional development should be built on genuine contributions and ethical practices, not on engagement with a journal that promotes such dubious methods [42](#page=42). --- ## Common mistakes to avoid - Review all topics thoroughly before exams - Pay attention to formulas and key definitions - Practice with examples provided in each section - Don't memorize without understanding the underlying concepts

| Term | Definition | |---|---| | Authorship | The act or status of being credited as an author of a published work, carrying significant academic, social, and financial implications for individuals involved in research. | | First author | The author who typically made the most substantial intellectual contributions to the work, often including drafting the initial manuscript and playing a lead role in the research design and execution. | | Senior author | The author, usually the principal investigator or lab head, who often conceives the original hypothesis, provides intellectual resources, supervises the research, and takes ultimate responsibility for the findings and publication. | | Middle authors | Authors listed between the first and last authors, whose contributions vary but are generally less significant than those of the first or senior authors; often includes students and technical staff. | | Guest author | An individual listed as an author who does not meet the standard authorship criteria but is included due to their seniority, reputation, or perceived influence, which is generally considered unethical. | | Gift author | An individual listed as an author who has not met the authorship criteria but is included as a personal favor or in exchange for payment, which is a form of authorship misconduct. | | Ghost author | An individual who meets the authorship criteria and has made substantial intellectual contributions but is not listed as an author on the publication, often to omit them or for undisclosed reasons. | | ICMJE criteria | A set of four recommended criteria by the International Committee of Medical Journal Editors that define authorship, requiring substantial contributions to conception/design or data acquisition/analysis/interpretation, drafting/revising critically, final approval, and accountability for all aspects of the work. | | CRediT system | Contributor Roles Taxonomy, a system designed to provide standardized descriptions of the contributions of each author to a research publication, moving towards contributorship rather than solely authorship. | | Conceptualization | The process of originating or evolving the overarching research goals, ideas, and aims that guide a study. | | Investigation | The active process of conducting research, which specifically involves performing experiments, gathering data, or collecting evidence relevant to the study's objectives. | | Formal analysis | The application of quantitative methods, such as statistical, mathematical, or computational techniques, to analyze and synthesize the data collected during a research study. | | Data curation | The management activities required to annotate, clean, and maintain research data, including software code, to ensure its usability for initial research purposes and for future reuse. | | Supervision | The role of overseeing and providing leadership for the planning and execution of research activities, which often includes mentorship for researchers involved in the project. | | Funding acquisition | The process of securing the financial resources necessary to support and conduct a research project leading to a publication. | | Scientific misconduct | Intentional deviation from accepted practices in research, such as fabrication, falsification, or plagiarism, which can undermine the integrity of scientific research. | | Coercive citation | The practice of pressuring authors to cite previous publications from a specific journal to artificially inflate its impact factor, which is an unethical practice. |

# Data representation and its misrepresentation This topic explores how data can be altered during its representation, leading to potential inaccuracies, and defines key forms of research misconduct [1](#page=1) [2](#page=2) [3](#page=3). ### 1.1 Understanding data representation Data representation refers to the process where information is captured and recorded. However, during this process, several issues can arise [1](#page=1) [2](#page=2): * **Loss of data:** Information can be omitted or not captured [1](#page=1) [2](#page=2). * **Addition of data:** Extraneous or incorrect information may be introduced [1](#page=1) [2](#page=2). * **Change of data:** Original data can be altered [1](#page=1) [2](#page=2). * **Manipulation of data:** Data can be intentionally altered to create a misleading impression [1](#page=1) [2](#page=2). It is crucial to recognize that data itself is not the same as its representation. The way data is presented can significantly distort its original meaning or accuracy [2](#page=2). > **Tip:** Think of data representation like making a photocopy of a document. The photocopy is a representation, but it might be darker, lighter, have smudges, or miss certain details compared to the original document. The representation is not the original itself [1](#page=1). ### 1.2 Research misconduct definitions The Department of Health and Human Services defines research misconduct as the fabrication, falsification, or plagiarism that occurs when proposing, performing, or reviewing research results [3](#page=3) [4](#page=4). #### 1.2.1 Fabrication Fabrication involves creating entirely made-up data or results and then recording or reporting them as if they were genuine [4](#page=4). #### 1.2.2 Falsification Falsification entails manipulating research materials, equipment, or processes. It also includes the act of changing or omitting results in a way that leads to the research record not accurately reflecting the true findings [4](#page=4). #### 1.2.3 Plagiarism Plagiarism is defined as the unauthorized use of another person's ideas, processes, results, or words without giving them proper acknowledgment. This can manifest in various forms, sometimes referred to as "citation amnesia," "disregard syndrome," or "bibliographic negligence." [4](#page=4). --- # Scientific image manipulation: examples and consequences Scientific image manipulation refers to the alteration of digital images used in research, which can range from acceptable enhancement to serious misconduct, leading to significant consequences. Digital images are inherently easy to manipulate due to their digital nature [15](#page=15) [5](#page=5) [6](#page=6) [7](#page=7) [9](#page=9). ### 2.1 Types of image manipulation and their implications Image manipulation can create false impressions and mislead the scientific community and the public. While some adjustments may be made to improve clarity, others can obscure or fabricate data. The core ethical issue lies in whether the manipulation alters the scientific integrity of the data [10](#page=10) [11](#page=11) [12](#page=12) [13](#page=13) [14](#page=14) [24](#page=24) [9](#page=9). #### 2.1.1 Removal and addition of data A common form of manipulation involves the selective removal or addition of