SIPS 2026 DC

In research comparing gender, sexuality, or racial groups, participants identifying outside traditional categories are frequently excluded from analyses or reassigned to more “convenient” categories without their input. These practices are often justified on pragmatic grounds, e.g., concerns about statistical power or appropriateness of drawing inferences from small subgroups. However, such decisions systematically remove marginalized identities from empirical evidence, producing gaps in the literature about some of the most vulnerable populations.

This unconference will present reflections on these analytic decisions and the tensions they raise between research ethics, statistical validity, and practical constraints. Using original data, we will show example strategies we have used to retain participants with low-frequency identities by giving them choice in how their data are used. The session will conclude with a structured discussion with attendees to generate ideas for improving study design, analytic practices, and reporting standards in research involving identity-based group comparisons.

Health Service Psychology trainers promote rigorous research methods and statistical training, yet causal inference is often treated implicitly or deferred to advanced electives. We contend that causal inference should be viewed as part of the common core of doctoral training – a universal intervention that benefits all students, whether it is specific to students’ research focus or simply supports better conceptualization of research problems. We revisit the central role of theory and articulate problems guiding research designs, which are often taught to students. We plan to illustrate how trainers can help students better appreciate when and why psychologists should care about causation, not merely association. We aim to develop discussion of pedagogical strategies for “hand-over-hand” training that moves beyond rote warnings (e.g., “correlation is not causation”) toward teaching of more principled causal thinking. Finally, we consider the downstream effects of including more causal inference in graduate training.