Applying a Mixed Methods Framework to Differential Item Function Analyses

Understanding the reason(s) for Differential Item Functioning (DIF) in the context of measurement is difficult. Although identifying potential DIF items is typically a statistical endeavor, understanding the reasons for DIF (and item repair or replacement) might require investigations that can be informed by qualitative work. Such work is consistent with investigating DIF via expert (or panel) review and cognitive interviews (or think alouds), but there is little guidance for mixing qualitative and statistical techniques. This methodological article articulates a mixed methods research framework that can be used to understand DIF by reviewing how qualitative techniques can be applied after DIF detection so as to yield a better understanding of the measurement scenario. The article does not focus on statistical DIF methods but rather on ideas for how qualitative work can be used to understand DIF once it is detected.