parts of an image. This can include [24](#page=24) [25](#page=25): * **Selective removal of parts:** Removing specific features or data points that do not support a hypothesis or that are considered artifacts [24](#page=24) [25](#page=25). * **Selective addition of parts:** Introducing elements or data that were not originally present in the image [25](#page=25). > **Tip:** The deliberate removal or addition of data fundamentally distorts the original findings and is considered scientific misconduct. #### 2.1.2 Inappropriate replication of data Another problematic form of manipulation is the inappropriate replication of data, where elements from one part of an image are duplicated to represent data in another part, creating the illusion of multiple, independent experimental results. This is particularly egregious when it involves 'loading controls' in experiments such as Western blots [26](#page=26). > **Example:** In a Western blot analysis, if the same band pattern for a loading control is presented for two different experiments (Experiment A and Experiment B) when they should be distinct, it suggests the data has been fabricated or manipulated to appear consistent [26](#page=26). #### 2.1.3 Selective adjustment of image parts Manipulation can also involve selectively adjusting specific parts of an image, such as altering contrast or brightness in particular regions to highlight certain features or obscure others. While global adjustments might be acceptable for improving overall image quality, localized adjustments can disproportionately affect specific data points and lead to misinterpretation [29](#page=29). #### 2.1.4 Exaggerated global adjustments Exaggerated global adjustments, such as drastically altering contrast or brightness across the entire image, can also distort the original data. This might make faint signals appear stronger or suppress background noise excessively, leading to an inaccurate representation of the experimental results [31](#page=31). #### 2.1.5 Ambiguity in image selection Sometimes, the manipulation lies in the selection of which image or part of an image to present, offering multiple options that may not accurately reflect the full data set [32](#page=32). ### 2.2 Examples of scientific image manipulation in publications Several published cases highlight instances of scientific image manipulation: * **Maile et al., Science, 2004** [16](#page=16) [17](#page=17). * **Beisel et al., Nature, 2002** [16](#page=16) [17](#page=17). * **Sanchez-Elsner et al., Science, 2006**: This case involved issues with image similarity and potential manipulation of gamma settings and aspect ratios (V/H stretching) [18](#page=18) [19](#page=19) [20](#page=20) [21](#page=21) [22](#page=22) [23](#page=23). * **Sud N et al., Am J Physiol Lung Cell Mol Physiol, 2008**: This study presented cases of inappropriate replication of data across different figures and panels [27](#page=27) [28](#page=28). * **Uittenbogaard A, et al., J Biol Chem, 2002** [38](#page=38) [39](#page=39) [40](#page=40). * **Sawada M, et al., Nat Cell Biol, 2003** [41](#page=41) [42](#page=42) [43](#page=43). * **Hwang WS, et al., Science, 2005** [44](#page=44). * **Kang J et al., Cell, 2005**: This involved enhanced contrast in original images, potentially obscuring details or making weak signals more prominent [45](#page=45) [46](#page=46) [47](#page=47). ### 2.3 Consequences of scientific image manipulation The consequences of scientific image manipulation can be severe, undermining the integrity of research and potentially leading to: * **Retractions of publications:** Journals may retract papers containing falsified or manipulated images [41](#page=41) [45](#page=45). * **Damage to scientific reputation:** Researchers involved in image manipulation face significant damage to their credibility and career prospects [37](#page=37). * **Loss of public trust:** Scientific misconduct involving image manipulation erodes public trust in science and its findings. * **Misguided future research:** Manipulated data can lead other researchers down incorrect paths, wasting resources and time. ### 2.4 Acceptable image manipulation practices Not all image manipulation is unethical. Certain practices are considered acceptable when they are used to enhance data clarity and are properly disclosed. These include: * **Adding arrows:** Using arrows to point out specific features of interest in an image [50](#page=50) [51](#page=51). * **Pseudocoloring particles:** Applying pseudocolors to highlight features like immunogold particles without altering the brightness of individual pixels [50](#page=50) [51](#page=51). It is crucial that any such enhancements are disclosed in the figure legend [50](#page=50) [51](#page=51). > **Note:** The key distinction between acceptable and unacceptable manipulation lies in whether the alterations distort the original data or introduce false information. Transparency and disclosure are paramount [50](#page=50) [51](#page=51). --- # Guidelines and prevention of scientific image manipulation This section outlines journal guidelines for acceptable scientific image manipulation and details preventive measures and checks implemented by publishers to ensure data integrity. ### 3.1 Publisher scrutiny and prevention measures All digital images submitted for publication are scrutinized by the production department for any indications of improper manipulation. If questions arise, the production department refers them to the editors, who then request the original data from authors for comparison with the prepared figures. Failure to produce original data can lead to the revocation of manuscript acceptance. Deliberate misrepresentation of data will result in acceptance revocation and reporting to the author's institution or funding agency [52](#page=52). > **Tip:** Publishers employ image checking systems, such as those used by JCB, JEM, and JGP (Rockefeller University Press), to detect image manipulation [52](#page=52) [54](#page=54). ### 3.2 General guidelines for image manipulation #### 3.2.1 Acceptable manipulations * Adjustments of brightness, contrast, or color balance are acceptable if they are applied to the *entire* image and do not obscure or eliminate any information present in the original [53](#page=53). * Cropping an image is generally considered acceptable. However, cropped gels in a paper must retain important bands and cropped blots should retain at least six band widths above and below the band of interest [56](#page=56). * Simple adjustments to the entire image are usually acceptable . * Manipulation of images should only be performed on a copy of the unprocessed image . #### 3.2.2 Unacceptable manipulations * No specific feature within an image may be enhanced, obscured, moved, removed, or introduced [53](#page=53). * The use of touch-up tools, such as cloning and healing tools in software like Photoshop, or any feature that deliberately obscures manipulation, is to be avoided [55](#page=55). * Contrast should not be adjusted to the point where data disappear [55](#page=55). * Threshold manipulation, expansion or contraction of single ranges, and altering high signals should be avoided [57](#page=57). * Manipulations