Abstract

Background

Scientific dishonesty frequently refers to actions or omissions in connection with research, which leads to false or distorted scientific results or gives misleading information about an individual contribution to research [1-3]. Scientific dishonesty is problematic for a number of reasons. It may directly or indirectly harm vulnerable research participants. It may undermine the general trust in science and scientists, and it may cause harm if future research or therapy attempts to rely on fraudulent results.

Traditionally the norms of science are learned by witnessing exemplary behaviour [4]. Mentors, supervisors and institutions played a significant role in promoting such norms. This has changed significantly as the number of researchers and time pressure has increased, as research has become international and interdisciplinary, as there are tight ties between academia, private industry, and governmental research agencies, and as there is an experienced increase in pressure for publications and achieving grants. Although only the serious cases grab the media headlines, such as the Sudbø case in Norway [5], it remains important to be aware of other types of questionable behavior that threaten the integrity of science [6,7].

How widespread is scientific dishonesty? This is a difficult question with significant methodological challenges [8], and divergent answers can be found within empirical research [9-11]. Not surprisingly there seems to be significant underreporting of dishonesty and misconduct [12,13]. A systematic review and meta-analysis [11] showed that an average of 2% of scientists admitted serious forms of misconduct (fabrication, falsification or modification of data or results) at least once. Up to 34% admitted other questionable research practices, and in surveys asking about the behaviour of colleagues, ‘admission’ rates were 14% for falsification, and up to 72% for other questionable research practices. Medical/pharmacological researchers reported misconduct more frequently than others, although not significantly so, and it is concluded by the author that it is likely that the result is a conservative estimate of the true prevalence of scientific misconduct [11]. In a Nordic context population based calculations indicate that Denmark has 1–2 cases/million inhabitants per year, Finland 1–2 cases/million per year, and Norway 1 case/million per year [14].

It is argued that there is no clear relationship between attending courses in research ethics or mentoring and scientific dishonesty. Training has not been shown to be effective in preventing problematic behaviour, and mentoring both increased and decreased the likelihood of problematic behaviours, depending on the kind of mentoring (research, financial, survival, personal, ethics) [15].

A study of undergraduate and postgraduate pharmacy students revealed widespread deficiencies in student knowledge of, and attitudes towards, plagiarism [16]. Many students did not perceive plagiarism as a serious issue and the use of inappropriate strategies for sourcing and acknowledging material was common. Similar results are found elsewhere [17].

Although medical students consider dishonest behaviour to be wrong, a substantial number of students report that they engage in dishonest behaviour [17,18]. Students’ attitudes appear to be stable throughout the years of the study [19], and their willingness to cheat may be culturally dependent and starts before medical school [20]. The baseline knowledge of responsible conduct of research for biomedical sciences students is shown to be inadequate and inconsistent [21].

There appear to be several motives for scientific dishonesty: the universities’ admission requirements, the system for financing, access to scholarships [3], and the perceived need for personal progress and success.

Scientific dishonesty has been studied in the Nordic countries [22-25], but our knowledge is still limited and of medium or poor quality. One study has been carried out in Sweden, [3] but comparative knowledge is necessary to assess the extension of dishonesty and postgraduate students’ attitudes towards scientific dishonesty.

What are PhD-students experiences with scientific dishonesty, what are their attitudes towards various forms of scientific dishonesty, and what do they know about regulations and policies? These are the key questions of our survey carried out among PhD-students at medical faculties in Norway.

Methods

A two-page questionnaire combining a survey developed at the Department of Medical Ethics in Lund, Sweden [3] with a survey developed by Kalichman was applied [9,26]. Terms such as ‘scientific dishonesty’, ‘plagiarism’, ‘fabrication of data’, and ‘falsification’ were given standard definitions in the introduction. The questionnaire was piloted with a group of students similar to the intended respondents, and a few ambiguous statements were identified and re-written.

The participants in the study were post-graduate students being enrolled in PhD-programs at all medical faculties in Norway, i.e., at the universities in Oslo, Bergen, Trondheim, and Tromsø. The questionnaires were distributed to doctoral students attending basic, obligatory courses in research methodology, philosophy of science, and research ethics at the four medical faculties in Norway in the fall 2010 and in the winter 2011 in Tromsø and Trondheim. There were two courses in Oslo in the fall 2010, one of the courses was held in Norwegian (Oslo 1) and one in English (Oslo 2). The questionnaires were anonymous and participation was voluntary. The questionnaires were distributed and collected before any discussion of scientific misconduct had taken place in the course. Answer categories were Yes/No/Uncertain, and Strongly Disagree/Disagree/Neither Agree nor Disagree/Agree/Strongly Agree.

Descriptive statistics was used to find differences in post graduate students’ attitudes towards and experiences with scientific dishonesty. Categorical variables were compared using the Student t test, Fisher exact test, Mann–Whitney test (p<0.05 was considered significant).

