Interpreting discordant results in mixed-method research: data triangulation, participant voices, and epistemic issues in health research.

In this article, we discuss unexpected methodological challenges revealed during interim analysis of data from a mixed-methods study with female Mexican immigrants. 'Tertulias' was a randomized controlled trial of an innovative peer support group model for reducing social isolation and depression among women immigrants from Mexico using a Community-Driven, Community-Engaged Research (CD-CEnR) approach. We describe discordant data obtained using different methods and how we went about triangulating and integrating analyses. In the process, we identified cultural, contextual, and methodological factors that influenced the nature and content of the data, and issues of power and epistemic justice in the broader research enterprise. Based on this experience, we offer insights for data integration to improve both scientific rigor and epistemic justice in health disparities research.