performed on one area of an image but not on other areas are questionable . * Cloning or copying objects from other parts of the same image or from a different image is very questionable . * Use of software filters to improve quality is not recommended for biological images . * Avoid the use of lossy compression . #### 3.2.3 Disclosure requirements * Nonlinear adjustments, such as changes to gamma settings, must be disclosed. Pseudo-coloring and nonlinear adjustments are only allowed if unavoidable and must be disclosed [53](#page=53) [57](#page=57). * The grouping of images from different parts of the same gel, or from different gels, fields, or exposures, must be made explicit by the figure's arrangement (e.g., using dividing lines) and in the figure legend [53](#page=53). * If juxtaposing images is essential, the borders should be clearly demarcated in the figure and described in the legend [54](#page=54). * When submitting revised final figures, authors may be asked to submit original, unprocessed images [55](#page=55). * Authors should list all image acquisition tools and image software used [54](#page=54). * Key image-acquisition settings and processing manipulations should be documented in the Supplementary Information [54](#page=54). * Adjustments of individual color channels on "merged" images should be noted in the figure legend [57](#page=57). ### 3.3 Guidelines for gels and blots * Vertically sliced gels that juxtapose lanes not contiguous in the experiment must have a clear separation or a black line delineating the boundary [56](#page=56). * Cropped gels in the paper must retain important bands [56](#page=56). * Cropped blots in the body of the paper should retain at least six band widths above and below the band [56](#page=56). * High-contrast gels and blots are discouraged. Multiple exposures should be presented in supplementary information if high contrast is unavoidable [56](#page=56). * Immunoblots should be surrounded by a black line to indicate borders if the background is faint [56](#page=56). ### 3.4 Quantitative imaging and data integrity * For quantitative comparison, appropriate reagents, controls, and imaging methods with linear signal ranges should be used [57](#page=57). * Adjustments should be applied to the entire image [57](#page=57). * Intensity measurements should be performed on uniformly processed image data, and the data should be calibrated to a known standard . ### 3.5 Image checking systems and software Publishers like Rockefeller University Press utilize image checking systems to verify data integrity. Authors are required to list all image acquisition tools and image software used [52](#page=52) [54](#page=54). ### 3.6 Understanding pixel images Scientific images are fundamentally data that can be compromised by inappropriate manipulations [66](#page=66). #### 3.6.1 Pixel images and resolution * Pixel images are composed of a grid of pixels, where each pixel has a specific position (x, y) and intensity value. The total number of pixels ($x \times y$) determines the resolution of the image [68](#page=68). * Different pixel counts and pixel sizes affect image clarity and resolution. A larger pixel count with the same pixel size leads to better resolution [64](#page=64). #### 3.6.2 Data format and encoding * Pixel intensity values range from 0 (no light/signal) to a maximum value, often 255 for 8-bit images [74](#page=74) [75](#page=75) [76](#page=76). * Bit depth determines the number of shades of grey or colors an image can represent. For example, 8 bits allow for 256 tones, while 16 bits allow for 65,536 tones [77](#page=77) [78](#page=78). #### 3.6.3 Dynamic range and Look-Up Tables (LUTs) * Dynamic range refers to the range of intensity levels (pixel values) between the lowest and highest detectable values. It can be adjusted during acquisition (sensitivity/gain) or post-processing (brightness/contrast) [81](#page=81) [82](#page=82). * Look-Up Tables (LUTs) are used to match intensity levels to specific displayed information, affecting how image data is visualized [85](#page=85) [86](#page=86) [88](#page=88) [89](#page=89) [90](#page=90) [91](#page=91) [92](#page=92) [93](#page=93). #### 3.6.4 Saturation Saturation refers to the intensity levels of pixels, with "under-saturated" and "over-saturated" indicating extremes of the intensity range [94](#page=94). ### 3.7 Image file formats * Different image file formats (e.g., TIFF, JPEG) have varying characteristics and levels of compression [100](#page=100) . * JPEG is a lossy compression format that can degrade image quality with repeated saving or manipulation. TIFF is a lossless format . * JPEG compression involves techniques like the Discrete Cosine Transform (DCT) . ### 3.8 Non-linear adjustments Non-linear adjustments, such as gamma corrections, alter the relationship between pixel values and their displayed intensity, impacting the image's appearance. These should be disclosed when used [53](#page=53) [57](#page=57). ### 3.9 Summary of rules for scientific images 1. Scientific images are data that can be compromised by inappropriate manipulations . 2. Manipulation of images should only be performed on a copy of the unprocessed image . 3. Simple adjustments to the entire image are (usually) acceptable . 4. Cropping an image is (usually) acceptable . 5. Images that will be compared to one another should be acquired under identical conditions, and any post-acquisition image processing should also be identical . 6. Manipulations that are performed on one area of an image but not on other areas are questionable . 7. Use of software filters to improve quality is not recommended for biological images . 8. Cloning or copying objects from other parts of the same image or from a different image is very questionable . 9. Intensity measurements should be performed on uniformly processed image data, and the data should be calibrated to a known standard . 10. Avoid the use of lossy compression . 11. Magnification and resolution are important . 12. Be careful when changing the size (in Pixels) of a digital image . --- ## Common mistakes to avoid - Review all topics thoroughly before exams - Pay attention to formulas and key definitions - Practice with examples provided in each section - Don't memorize without understanding the underlying concepts

| Term | Definition | |------|------------| | Data Representation | The way in which data is structured, organized, and presented, which can influence its interpretation and may be subject to changes or manipulations. | | Research Misconduct | Defined as fabrication, falsification, or plagiarism in proposing, performing, or reviewing research results, leading to compromised scientific integrity. | | Fabrication | The act of making up data or results and then recording or reporting them as if they were genuine, which is a form of research misconduct. | | Falsification | The manipulation of research materials, equipment, or processes, or the alteration or omission of results, such that the research is not accurately represented in the record. | | Plagiarism | The appropriation of another person's ideas, processes, results, or words without giving proper credit, which includes