Misinformation in sociological research can arise by: (1) presenting highly selective literature reviews; (2) investigating unrepresentative or inappropriate data (including “target population creep”); (3) analyzing indicators or variables with notably inadequate measurement; (4) misrepresenting or selectively considering the statistical results; (5) failing to note the possibility of omitted variable bias in statistical models that likely under-control for other relevant factors; (6) confusing direct effects for total effects in statistical models that over-control for intervening variables; (7) “p-hacking” and assuming that statistical significance constitutes substantive significance; and (8) making exaggerated claims about putatively deterministic empirical regularities based on statistical models with limited explanatory power (e.g., a low R-squared).

The careers of individual sociologists are often rewarded for producing findings that support ideological views, but the deceptive usage of these research practices is further promoted by publication bias and by unprofessional manuscript review processes. By eroding the norms of science, as discussed by Robert Merton, the credibility of the sociology of race and ethnicity―and hence its public policy relevance―is undermined.

The session was led by Arthur Sakamoto, HxA member and Professor of Sociology at Texas A&M University. Please note that the presenter sought to make the concepts accessible to all, even those without a formal background in the topics being discussed. All were welcome to attend.