Search

Table S2. Pattern matrix with 3 components. (DOCX 18 kb)

Additional file 13. List of 60 misbehaviours with alternative formulations and newly added items from the focus group interviews printed in italics. Newly added and preferred formulations of existing items are printed italic-bold and the former formulation is crossed out.

Additional file 1. Full list of 60 items.

Table S3. Correlations PPQr items with PPQr subscales. (DOCX 15 kb)

Additional file 2. Privacy policy.

Additional file 9. 60 items ranked for aggregate impact per disciplinary field and academic rank. M = mean score per subgroup, SD = standard deviation, N = number of responses. PD & Asis Prof = Postdocs and assistant professors. Asso & Full Prof = Associate and full professors. Biomed = biomedical researchers, Nat = natural sciences researchers, Soc = social sciences researchers and Hum = humanities. Note that we used a missingness-by-design, so the actual number of respondents from that particular subgroup should be multiplied by three. aggregate impact was computed as the product score of frequency (1-3) and impact (1-5) and thus ranged from 1 to 15. The top 5 research misbehaviours on the aggregate level are printed in bold.

Additional file 12. Qualitative rankings of research misbehaviours according to the focus group participants. * Italics denote research misbehaviours that came from the quantitative survey. Focus groups were conducted over 3 academic ranks (see rows, PhD students, postdocs and assistant professors and associate and full professors) and 4 academic ranks (see columns, biomedical sciences, natural sciences, social sciences and humanities). The higher the misbehaviour in the rows; the more detrimental focus group participants indicated this research misbehaviour to be.

Table S5. Pattern matrix with PPQ items and 3 components taken from the reliability sample. (DOCX 16 kb)

Additional file 7. Top 5 most impactful research misbehaviours by disciplinary field and academic rank. M = mean score per subgroup, SD = standard deviation. Impact was scored on a 1 = ‘negligible’ and 5 = ‘major’ impact scale. The higher the mean score, the more impact the misbehaviour was perceived to have on the validity of the study’s findings.

Additional file 10. Adjusted rankings for most detrimental research misbehaviours on the aggregate level by disciplinary field.

Table S4. Regression model to predict emotional exhaustion (MBI). (DOCX 14 kb)

Additional file 4. Focus group topic guide.

Additional file 5. Code tree. Focus group themes are colour coded where pink is social sciences, purple is humanities, blue is natural sciences and green is biomedical sciences. Higher order themes are dark purple and connect to the overall theme, which is light green. Lines signify relations (e.g. is one example of/is part of).

Additional file 6. Top 5 most frequent research misbehaviours by disciplinary field and academic rank. M = mean score per subgroup, SD = standard deviation. The frequency response scale ranged from 1 (‘never’), 2 (‘once or twice’) to 3 (‘three times or more’). A mean value of 2.03 (“Insufficiently mentor or supervise junior co-workers”) thus means that on average, our respondents stated seeing this research misbehaviour once or twice within the last three years.

Additional file 8. Top 5 most detrimental research misbehaviours on the aggregate level by academic rank. M = mean score per subgroup, SD = standard deviation. Detrimental impact on the aggregate level was computed as the product score of frequency (1-3) and impact (1-5) and thus ranged from 1 to 15. The higher the mean score, the higher the perceived aggregate impact.

Additional file 3. Research protocol.

Table S1. Newly developed PPQr items using the facet method. (DOCX 14 kb)