forms like citation amnesia and bibliographic negligence. | | Digital Images | Visual data captured or created using digital technology, which are susceptible to easy manipulation due to their digital nature. | | Image Manipulation | The alteration of digital images, which can range from simple adjustments to complex modifications, and can be used both legitimately and illegitimately in scientific contexts. | | Pixel | The smallest addressable element in a raster image or the smallest controllable element of a picture on a screen; a picture element. | | Vector Graphics | Images that are composed of mathematical equations that define points, lines, and curves, allowing for infinite scalability without loss of quality. | | Raster Graphics | Images that are composed of a grid of pixels, where each pixel has a specific color and position, and scaling can lead to a loss of quality. | | Charge-Coupled Device (CCD) | An electronic sensor used in digital cameras and imagers to capture light and convert it into an electrical signal, forming the basis of pixel data. | | Resolution | The level of detail in an image, determined by the number of pixels per unit of area; higher resolution means more pixels and thus more detail. | | Bit Depth | The number of bits used to represent the color of a single pixel in a bitmap image or the number of shades of gray in a grayscale image, determining the range of tones. | | Dynamic Range | The ratio between the maximum and minimum signal values that a system can detect or display, indicating the range of light intensities that can be captured in an image. | | Look-Up Table (LUT) | A data table used to transform input values into output values, often used in image processing to adjust display characteristics like color or brightness. | | Saturation | The intensity or purity of a color in an image; high saturation means a vivid color, while low saturation means a duller color. | | JPEG | A common file format for digital images that uses lossy compression to reduce file size, often by discarding some image information. | | TIFF | A flexible image file format that supports lossless compression and is often used for high-quality images in professional contexts, preserving more image data than JPEG. | | Convolution Kernel | A small matrix used in image processing to apply filters and effects to an image by performing a convolution operation with the image pixels. | | Cloning | A digital image editing technique where pixels from one area of an image are copied and pasted onto another area, often used to conceal defects or create illusions. | | Lossy Compression | A data compression method that permanently discards some data to achieve smaller file sizes, which can lead to a loss of image quality. | | Nonlinear Adjustments | Image processing operations that do not scale linearly with the input values, such as gamma correction, which can alter the perceived brightness and contrast of an image. | | Gamma Correction | A nonlinear adjustment that modifies the luminance of an image by altering the gamma curve, affecting the overall brightness and contrast. | | Pseudocoloring | The assignment of artificial colors to grayscale images or data to enhance visualization and interpretation, especially for highlighting specific features or ranges. | | Figure Legend | A descriptive text accompanying a figure or image that explains its content, methods, and any relevant details, essential for proper interpretation. | | Image Acquisition Tools | Software and hardware used to capture or generate digital images, including cameras, microscopes, and scanners. | | Image Software | Applications used to process, edit, and manipulate digital images, such as Photoshop or GIMP. |

# La démarche expérimentale et ses fondements La démarche expérimentale constitue le pilier de la science et de la recherche, s'appuyant sur des outils fondamentaux tels que l'observation et la mesure pour vérifier des hypothèses et construire le savoir. ### 1.1 Introduction à la science et à la recherche La science est définie comme un ensemble organisé de connaissances construit collectivement. La recherche, quant à elle, est le processus par lequel un chercheur élabore et enrichit ce savoir [1](#page=1). ### 1.2 Les outils de la démarche scientifique La démarche scientifique repose sur plusieurs outils essentiels : * **L'observation**: Elle consiste à noter et rapporter de manière systématique les événements [1](#page=1). * **La mesure**: Il s'agit d'attribuer une valeur numérique à des objets, des événements ou des phénomènes étudiés [1](#page=1). * **L'expérimentation**: C'est un procédé spécifique utilisé pour vérifier une hypothèse, où l'observation joue un rôle intrinsèque [1](#page=1). ### 1.3 Claude Bernard et l'approche expérimentale Claude Bernard, grâce à sa pratique de médecin confronté à la complexité de l'être humain, a contribué de manière significative à l'émergence d'une approche expérimentale rigoureuse en science. Selon lui, le raisonnement scientifique se construit en trois étapes indispensables [1](#page=1): 1. L'observation d'un fait particulier [1](#page=1). 2. La formulation d'une hypothèse pour expliquer ce fait [1](#page=1). 3. La mise en place d'une expérience afin de confirmer ou d'infirmer cette hypothèse [1](#page=1). > **Tip:** L'approche de Claude Bernard met en évidence l'interaction dynamique entre l'observation du réel et la construction théorique, validée ensuite par l'expérimentation. ### 1.4 Expérimenter : provoquer et observer Expérimenter peut être résumé par l'action de provoquer des phénomènes dans le but de les observer ensuite. Cette double action est au cœur de la méthode expérimentale [1](#page=1). ### 1.5 Les phases expérimentales La démarche expérimentale se décompose en plusieurs phases distinctes [1](#page=1). --- # Les hypothèses en recherche expérimentale Les hypothèses constituent des énoncés relationnels supposés entre des faits, servant de fondement à la démarche scientifique expérimentale [2](#page=2). ### 2.1 L'origine des hypothèses Les hypothèses peuvent émerger de deux types de raisonnement [2](#page=2): * **Raisonnement inductif:** Il s'agit de la formulation d'hypothèses générales à partir d'observations spécifiques [2](#page=2). * **Raisonnement déductif:** Inversement, ce processus consiste à formuler des observations spécifiques à partir d'hypothèses générales [2](#page=2). La **démarche hypothético-déductive** combine ces deux approches en formulant des prédictions précises à partir d'une hypothèse théorique afin de la confirmer ou de l'infirmer. Dans le contexte de la recherche expérimentale, les faits de "cause" précèdent logiquement les faits de "conséquences". Il est important de noter que toutes les hypothèses ne peuvent pas être directement mises à l'épreuve des faits [2](#page=2). ### 2.2 Critères d'acceptabilité des hypothèses expérimentales Pour être jugée pertinente et utilisable dans un cadre expérimental, une hypothèse doit satisfaire plusieurs critères essentiels [3](#page=3): #### 2.2.1 L'opérationnalité Ce critère assure le caractère intersubjectif de la recherche expérimentale et améliore la précision de la communication des intentions du chercheur. Une hypothèse