| Abstract: | Although many insights about political conversation have been gleaned from survey data, researchers have yet to engage in serious and intense scrutiny of the measures used to assess political discussion habits. Analyzing three types of data (two three-wave panel studies, social network data from 25 intact groups, and a cognitive interview), this paper examines the reliability of (and meaning ascribed to) self-reported political discussion frequency measures. Results of Study 1 and Study 2 suggest that self-report measures of political discussion exhibit strong internal consistency at a given time point, but substandard reliability within individuals over time and between members of a dyad. Study 3 demonstrates substantial variability in survey respondents’ definitions of political discussion, with more expansive definitions of politics related to higher reports of political discussion frequency. Directions for future research to enhance understanding and improve the reliability of political discussion measures are addressed. |