Answers were not traceable to participants. The project is registered at the Norwegian Social Science Data Service, overseeing privacy and data protection according to the Personal Data Act.

Results

The total number of questionnaires distributed and returned at the various universities, together with the participants’ answers to questions about academic background is displayed in Table1. 262 questionnaires were distributed, of which 189 were returned, giving an overall response rate of 72.1%. The doctoral candidates’ voluntary participation and anonymity were emphasised in the covering letter and when handing out the questionnaires.

The response rate varied greatly between the five Norwegian universities—from about 51% (Trondheim) to 98% (Oslo). Most of the respondents had taken their undergraduate studies in Norway (72%), except for the course in Oslo given in English (Oslo 2) where 53% of the respondents had taken their undergraduate studies outside of Norway.

The largest group of respondents were doing clinical research (45%), and 29% were doing basic research. 62% and 11% of the respondents had been a doctoral student less than one year and more than two years respectively. On average, the respondents in Tromsø had been doctoral students longer compared to respondents from the other universities.

59% of the respondents had attended lectures or courses in science ethics as part of their undergraduate studies while 31% had not.

The pilot study was carried out in Oslo including 26 respondents with a response rate of 55,3%.

To the questions about the experience of scientific dishonesty and other unethical behaviour in connection with research during the previous 12months, on average 23% of the respondents answered that they had heard or read about researchers who had cheated (nationally or internationally), 65% had not heard about this, and 12% were uncertain. 29% had heard about fabricated data (see Table2).

Table 2

Answers to questions about scientific dishonesty and other unethical behaviour in connection with research (Those who have answered YES in percent)

Questions	Bergen	Oslo 1	Oslo 2	Tromsø	Trondheim	All Norway	All Sweden
Have you, nationally or internationally, heard about anyone who during the last 12months that has
Fabricated data	21,1	28,3	33,3	36,7	29,3	29,2	29
Falsified data	18,4	23,9	23,3	30	24,4	23,8	31,8
Plagiarised data	13,2	19,6	20	23,3	29,3	21,1	24,2
Plagiarised publications	5,3	17,4	16,1	31,3	29,3	19,7	-
Have you yourself during the last 12months been the object of pressure to
Fabricate data	0	2,1	0	0	0	0,5	0
Falsify data	0	2,1	0	0	0	0,5	5,4
Plagiarise data	0	0	3,2	0	0	0,5	0
Plagiarise publications	0	0	0	0	0	0	-
Have you during the last 12months been exposed to unethical pressure concerning
Ordering of authors	13,2	8,7	12,9	12,5	7,3	10,6	8,5
Design/method	0	2,2	6,5	3,1	2,4	2,7	3,1
Results	0	0	12,9	0	2,4	2,7	0,8
Harassment	0	0	0	3,1	0	0,5	0,8
Have you during the last 12months been affected by any consequences of scientific dishonesty
Ethical	0	6,5	3,2	12,5	7,3	5,9	0
Legal	0	0	3,2	3,1	0	1,1	0
Methodological	0	4,3	0	3,1	7,3	3,2	-
Any other aspect	2,6	4,3	3,2	0	4,9	3,2	0

Open in a separate window

Data from Sweden reproduced from Nilstun 2010.

One respondent had experienced pressure to fabricate and to falsify data, and one had experienced pressure to plagiarize data, while none of the participants had felt pressure to plagiarize publications. Five respondents were uncertain if they had experienced pressure to fabricate data.

None of the respondents reported to have fabricated, falsified or plagiarized data or to have plagiarized publications. Only one respondent was uncertain whether he or she had plagiarized data and two were uncertain whether they had plagiarized publications.

About 11% of the respondents had experienced unethical pressure concerning the order of authors on publications during the last 12months. In addition 7% stated that they didn’t know whether they have experienced such pressure. Five respondents had experienced unethical pressure concerning design/method and five had experienced unethical pressure concerning results, while one had experienced harassment.

About 6% of the respondents had been affected by ethical consequences of scientific dishonesty, and 3% had been affected by methodological consequences of scientific dishonesty during the last 12months.

60% of the respondents were uncertain whether their department had a written policy about scientific conduct. 28% stated that their department had a written policy, and 12% that their department did not. Written policies on funding and on fabrication and falsification of data reached the highest reported awareness amongst the respondents (38%, 31%, and 31%). See Table3.