est opérationnelle lorsqu'elle est suffisamment concrète et précise pour qu'un tiers puisse la tester, fournissant ainsi suffisamment d'informations pour son évaluation [3](#page=3). > **Tip:** L'opérationnalité est renforcée chaque fois que le chercheur communique les décisions qu'il a prises lors de l'élaboration de son plan d'action [3](#page=3). #### 2.2.2 La vérifiabilité L'hypothèse doit être réellement testable, c'est-à-dire qu'il doit être possible de la soumettre à une épreuve empirique. Il est crucial de comprendre que si une hypothèse est opérationnelle, elle n'est pas nécessairement vérifiable [3](#page=3). #### 2.2.3 La vraisemblance Le chercheur doit procéder à un examen raisonné de l'hypothèse proposée. La mise à l'épreuve ne devrait être entreprise que si des éléments portent à considérer l'hypothèse comme plausible. Une hypothèse vraisemblable s'articule avec les connaissances actuelles [3](#page=3). #### 2.2.4 La cohérence externe Une bonne cohérence externe implique que le chercheur établit un lien avec l'ensemble des connaissances disponibles. Il ne s'agit pas uniquement de suppositions d'accord avec des recherches connues, mais plutôt d'une articulation sur celles-ci; il est possible d'avoir une bonne cohérence externe même si l'hypothèse contredit des résultats antérieurs [3](#page=3). #### 2.2.5 La productivité épistémique Ce critère évalue la valeur ajoutée potentielle de l'hypothèse à la science. La vérification de l'hypothèse doit permettre un véritable accroissement des connaissances dans le domaine étudié. Une hypothèse peut être opérationnelle, vérifiable et vraisemblable, mais sa valeur scientifique dépend de son potentiel de créativité et de génération de savoir nouveau [3](#page=3). #### 2.2.6 L'acceptabilité Lorsqu'une hypothèse est énoncée, elle doit être conforme aux exigences légales, déontologiques (en lien avec les professions) et éthiques (de manière générale) [3](#page=3). --- # L'élaboration et la structure des plans expérimentaux L'élaboration d'un plan expérimental rigoureux est une étape fondamentale pour assurer la validité des conclusions scientifiques, impliquant la définition précise des variables, la sélection adéquate des participants et la structuration des groupes de recherche. ### 3.1 Définition des variables La première étape de l'élaboration d'un plan expérimental consiste à définir clairement les variables qui seront étudiées [4](#page=4). #### 3.1.1 Variable indépendante (VI) La variable indépendante est celle que le chercheur manipule intentionnellement afin de provoquer l'apparition de phénomènes. Les différentes valeurs qu'une VI peut prendre sont appelées ses "niveaux". Une VI doit comporter au minimum deux niveaux. On distingue deux types de VI [4](#page=4): * **Invoquées:** Caractéristiques intrinsèques qui ne peuvent pas être manipulées par le chercheur, comme l'âge, le sexe ou le quotient intellectuel (QI) [4](#page=4). * **Provoquées:** Caractéristiques qui peuvent être activement manipulées par le chercheur [4](#page=4). #### 3.1.2 Variable dépendante (VD) La variable dépendante est l'objet de l'observation et n'est pas directement manipulée par le chercheur. Ses variations sont censées dépendre de la manipulation des niveaux de la VI [4](#page=4). #### 3.1.3 Variables parasites (VP) Les variables parasites sont celles qui exercent un effet indésirable sur la VD, soit directement, soit en influençant la VI. Il est crucial d'essayer de contrôler toutes les variables parasites pour garantir la validité interne de l'expérience [4](#page=4). * **Contrôle statistique:** Utilisation de procédures quantitatives pour prendre en compte les variations introduites par les VP [4](#page=4). ### 3.2 Sélection des sujets La manière dont les participants sont sélectionnés peut grandement influencer la généralisabilité des résultats. Plusieurs méthodes d'échantillonnage existent [4](#page=4): #### 3.2.1 Aléatoire simple Cette méthode vise une représentativité garantie et offre généralement une bonne validité externe. Elle implique un tirage au sort où tous les sujets ont une chance égale d'être sélectionnés, sans aucun critère de sélection spécifique. Sa mise en place est cependant très difficile car elle exige la disponibilité de tous les sujets potentiels [4](#page=4). #### 3.2.2 Occasionnel L'échantillonnage occasionnel se base sur la disponibilité des sujets. Bien que fréquent en sciences humaines, il est critiquable en termes de représentativité [4](#page=4). #### 3.2.3 Empirique Cette méthode, qui présente une très faible probabilité d'exister, repose sur un groupe de sujets possédant déjà toutes les caractéristiques recherchées pour l'étude [4](#page=4). #### 3.2.4 Stratifié Cette approche consiste à établir une répartition des sujets dans des classes basées sur un critère prédéfini. L'échantillonnage aléatoire stratifié vise à obtenir des groupes équivalents, mais il nécessite un accès à l'ensemble de la population [4](#page=4) [5](#page=5). * **Groupes appariés:** Identification d'une VP redoutée, mesure de celle-ci chez chaque sujet, puis association des sujets dans les groupes en fonction de leurs valeurs sur cette VP [5](#page=5). * **Groupes parallèles:** Figer la VP redoutée et sélectionner plusieurs valeurs de cette variable pour constituer les groupes [5](#page=5). #### 3.2.5 Groupes contrastés (non équivalents) Cette méthode implique : 1. Déterminer une variable dont on souhaite étudier l'effet (VI) [5](#page=5). 2. "Mesurer" cette variable sur les sujets [5](#page=5). 3. Composer les groupes en caractérisant des valeurs volontairement différentes de cette VI [5](#page=5). > **Tip:** La variable utilisée pour créer le contraste n'est pas une variable parasite, mais plutôt une VI dont on cherche à étudier l'effet [5](#page=5). ### 3.3 Constitution des groupes La manière dont les groupes sont constitués est essentielle pour comparer les effets des différentes conditions expérimentales. #### 3.3.1 Groupes équivalents L'objectif est de former des groupes dont les caractéristiques sont aussi similaires que possible avant l'application du traitement expérimental. Ceci est idéalement obtenu par un échantillonnage aléatoire, mais des méthodes comme l'appariement ou la stratification peuvent aussi être utilisées [5](#page=5). #### 3.3.2 Groupes non équivalents Dans ce cas, les groupes présentent des différences préexistantes qui ne sont pas contrôlées par l'expérimentateur. Les groupes contrastés sont un exemple de groupes non équivalents où les différences sont intentionnellement créées sur la VI [5](#page=5). ### 3.4 La validité de l'expérimentation La validité d'une expérience fait référence à sa capacité à tester fidèlement ses propres hypothèses [5](#page=5). #### 3.4.1 Validité interne La validité interne est assurée lorsque seules les variables indépendantes manipulées par le chercheur sont responsables des variations observées dans les variables dépendantes. Une menace à la validité interne est la "mortalité" dans l'expérience, c'est-à-dire la perte de sujets au cours de l'étude [5](#page=5). #### 3.4.2 Validité externe La validité externe est atteinte si les résultats d'une expérience peuvent être généralisés à l'ensemble des individus potentiellement