Perspectives of Singaporean biomedical researchers and research support staff on actual and ideal IRB review functions and characteristics

The file “raw_data_repo.xlsx” contains all the data that was collected in the survey. A detailed description of the data is found in the file “raw_data_description.xlsx”. This includes the wording of each question and the contents of each column. Column A contains the participant ID, which is automatically generated by the survey system. Columns B and C contain the screening questions. Columns D to G contain questions on the general perception of IRBs. Columns H to CS contain the individual responses for the 45 IRB-RAT items. Each IRB-RAT item required 2 consecutive responses, the first for what constituted an ideal IRB, and second for what the actual IRB looked like. Within these columns, an underscore (“_”) means that the option of “I don’t know/ I have no experience” is chosen for the actual ratings. This is in contrast to an empty cell, which appears when the question was not answered. Columns CT to DJ contains demographic information of respondents, such as the type of research conducted, the length of time involved in biomedical/clinical research, affiliation with IRBs, and experience with submitting research protocols to an IRB in a foreign jurisdiction. Research institutions which were represented with at least one response are noted in S1 table. The number of responses from each research institution are not provided here, as we have provided assurances to respondents that potentially identifiable information will not be released. Release of individual level responses regarding affiliation with participating institutions could be used to identify some individual participants due to some institutions having a very small number of researchers affiliated with them, particularly when combined with other demographic data provided. Columns DK to DO contains the timestamps and metadata that are automatically generated by the e-survey system. Respondents with a timestamp in the “Branched Out” column (column DM) have either not met the inclusion criteria (the screening questions), or have met the inclusion criteria but have not completed any other question. Comparing our results with the results in other published studies The file "KS_Chenn_comparison_ideal_actual.xlsx" contains the mean, standard deviation, and sample size of the 45 IRB-RAT items in Keith Spiegel et al. (2006), Chenneville et al. (2014) and our study. This allows our ideal and actual scores to be compared with the scores in Keith Spiegel et al. (2006) and Chenneville et al. (2014) by computing Welch’s t-statistic, and the 2-tailed p-value based on the normal distribution. The file “Reeser_comparison_ideal_actual.xlsx” contains the mean of the 45 IRB-RAT items in Reeser et al. (2008), and the mean, standard deviation and sample size of the 45 IRB-RAT items in our study. Because Reeser et al. (2008) did not report the standard deviation, the comparison could not take into account the distribution of the data from Reeser et al. (2008). The Student’s t-score for the null hypothesis that our population mean is equivalent to the scores reported by Reeser et al. (2008) can be computed. The 2-tailed p-value can be calculated based on the normal distribution. The file “difference_comparison.xlsx” contains the mean difference between the ideal and actual scores for Chenneville et al. (2014) and Reeser et al. (2008). It also contains the mean, standard deviation, and sample size of the difference between ideal and actual scores for our study.