Table 3

How many PhD-students were uncertain about whether their department had written policies (in percent)

Questions	Bergen	Oslo 1	Oslo 2	Tromsø	Trondheim	All Norway	All Sweden
Does your department have a written policy about
Application for funds	63,9	44,7	48,4	62,5	53,7	54	59,2
Use of funds	63,9	36,2	41,9	65,6	61	52,9	57,7
Changes in design/method	72,2	63,8	61,3	77,4	68,3	68,3	47,3
Changes in results	77,8	59,6	71	77,4	65,9	69,4	43,1
Fabrication of data	75	38,3	48,4	71	61	57,5	44,6
Falsification of data	75	38,3	48,4	71	61	57,5	43,8
Ordering of authors	72,2	38,3	54,8	75	63,4	59,4	50,8
Plagiarism of others	75	46,8	48,4	71,9	53,7	58,3	49,6
Publishing the same twice	77,8	44,7	61,3	75	61	62,6	46,2
Harassment	83,3	53,2	41,9	68,8	65,9	62,6	43,8

Open in a separate window

Data from Sweden reproduced from [3].

The results on attitudes towards scientific misconduct (Table4) show that 10% did not agree with the statement that it was never appropriate to report experimental data that have been created without actually having conducted the experiment, and 4% did not agree that it is never appropriate to alter experimental data to make an experiment look better than it actually was.

38% did not agree with the statement that it is never appropriate to try a variety of different methods of analysis until one is found that yields a result that is statistically significant, and 5% did not agree that it is never appropriate to take credit for the words or writing of someone else.

12% did not agree that it is never appropriate to take credit for data generated by someone else, and 8% did not agree that it is never appropriate to take credit for the ideas generated by someone else.

13% agreed that it is acceptable to selectively omit contradictory results to expedite publication and 10% found it acceptable to falsify or fabricate data to expedite publication, if they were confident of their findings, and 29% agreed that it is more important that data reporting be completely truthful in a publication than in a grant application. 13% agreed that you have an ethical obligation to act if you witness someone committing research misconduct.

79% agreed they would be willing to report that misconduct to a responsible official, if they had witnessed a co-worker or peer committing research misconduct, and 75% agreed that they would be willing to report misconduct to a responsible official if they had witnessed a supervisor or principal investigator committing research misconduct.

46% agreed that all co-authors must equally share in the blame if fabricated data are discovered in a published paper, and 29% agreed that all co-authors must receive the same punishment if fabricated data are discovered in a published paper.

There were some differences between the universities. E.g., only half of the PhD-students in Oslo disagreed that it was never appropriate to try a variety of different methods of analysis until one is found that yields a result that is statistically significant. More PhD-students at the course in English in Oslo (Oslo 2) than the average found it acceptable to selectively omit contradictory results to expedite publication if they were confident of their findings. Fewer PhD-students at Oslo 1 and more in Bergen agreed that all co-authors must equally share in the blame if fabricated data are discovered in a published paper than the average. Fewer of the PhD-students in Trondheim agreed that it is more important that data reporting be completely truthful in a publication than in a grant application than the average. In Tromsø fewer respondents than the average reported that it is never appropriate to try a variety of different methods of analysis until one is found that yields a result that is statistically significant. More PhD-students with undergraduate studies outside Norway answered that they had been exposed to unethical pressure concerning results during the last 12months than those who had studied in Norway (p=0.045, Fischer’s exact test), and that they were uncertain whether they had been exposed to unethical pressure concerning harassment (p=0.02, Fischer’s exact test). The PhD-students with undergraduate studies outside Norway differed also from those that had studied in Norway in that they found it more acceptable to selectively omit contradictory results to expedite publication if they were confident of their findings (p=0.008, Mann–Whitney).

Discussion

The findings show that scientific dishonesty is not unknown to Norwegian doctoral students and that they found some actions acceptable which are considered to be misconduct in the science ethics literature. There was little awareness of relevant policies for scientific conduct, but a high level of willingness to report misconduct.

Conclusion

The survey shows that scientific dishonesty is not unknown to PhD students in Norway. Very few stated that they were exposed to pressure to fabricate, falsify, or plagiarise data, while pressure put upon post graduate students regarding the order of authors was more common. Some forms of scientific misconduct were considered to be acceptable by a substantial minority. e.g., almost two of five respondents found data fishing acceptable. There was little awareness of relevant policies for scientific conduct, but a high level of willingness to report misconduct.

With the intention to build more awareness and improved attitudes towards various forms of scientific dishonesty our suggestions are:

• Increase the quality of teaching, use cases, focus on grey areas, and communicate the norms of good scientific practice repeatedly and in a variety of settings.

• Provide better training for supervisors and ensure that they have the appropriate knowledge and promote attitudes.

• Clarify institutional policies towards scientific dishonesty and communicate them more effectively.

Competing interests

All authors are involved in teaching science ethics to PhD-students.

Authors’ contributions

Søren Holm (SH) and Bjørn Hofmann (BH) elaborated the questionnaire which was revised by all authors. All authors participated in data collection. BH performed data analysis and developed the first draft of the manuscript. All authors contributed significantly to the manuscript and its revision and have approved the final version.

Pre-publication history

The pre-publication history for this paper can be accessed here:

http://www.biomedcentral.com/1472-6939/14/3/prepub

References