Researchers sometimes ask consulting statisticians to interpret data in unethical ways during biomedical research projects, according to a study published in the Annals of Internal Medicine.
The researchers asked a random sample of 522 members of the American Statistical Association, all of whom identified as consulting biostatisticians, to participate in an online survey. Nearly 400 statisticians responded to the survey, which asked about the frequency and severity of researchers approaching them with requests for inappropriate analysis or reporting of data during consultations.
The four most common "inappropriate requests" that at least 20 percent of respondents rated as "most severe" involved:
1. Removing or altering some data records to better support the research hypothesis
2. Interpreting the statistical findings on the basis of expectation, rather than actual results
3. Not reporting the presence of key missing data that might bias the results
4. Ignoring violations of assumptions that would change results from positive to negative
These types of requests were most often reported by younger biostatisticians, according to the researchers. The researchers noted the study only outlines the frequency of inappropriate requests, not how the requests were handled or the reasoning behind the requests — which might simply be inadequate knowledge of statistical research methods.
"This survey suggests that researchers frequently make inappropriate requests of their biostatistical consultants regarding the analysis and reporting of their data," the study authors concluded. "Understanding the reasons for these requests and how they are handled requires further study."