soumis aux mêmes influences de la (ou des) VI. Une bonne validité externe permet de généraliser les résultats à un univers d'individus et à un univers de situations spécifiquement décrits [5](#page=5). #### 3.4.3 Validité écologique La validité écologique est satisfaite lorsque les résultats obtenus sont généralisables et prédictifs des comportements observables dans des situations du monde réel. Si les résultats ne sont pas prédictifs des comportements réels, l'expérience a une mauvaise validité écologique [5](#page=5). #### 3.4.4 Paradoxe de l'observateur/de l'expérimentateur Dès qu'un dispositif d'observation est introduit dans une situation réelle, il la déforme, rendant son état initial inobservable [5](#page=5). ### 3.5 Anticiper les problèmes : variables parasites et biais Les variables parasites et les biais constituent des menaces importantes pour la validité interne et externe des expériences [5](#page=5). #### 3.5.1 Des biais Un biais est une action non contrôlée d'une variable parasite sur la variable dépendante, compromettant ainsi les validités interne et externe. Si un biais est présent, il y a nécessairement une variable parasite agissant sur la VD, mais l'inverse n'est pas toujours vrai. Si une variable autre que la VI agit sur la VD, soit elle est ignorée (résultats biaisés), soit elle est connue mais non contrôlée, introduisant une suspicion de biais. Lorsque le contrôle total n'est pas possible, une démarche de réflexion critique est nécessaire [5](#page=5). > **Tip:** Les biais peuvent être initiés par le sujet, l'expérimentateur, ou le dispositif expérimental lui-même [6](#page=6). #### 3.5.2 L'effet Hawthorne Ce biais survient lorsque la simple conscience de faire partie d'une expérience amène les sujets à modifier leur comportement. Pour y remédier, on peut choisir de ne pas informer le sujet avant le test (en tenant compte de la déontologie) ou de demander son accord après pour utiliser les données. Il est conseillé de ne pas tout révéler avant le test [6](#page=6). #### 3.5.3 L'effet placebo L'effet placebo se manifeste par une amélioration de l'état du sujet due à l'administration d'un produit sans effet thérapeutique réel. Pour le contrôler, on utilise généralement deux groupes: un groupe expérimental et un groupe témoin. L'expérimentation en "double aveugle" est une méthode efficace pour minimiser cet effet [6](#page=6). #### 3.5.4 L'effet John Henry Les sujets, conscients de leur statut expérimental, peuvent sur-performer dans le but de prouver leurs capacités. Pour atténuer cet effet, on peut ignorer les sujets ou les prévenir, tout en évitant de créer une hiérarchie [6](#page=6). #### 3.5.5 L'effet Rosenthal (ou effet Pygmalion) Cet effet décrit la tendance de l'expérimentateur à façonner involontairement le comportement des sujets en fonction de ses attentes, basées sur des connaissances préalables sur eux. Des stratégies pour minimiser cet effet incluent l'intervention d'une personne extérieure pour administrer le test (bien que potentiellement coûteuse), la minimisation des contacts entre l'expérimentateur et les sujets, et un examen de conscience rigoureux de l'expérimentateur pour assurer son impartialité [6](#page=6). #### 3.5.6 L'effet de Halo L'évaluateur est influencé dans ses notations par des informations externes aux performances mesurées, telles que la personnalité ou l'écriture du sujet. Pour contrer cet effet, il est possible de filtrer les informations accessibles à l'évaluateur, d'utiliser plusieurs évaluateurs, ou de procéder à un examen de conscience de leur part [6](#page=6). #### 3.5.7 L'effet de mortalité expérimentale Ce biais se manifeste lors d'études longitudinales par la perte de sujets au cours de l'expérience pour diverses raisons. Pour maintenir la motivation des sujets, des récompenses peuvent être utilisées, mais cela soulève des questions éthiques et peut introduire d'autres biais, tout en nuisant à la validité écologique [6](#page=6). #### 3.5.8 L'effet animateur Des données externes ou les interactions avec l'expérimentateur peuvent influencer les comportements des sujets. Il est recommandé d'adopter une attitude neutre et, en présence de plusieurs groupes, de permuter les expérimentateurs entre les groupes [7](#page=7). #### 3.5.9 L'effet de l'expérience personnelle Les influences externes au plan expérimental peuvent s'ajouter aux effets de la ou des VI lorsque le sujet n'est pas dans le contexte expérimental. L'isolement en quarantaine est une solution radicale pour soustraire le sujet aux influences externes, mais d'autres approches impliquent une réflexion sur les sources d'informations externes et leur contrôle quantitatif [7](#page=7). #### 3.5.10 L'effet de maturation Des modifications internes au sujet (aptitudes, humeur, évolution cognitive) peuvent affecter les résultats. Pour atténuer cet effet, on peut diminuer la durée entre les observations ou évaluer une éventuelle progression des sujets [7](#page=7). #### 3.5.11 L'effet de nouveauté Le matériel ou la situation expérimentale peuvent susciter des réactions chez les sujets, positives (engouement) ou négatives (résistance, aversion). Présenter le nouveau matériel aux sujets avant l'expérience peut aider à mitiger cet effet [7](#page=7). #### 3.5.12 L'effet pré-test Dans les plans expérimentaux incluant une observation avant et après le traitement, l'observation pré-expérimentale peut influencer la VD. Cela peut se manifester par un effet d'apprentissage (utilisation d'épreuves parallèles) ou un effet de contamination, où le sujet comprend les intentions du chercheur. Pour éviter cela, on peut proscrire les plans expérimentaux avec pré-test ou utiliser des épreuves parallèles aux différentes phases [7](#page=7). #### 3.5.13 L'effet de régression statistique Cet effet se caractérise par une diminution du caractère contrasté des groupes après la sélection, les notes ayant tendance à se rapprocher de la moyenne globale du groupe. Pour y pallier, il est recommandé de mesurer le trait visé un grand nombre de fois et de calculer la moyenne [7](#page=7). ### 3.6 Organisation du plan d'action L'organisation d'une expérience repose sur la définition de plans expérimentaux précis [7](#page=7). #### 3.6.1 Les plans expérimentaux Un plan expérimental regroupe toutes les informations relatives à l'organisation d'une expérience. Il découle des décisions prises par le chercheur concernant le nombre de variables, leurs relations et le nombre de niveaux pour chaque variable [7](#page=7). #### 3.6.2 Les plans expérimentaux classiques Un plan expérimental est une "structure d'action et d'observation". Cette structure peut caractériser des actions variées, d'où la nécessité d'une codification. Dans les plans les plus simples, une seule VI est manipulée, souvent en comparant l'application d'un traitement à son absence. L'ensemble du plan peut être résumé par une formule simple [7](#page=7). #### 3.6.3 Codification et syntaxe Une codification standard est utilisée pour représenter les éléments d'un plan expérimental [8](#page=8): * $O$ = Observation [8](#page=8). * $X$ = Traitement expérimental [8](#page=8). * $R$ = Échantillonnage aléatoire des sujets [8](#page=8). * $R'$ = Aléatorisation des groupes [8](#page=8). * $=$ = Séparation de groupes équivalents [8](#page=8). * $.......