![Sample our Humanities journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5939/TOC-Subject-Sample-2021-Humanities.jpg"})
Abstract
This study examines the mediating role of media indignation, a set of negative emotional reactions to media coverage perceived to have partisan bias, between hostile media perception (HMP) and its consequences in behavioral willingness, more specifically, willingness to engage in discursive activities. It is examined together with a cognitive pathway through individuals' inference on opinion climate. Data from a web-based survey of university students (N = 696) on three controversial issues were analyzed. Findings showed that the degree of perceived partisan bias in mainstream media coverage on an issue was positively related to the intensity of media indignation and levels of incongruity between one's own opinion and perceived majority opinion. Media indignation in turn had a significant and positive effect on willingness to engage in discursive activities across all three issues, mediating the effect of HMP on behavioral willingness, whereas self-majority opinion incongruity did not. Theoretical implications of the findings are discussed.
Notes
1. Reported coefficients are estimated by standardized solution. Demographics (gender and year of school), news media uses (print news, TV news, and Web news), ideological extremity, extremity of issue attitude, value-involvement, and news media distrust were controlled.
2. Due to relative large sample size, the Sobel z-test (CitationSobel, 1982) was used for statistical tests of the specific indirect effects. N = 696.
∗p < .05
∗∗p < .01
∗∗∗p < .001.
a. Parenthesized entries are control variable.
∗p < .05
∗∗p < .01
∗∗∗p < .001.
1. Two studies in the extant literature measured media slant perception using perceivers' issue position as a judgmental anchor (CitationEveland & Shah, 2003; CitationHwang, Schmierbach, Paek, Gil de Zuniga, & Shah, 2006).
2. A separate survey from the same subject pool showed a correlation between age and years in school at .85 and more than 80% between 19 and 21 and an additional 16% between 22 and 23 years of age. The subject pool is also homogeneous in racial and ethnic composition. The same survey showed that more than 90% were Caucasians.
3. Although our three-factor model revealed satisfactory model fit, a one-factor model in which all indicators form a single latent factor may also be conceptually justifiable given that positions on all of these issues could fall on a single liberal-conservative continuum. To test this alternative measurement model, we performed a CFA for a one-factor solution with the same correlated error allowance as the adopted model. The results showed that the one-factor model did not produce a satisfactory fit [χ2 (18) = 779.45, p = .00, RMSEA = .25, SRMR = .11, CFI = .76, and TLI = .52].
4. The large chi-square statistic could be due to the large sample size (see CitationMacCallum, Browne, & Sugawara, 1996). Fitting the same model (df = 40) with reduced sample sizes to the level of sufficient statistical power (e.g., N = 300, power = .87) led to significant reductions in χ2 statistic (χ2 (39) = 50.99, p = .04), but no noticeable improvement in other fitness statistics.
5. This involves regressing all of the study variables on the control variables and then using the residuals of the study variables from this predictive model in the subsequent model fitting.
