Table 1 Questions to assess how strongly university codes of research conduct (1) defined research integrity, quality and misconduct, (2) required ethics approval, (3) endorsed 9 responsible research practices and (4) discouraged 5 research misbehaviours and questionable research practices. Variable names used in Fig. 1 are in italics
Definitions. Does the code define: | |
'research integrity'? | research integrity |
'research quality'? | research quality |
'research misconduct'? | research misconduct |
Ethics approval. Does the code state that: | |
all research on humans must be approved by an appropriate ethics committee? | NA |
all research on animals must be approved by an appropriate ethics committee? | NA |
Responsible research practices. Does the code state that: | |
study protocols of clinical trials should be publicly registered? | register trial protocol |
study protocols of other study designs should be publicly registered? | register other protocol |
analysis protocols should be publicly registered? | register analysis protocol |
study data should be made publicly available? | open data |
analysis code should be made publicly available? | open code |
findings should be published on open access platforms? | open publishing |
reporting guidelines (e.g. from journals, professional associations, or the EQUATOR network) should be used in reporting research? | reporting guidelines |
conflicts of interest should be declared? | conflicts of interest |
researchers should receive training in research integrity and other areas (e.g. research supervision, data management, peer review, publishing 'negative' findings)? | researcher training |
Research misbehaviours and questionable research practices. Does the code state that: | |
fabricating data should be discouraged? | fabricate data |
selectively deleting or modifying data after performing initial data analysis should be discouraged? | select data |
selectively reporting results (e.g. not publishing a valid 'negative' finding) should be discouraged? | select results |
performing analyses until statistically significant results are obtained (i.e. p should be discouraged? | p-hacking |
hypothesizing after results are known should be discouraged? | harking |