$ = Séparation des strates d'un échantillon stratifié [8](#page=8). La syntaxe suit une logique de lecture de gauche à droite pour représenter l'évolution temporelle. La superposition de lignes indique la présence de différents groupes [8](#page=8). #### 3.6.4 Plan à observations pré et post expérimentales Ce plan implique une observation initiale ($O$), suivie d'un traitement expérimental ($X$), puis d'une observation finale ($O$) [8](#page=8). $$ O_1 \quad X \quad O_2 $$ * **Avantage:** Permet une comparaison avant et après, et aide à réduire les variations inter-individuelles [8](#page=8). * **Inconvénient:** Sensible aux effets pré-test, aux effets liés au temps, à l'effet Hawthorne et à la mortalité expérimentale [8](#page=8). > **Tip:** Transformer une double liste de données appariées en une simple liste d'écarts permet de quantifier les "GAINS" [8](#page=8). #### 3.6.5 Plan à observations pré et post expérimentales avec groupe contrôle Ce plan ajoute un groupe contrôle au schéma précédent, permettant de comparer les changements observés dans le groupe expérimental à ceux du groupe contrôle [9](#page=9). $$ \begin{array}{ccc} R & O_1 & X & O_2 \\ R & O_3 & \quad & O_4 \end{array} $$ * **Avantages:** Permet de statuer sur l'influence réelle du traitement et de contrôler l'effet pré-test par une exposition équivalente des deux groupes au pré-test [9](#page=9). * **Limitation:** Ne permet pas de déterminer l'action éventuelle du pré-test sur les observations réalisées, mais assure que si l'effet pré-test existe, il se manifeste de manière similaire dans les deux groupes [9](#page=9). #### 3.6.6 Plans à groupes parallèles Ces plans cherchent à échapper à l'effet pré-test [9](#page=9). #### 3.6.7 Le plan avant-après sur deux groupes indépendants Ce plan est souvent considéré comme assez invasif et ne peut donc pas être appliqué deux fois aux mêmes sujets [9](#page=9). #### 3.6.8 Plan à pré-tests multiples Ce type de plan permet de contrôler les effets liés au temps. Si la différence entre deux observations consécutives ($O_{n-1}$ et $O_n$) est similaire à celles observées entre d'autres pré-tests, cela indique la présence d'effets temporels [9](#page=9). #### 3.6.9 Dispositif séquentiel : élargissement du plan "avant-après" Ces dispositifs visent à confronter les variations expérimentales aux variations naturelles [9](#page=9). #### 3.6.10 Dispositif de Solomon Ce dispositif combine plusieurs plans pour mesurer l'ampleur de l'effet pré-test et distinguer ses effets de ceux du traitement [9](#page=9). $$ \begin{array}{cccccc} R & O_1 & X & O_2 \\ R & O_3 & \quad & O_4 \\ R & \quad & X & O_5 \\ R & \quad & \quad & O_6 \end{array} $$ * **Avantages:** Permet de mesurer l'ampleur de l'effet pré-test et de différencier les effets du pré-test de ceux d'autres facteurs [9](#page=9). * **Comparaisons clés :** * Comparer $O_2$ et $O_5$ permet de mesurer l'ampleur de l'effet pré-test combiné à $X$ [9](#page=9). * Comparer $O_4$ et $O_6$ permet d'estimer l'effet du pré-test en l'absence de $X$ [9](#page=9). --- # Les relations entre variables dans les plans factoriels Ce chapitre explore les différentes manières dont les variables peuvent interagir dans le cadre des plans factoriels, en se concentrant sur les relations de croisement, de nichage, de confusion et des plans expérimentaux spécifiques comme le carré latin. ### 8.1 Relation de croisement Deux variables indépendantes (VI) sont dites croisées si chaque niveau de l'une des VI est systématiquement associé à tous les niveaux de l'autre VI. Cette structure est fondamentale dans les plans factoriels, où l'on s'intéresse non seulement aux effets de chaque VI prise individuellement, mais aussi à leurs effets conjoints, appelés effets d'interaction [10](#page=10). #### 8.1.1 Effets testables dans une relation de croisement Dans un plan factoriel avec deux VI croisées, il y a trois sources principales de variation à tester : 1. L'effet de la première VI sur la variable dépendante (VD). 2. L'effet de la seconde VI sur la VD. 3. L'effet d'interaction, c'est-à-dire l'effet conjoint des deux VI sur la VD [10](#page=10). Un exemple concret de variation de la VD (score au test de maîtrise de la lecture) pourrait être dû à la VI "méthode" (par exemple, méthode A vs méthode B), à la VI "spatialisation" (par exemple, en ligne vs en colonne), et à l'effet d'interaction entre "méthode" et "spatialisation" [11](#page=11). #### 8.1.2 Représentations des relations de croisement Une représentation graphique ou schématique peut montrer les différentes combinaisons de niveaux des VI, menant à un nombre total de sources de variation discernables. Par exemple, avec deux VI ayant chacune deux niveaux, on obtient $2 \times 2 = 4$ conditions expérimentales, plus les effets principaux de chaque VI, ce qui fait un total de $4 + 2 = 6$ effets à tester si l'on considère aussi la variabilité résiduelle, ou 7 sources de variation discernables dans une analyse complète [10](#page=10) [11](#page=11). ### 8.2 Relation de nichage Deux VI entretiennent une relation de nichage lorsque les niveaux d'une VI (la VI nichée) sont associés à un et un seul niveau de l'autre VI (la VI nichante). Chaque niveau de la VI nichante doit être associé à au moins un niveau de la VI nichée. Il s'agit d'une relation hiérarchique, où une VI est "contenue" dans les niveaux de l'autre [11](#page=11). #### 8.2.1 Exemple de relation de nichage Considérons une étude sur l'effet de deux méthodes de relaxation (sophrologie et training autogène) sur la résistance au stress (VD). Les patients sont suivis par sept thérapeutes: trois pratiquant la sophrologie et quatre le training autogène. Dans ce cas, la VI "thérapeute" est nichée dans la VI "méthode de relaxation". Savoir qu'un patient a eu le thérapeute 1 implique qu'il a reçu la méthode de sophrologie, mais savoir qu'un patient a reçu la sophrologie ne permet pas d'identifier avec certitude le thérapeute spécifique qu'il a eu. Ici, "méthode" est la VI nichante et "thérapeute" est la VI nichée [11](#page=11). #### 8.2.2 Représentations des relations de nichage Les représentations graphiques montrent cette structure hiérarchique. Dans l'exemple précédent, on testerait deux sources de variation: l'effet de la méthode et l'effet du thérapeute (au sein de la méthode) [12](#page=12). ### 8.3 Relation de confusion Une relation de confusion survient lorsque les niveaux de deux VI partitionnent l'ensemble des données de la même manière, rendant impossible de distinguer leurs effets respectifs. Cette situation est à éviter absolument dans la conception d'une expérience, car elle invalide l'interprétation des résultats [12](#page=12). #### 8.3.1 Exemple de relation de confusion Un exemple typique est une recherche en pédagogie portant sur 5 classes. Si chaque classe est enseignée par un seul instituteur et que chaque instituteur n'enseigne qu'une seule classe, alors la VI "classe" est en confusion totale avec la VI "instituteur". Il devient impossible de savoir si une différence observée est due à la classe elle-même ou à l'instituteur qui y enseigne [12](#page=12). #### 8.3.2 Représentations des relations