Scientific Fraud: Driven by the Culture of a Discipline?
All kinds of research need an ethical approach. From the big questions, such as deciding if embryonic stem cells should be used, to the simple integrity that is expected in data collection and analysis - ethics are omnipresent. These are the things I initially wanted to write about when I thought about research ethics. But as I researched the topic, I noticed something unexpected: when I scanned the internet for cases of scientific fraud, the majority seemed to be in the field of biology. Was this my personal bias or was it a fact?
This got me thinking. I decided to instead write about the different kinds of scientific fraud across the disciplines. My analysis is based on the only research paper I could find that provided quantitative data for various types of frauds across different disciplines: a 1993 American Scientist study titled Ethical Problems in Academic Research, by Judith Swazey and her colleagues. The authors conducted a survey of the prevalence of different types of scientific malpractices observed across 99 of the largest graduate departments in chemistry, civil engineering, microbiology and sociology.
Swazey and her team surveyed 2,000 doctoral students and their respective professors about their observations of different types of fraud they had witnessed among students and faculty. I compiled parts of their data obtained from information given by faculty: the percentage of faculty who acknowledged the prevalence of a particular type of fraud amongst their colleagues (other faculty). Based on this I ranked percentages for departments in each misconduct from 1 to 4. In cases where two departments had equal percentages, I gave both equal ranks. In the table below, the actual percentages from the paper are presented, with ranks that I gave in parenthesis.
Sl.
No.Type of misconduct
Chemistry
Microbiology
Sociology
1.
Falsification or cooking data
4 (3)
10 (1)
8 (2)
4 (3)
2.
Plagiarism
5 (4)
18 (1)
6 (3)
7 (2)
3.
Inappropriate assignment of authorship
23 (4)
44 (1)
28 (3)
33 (2)
4.
Overlooking of flawed data or interpreting data in questionable fashion
18 (4)
20 (3)
24 (2)
27 (1)
5.
Use of university resources for outside consulting or inappropriate purposes
38 (3)
61 (1)
32 (4)
42 (2)
6.
Overlooking data contradictory to one's own previous research
18 (1)
15 (3)
17 (2)
12 (4)
7.
Misuse of research funds
20 (2)
35 (1)
20 (2)
16 (3)
8.
Unauthorized use of privileged information in connection with their own research
11 (3)
12 (2)
14 (1)
4 (4)
9.
Failure to disclose commercial involvement relating to one's own research
5 (3)
12 (1)
8 (2)
2 (4)
10.
Sexual harassment
20 (2)
16 (3)
15 (4)
40 (1)
11.
Discrimination on the basis of race, ethnicity or gender
21 (3)
28 (2)
19 (4)
32 (1)
12.
Use of position to exploit/ manipulate others
45 (4)
53 (2)
47 (3)
57 (1)
Swazey and her colleagues' results indicate that the prevalence of fraud in a department depends strongly on the nature of fraud. Based on this, I looked for possible correlations between the prevalent form of misconduct and the nature of the discipline. From the table, it is obvious that civil engineering leads the other disciplines by a significant margin in instances of plagiarism, inappropriate assignment of authorship, use of university resources for outside consulting, misuse of research funds and failure to disclose commercial involvement relating to one's own research. Sociology is the fourth ranked in both cases of unauthorised use of privileged information in connection with research and failure to disclose commercial involvement relating to one's own research. However, it leads all other disciplines in the three forms of interpersonal misconduct: sexual harassment; distinction based on race, gender, ethnicity; and using power and position to exploit others. Microbiology is consistently ranked second, closely tailing the first, in misconduct related to data falsification, analysis and interpretation.
I believe that each of these trends portray some aspect of the nature of the discipline itself. For example, since civil engineering is an old discipline, it stands to reason that its structure is more hierarchical. This could lead to junior scientists trying to please their seniors by offering them authorship on their papers (although this does not explain away high cases of plagiarism). Since this discipline also has more immediate applications in the real world and strong ties with its industrial counterparts, it has more scope for misconduct resulting from commercial involvement.
Sociology on the other hand, is a discipline where very little commercial involvement seems to be required (though it is finding a lot of use in marketing these days). So it is reasonable to expect it to have the least number of cases of misconduct due to such involvement. But it is indeed surprising to see it having the maximum number of cases related to interpersonal misconduct. Surely the sexual harassment cases cannot be explained by the fact that sociology has a healthy gender ratio, since microbiology too has a similar one. Can this be attributed to the fact that people in this department are more aware of and sensitive to issues related to interpersonal dealings, thus leading to more such cases being reported? That needs to be looked into.
Traditionally, microbiology has been an observational science in which a majority of the results are qualitative and not quantitative. This creates scope for the subjective interpretation of results, and therefore, scope for misconduct related to the same. A majority of biologists also lack proper training in mathematical sciences (though this trend is fast changing). This could also lead to a lack of rigour in data analysis.
Different scientific disciplines come with their own distinctive cultures, norms, values and thinking patterns. They vary in the nature of questions that are asked and the level at which they are answered. While some fields may demand observational skills, others might require more abstract thinking. In general, though we all need certain skills to be well-equipped to address scientific questions, some skills are more valued in one discipline than in another.
I believe that there is a strong positive conditioning for a particular type of thinking and working style in any discipline, which may be correlated with the type of misconduct commonly observed in that discipline. This article however, is quite basic and its results need to be checked for statistical significance. My interpretation of these results is also very subjective. But I do think there is need for more such studies investigating the correlation between the nature of a discipline and the common form of misconduct in that discipline. Any remedial measures to curb scientific misconduct should also pay attention to the background of the discipline and take appropriate steps to improve the training of researchers.
(Edited by Aathira Perinchery)
Comments
I don't think we can
Dear Anna, You are very
Post new comment