Does the Method Affect the Outcome? How Measures of Partisan Slant of Media Outlets and Affective Polarization Drive Results About Polarization in the United States

There is abundant evidence that partisan media use is related to polarization in the United States. However, there is less agreement on whether the outcomes of this exposure are detrimental to democracies. Different methods used to measure both partisan media use and polarization may have contributed to the variation in findings in this area of research. The present study relied on survey data from the 2020 U.S. presidential election to systematically examine how methodological differences in determining the partisanship of news outlets and polarization can impact outcomes of interest, such as the correlation between outlet-level partisan slant and affective as well as belief polarization. A specification curve analysis of 504 different ways of operationalizing these independent and dependent variables provided evidence that methodological choices considerably affected the strength of the relationship between partisan media use and affective and belief polarization.