Keywords: Mixed methods, qualitative research, media trust, Uganda

This study explores the methodological benefits of mixed-method research designs using media trust polling in Uganda as a case study. Specifically, I use a combination of in-depth interviews followed by a nationally representative survey to illustrate the limitations of prevailing polling questions to measure media trust to the Ugandan context. Together, these complementary datasets provide a more nuanced understanding of how Ugandan respondents are thinking about and responding to polling questions related to trust in journalism and journalists. The project also invites researchers to reconsider the conceptual definitions used in cross-national research, for which more interpretive and qualitative methods may not only be beneficial but necessary.

Conventional understandings of media trust cast the concept as a cognitive and affective outcome in which individuals expect that interactions with news workers, news organizations, and/or content will lead to more positive than negative outcomes (Strömbäck et al., 2020). In turn, media trust is most often measured as an multi-dimensional construct—including single items related to credibility, fairness, accuracy, and bias—using scales developed primarily in WEIRD (Western, educated, industrialized, rich, and democratic; Henrich et al., 2010) contexts. However, Uganda, like many countries on the African continent and across the global periphery more broadly, is characterized by a relatively restrictive political climate, low press freedom, and poor social protections for individuals. Thus, while applying existing measurement tools to assess media trust would certainly return a set of results, such findings may not map onto how individual are actually evaluating trust in journalism and journalists in a situated context.

In contrast, the present study begins with face-to-face, in-depth interviews using a purposive sample of Ugandan adults (N=28). Fielded in May-June 2023, interview questions were designed to understand salient evaluative frameworks used to assess media trust. I then fielded a nationally representative CATI survey (N=607) operationalizing a set of variables developed from the qualitative dataset to explore how these context-specific considerations correlated with existing measures of media trust. Interview transcripts were analyzed thematically using qualitative analysis software Dedoose. Survey data was analyzed using a combination of factor analyses, correlation analyses, and OLS regressions with media trust as both a single- and multi-item outcome variable.

Findings suggest that context-specific considerations—such as whether journalists accept bribes for coverage, for example—were the strongest predictors of media trust. This was the case when media trust was measured as both a single- and multi-item concept and while controlling for theoretically important covariates, including political ideology, media use, and education, among others. Moreover, media trust correlated with distinct covariates when measured as a single- and multi-item construct, underscoring how unsuitable measurement tools may activate unique dispositions, particularly when used cross-nationally. Together, these insights illuminate the specific case of media-trust polling in Uganda; more broadly, however, findings also exemplify the significance, if not critical necessity, of mixed-method research to cross-national public opinion research.