de confusion Visuellement, les niveaux des deux VI se superposent parfaitement, rendant impossible leur désimbrication [12](#page=12). ### 9 Les plans factoriels : cas particuliers #### 9.1 Carré latin : un cas hybride Le carré latin représente une situation intermédiaire entre un plan à deux VI croisées et un plan à trois VI croisées. Il est utilisé pour étudier les effets de deux VI tout en neutralisant les effets d'une troisième variable, souvent considérée comme parasite ou source de variation non désirée. Si la neutralisation de deux variables parasites est nécessaire, on utilise un plan en carré gréco-latin [12](#page=12). #### 9.2 Du plan à 3 VI croisées au plan en carré latin Un plan factoriel impliquant trois VI croisées nécessite une représentation en trois dimensions, souvent conceptualisée comme un parallélépipède ou un cube composé de sous-unités. Par exemple, trois VI ayant chacune trois niveaux mèneraient à $3 \times 3 \times 3 = 27$ combinaisons expérimentales. En revanche, un plan en carré latin avec deux VI ayant trois niveaux (et neutralisant une troisième VI) ne comporterait que $3 \times 3 = 9$ groupes de données. Une condition nécessaire pour le carré latin est que les VI (les deux VI principales et la VI neutralisée) doivent avoir le même nombre de niveaux [13](#page=13). > **Tip:** Les plans en carré latin sont particulièrement utiles lorsque le nombre total de conditions devient prohibitif dans un plan factoriel complet à trois VI croisées, mais que l'on souhaite tout de même évaluer les effets principaux de deux facteurs tout en contrôlant une troisième variable. #### 9.3 Exemple de plan en carré latin Prenons l'exemple de patients hospitalisés pour dépression recevant un nouveau médicament. Les VI principales pourraient être l'"intensité de la dépression" (par exemple, légère, modérée, sévère) et la "dose du médicament prescrite" (par exemple, faible, moyenne, forte). La variable parasite (VP) à neutraliser pourrait être la "personnalité du patient" (par exemple, introverti, extraverti, neutre). Un plan en carré latin permettrait de tester les effets de l'intensité de la dépression et de la dose du médicament, tout en s'assurant que chaque niveau de personnalité est représenté de manière équilibrée à travers les combinaisons de dépression et de dose. Les plans en carré latin neutralisent les effets de la VP en permutant ses niveaux dans le plan expérimental [13](#page=13). --- ## Erreurs courantes à éviter - Révisez tous les sujets en profondeur avant les examens - Portez attention aux formules et définitions clés - Pratiquez avec les exemples fournis dans chaque section - Ne mémorisez pas sans comprendre les concepts sous-jacents

| Term | Definition | |------|------------| | Observation | L'acte de noter et de rapporter systématiquement les événements tels qu'ils se produisent. | | Mesure | L'attribution d'une valeur numérique à des objets, des événements ou des concepts pour quantifier des propriétés. | | Expérimentation | Un procédé scientifique utilisé pour vérifier une hypothèse en provoquant délibérément des phénomènes et en les observant. | | Hypothèse | Une relation supposée ou une explication provisoire entre des faits, servant de base à une investigation scientifique. | | Raisonnement inductif | Le processus de formulation d'hypothèses générales à partir d'observations spécifiques. | | Raisonnement déductif | Le processus de formulation d'observations ou de prédictions spécifiques à partir d'hypothèses générales. | | Opérationnalité | Le critère selon lequel une hypothèse doit être suffisamment concrète et précise pour qu'un tiers puisse la tester empiriquement. | | Vérifiabilité | La capacité d'une hypothèse à être testée et potentiellement confirmée ou infirmée par des données empiriques. | | Vraisemblance | Le critère selon lequel une hypothèse doit être plausible et s'articuler de manière cohérente avec les connaissances scientifiques actuelles. | | Cohérence externe | La mesure dans laquelle une hypothèse s'articule et s'intègre avec l'ensemble des connaissances disponibles dans un domaine de recherche donné. | | Productivité épistémique | La capacité d'une hypothèse à générer de nouvelles connaissances et à accroître de manière significative la compréhension dans un champ d'étude. | | Acceptabilité | Le critère d'une hypothèse qui doit respecter les normes légales, déontologiques et éthiques en vigueur. | | Variable indépendante (VI) | La variable manipulée intentionnellement par le chercheur dans une expérience pour observer son effet sur une autre variable. | | Variable dépendante (VD) | La variable dont les variations sont observées et mesurées par le chercheur, supposée être influencée par la manipulation de la variable indépendante. | | Variables parasites (VP) | Des variables indésirables qui peuvent influencer la variable dépendante, soit directement, soit en interagissant avec la variable indépendante, et qui doivent être contrôlées. | | Validité interne | La mesure dans laquelle les variations observées dans la variable dépendante sont effectivement causées par la (ou les) variable(s) indépendante(s) de l'expérience, et non par des variables parasites. | | Validité externe | La mesure dans laquelle les résultats d'une expérience peuvent être généralisés à d'autres populations, contextes ou situations. | | Validité écologique | La mesure dans laquelle les résultats d'une expérience sont généralisables et prédictifs des comportements observés dans des situations du monde réel. | | Biais | Une action non contrôlée d'une variable autre que la VI sur la VD, menaçant la validité interne et externe de l'étude. | | Effet Hawthorne | La modification du comportement des sujets parce qu'ils savent qu'ils sont observés dans le cadre d'une expérience. | | Effet placebo | L'amélioration de l'état d'un sujet due à la croyance en l'efficacité d'un traitement, même si celui-ci est inactif. | | Effet Rosenthal (ou Pygmalion) | L'influence involontaire de l'expérimentateur sur les comportements des sujets, basée sur ses attentes ou connaissances préalables à leur sujet. | | Plans expérimentaux | L'ensemble des décisions prises par le chercheur concernant l'organisation d'une expérience, incluant le nombre de variables, leurs relations et leurs niveaux. | | Plans factoriels | Des plans expérimentaux qui manipulent deux variables indépendantes ou plus simultanément pour étudier leurs effets principaux et leurs interactions. | | Effet d'interaction | L'effet sur la variable dépendante de la combinaison de deux ou plusieurs variables indépendantes, qui est différent de la somme de leurs effets pris séparément. | | Relation de nichage | Une relation entre variables où les niveaux d'une variable (nichée) sont associés à un seul niveau d'une autre variable (nichante), créant une hiérarchie. | | Relation de confusion | Une relation entre variables où les niveaux de deux variables indépendantes se confondent, rendant impossible de distinguer leurs effets respectifs sur la variable dépendante. | | Carré latin | Un type de plan expérimental utilisé pour étudier les effets de deux variables indépendantes tout en neutralisant l'influence d'une troisième variable considérée comme parasite. |