The Effects of Responsibility Frames and Stigmatizing Headlines in News on Support for COVID-19 Policies in Korea

ABSTRACT

This study examined the mechanisms through which responsibility frames and stigmatizing headlines influence support for governmental policies to address the pandemic. Based on a factorial design experiment, we examined the effects of 2 responsibility frames (individual vs. societal responsibility) and 4 headline types (non-stigmatized vs. name-stigmatized vs. characteristic-stigmatized vs. both-stigmatized). The results showed that the individual responsibility frame increased individual attribution of responsibility for the cause and spread of COVID-19 whereas reducing societal attribution of responsibility, compared to the societal responsibility frame. The headline that detailed both the stigmatized characteristic and name increased individual attribution of responsibility compared to the non-stigmatized headline. Furthermore, the effects of frames and headline types on policy support were sequentially mediated by attribution of responsibility and emotions. Individual attribution of responsibility led to anger whereas societal attribution of responsibility led to sympathy. Subsequently, anger increased support for punitive policies while sympathy increased support for assistive policies. This study contributes to the literature on news framing of pandemics by integrating cognitive and emotional mechanisms in forming policy attitudes.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplementary data

Supplemental data for this article can be accessed online at https://doi.org/10.1080/15205436.2023.2202657.

Data availability statement

The data that support the findings of this study are available from the corresponding author, C. Lee, upon reasonable request.

Notes

¹ The percentage of participants recruited from the areas mentioned in the news stories were as follows: Daejeon (3.3%), Jeju (0.2%), and Gangwon-do including Taebaek (2.5%).

² Across the eight conditions, there were no significant differences in participants’ gender, χ² (7, 632) = .00, p = 1.00, age, F (7, 632) = 0.04, p = 1.00, education level, F (7, 632) = 0.94, p = .47, political orientation, F (7, 632) = 0.57, p = .78, and COVID-19 infection, F (7, 632) = 1.04, p = .41.

³ A pretest (N = 200) was conducted in November 2020 prior to the main experiment to make sure there were no confounding factors, such as perceived message quality (Zhao et al., 2011), associated with our stimulus construction. One-way ANOVA results indicated that there was no significant difference across the eight conditions in measures including perceived message quality, Daejeon F (7, 192) = 0.75, p = .63, Jeju F (7, 192) = 0.95, p = .47, Taebaek F (7, 192) = 0.84, p = .56; perceived argument strength, Daejeon F (7, 192) = 1.37, p = .22, Jeju F (7, 192) = 1.38, p = .22, Taebaek F (7, 192) = 0.69, p = .68; perceived informativeness, Daejeon F(7, 192) = 0.42, p = .89, Jeju F (7, 192) = 0.66, p = .71, Taebaek F (7, 192) = 0.46, p = .86. Thus, no confounding issue was detected.

⁴ The manipulation items on the responsibility frame were checked both in the pretest and the main experiment. The results were significant as intended in both times. In this manuscript, results from the main experiment were reported.

⁵ Respectively for each stimulus, individual responsibility ratings in the individual versus societal frame conditions were: Daejeon M(SD)_Ind = 4.00(0.71), M(SD)_Soc = 3.58(0.91), t (638) = 5.90, p < .001; Jeju M(SD)_Ind = 3.85(0.84), M(SD)_Soc = 3.29(1.01), t (638) = 7.61, p < .001; Taebaek M(SD)_Ind = 4.06(0.75), M(SD)_Soc = 3.73(0.92), t (638) = 4.93, p < .001.

⁶ Societal responsibility ratings were: Daejeon M(SD)_Ind = 2.58(0.93), M(SD)_Soc = 3.63(1.04), t (638) = −12.11, p < .001; Jeju M(SD)_Ind = 2.28(1.04), M(SD)_Soc = 3.28(1.03), t (638) = −12.23, p < .001; Taebaek M(SD)_Ind = 2.36(1.09), M(SD)_Soc = 3.38(1.01), t (638) = −12.24, p < .001.

⁷ The manipulation check on the headline types was performed in the pretest and, because of its straightforwardness, was omitted in the main experiment. Statistical values reported here were from the pretest.

⁸ Descriptive statistics in the four conditions were as follows: Daejeon M(SD)_{Non-stigmatized} = 2.84(1.20), M(SD)_{Characteristic} = 4.36(0.63), M(SD)_Name = 3.42(1.18), M(SD)_Both = 4.20(0.64), F(3, 196) = 27.52, p < .001; Jeju M(SD)_{Non-stigmatized} = 3.02(1.23), M(SD)_{Characteristic} = 3.89(0.89), M(SD)_Name = 3.16(1.20), M(SD)_Both = 4.10(0.74), F(3, 196) = 14.22, p < .001; Taebaek M(SD)_{Non-stigmatized} = 3.44(1.05), M(SD)_{Characteristic} = 3.96(0.98), M(SD)_Name = 3.20(1.18), M(SD)_Both = 4.26(0.63), F(3, 196) = 12.01, p < .001.

⁹ Descriptive statistics of the rating on the given statement in the four conditions were as follows: Daejeon M(SD)_{Non-stigmatized} = 2.14(0.96), M(SD)_{Characteristic} = 2.58(1.37), M(SD)_Name = 4.32(0.87), M(SD)_Both = 4.38(0.73), F(3, 196) = 66.01, p < .001; Jeju M(SD)_{Non-stigmatized} = 2.10(1.13), M(SD)_{Characteristic} = 2.18(1.24), M(SD)_Name = 4.22(0.88), M(SD)_Both = 4.20(0.67), F(3, 196) = 70.58, p < .001; Taebaek M(SD)_{Non-stigmatized} = 2.10(1.13), M(SD)_{Characteristic} = 2.44(1.26), M(SD)_Name = 4.12(0.92), M(SD)_Both = 4.40(0.57), F(3, 196) = 66.01, p < .001.

¹⁰ Latent variables included individual attribution of responsibility, societal attribution of responsibility, anger, sympathy, punitive policy support, and assistive policy support.

¹¹ When the bootstrapping technique is used in a multiple-mediation model using AMOS version 21, the results are presented only for total mediation, and not for each specific mediation. To verify a specific indirect effect in the multiple-mediation model, it is necessary to create a phantom variable that does not affect the model fit or estimates (Macho & Ledermann, 2011). Following this guideline, we set phantom variables on the independent variables to create a modified multiple-mediation model.

¹² For reference, we also report the indirect effects based on the model where the interaction term was included. Of the interaction terms, one showed a significant indirect path in all three contexts: frame by characteristic-stigmatized headline to punitive policy support. The path was mediated by greater individual attribution of responsibility and then by anger, leading to greater support for punitive policies (Daejeon b_indirect = -.12, 95% CI[−0.31, −0.06]; Jeju b_indirect = -.20, 95% CI[−0.30, −0.07], Taebaek b_indirect = -.18, 95% CI [−0.33, −0.03]).

¹³ Another indirect path from the individual responsibility frame to assistive policy support was found in Daejeon only, which was mediated through increasing individual responsibility attribution and then decreasing sympathy, leading to less policy support (b_indirect = −.01, 95% CI [−0.03, −0.003]).

¹⁴ In Jeju, the name-stigmatized headline also increased individual responsibility attribution, and then anger, which in turn led to greater support for punitive policies (b_indirect = .05, 95% CI [0.004, 0.11]).

¹⁵ Only in Daejeon, the both-stigmatized headline increased individual responsibility attribution, which in turn decreased sympathy and lowered support for assistive policies (b_indirect = -.01, 95% CI [−0.02, −0.001]).

Additional information

Funding

This work was supported by the Institute of Communication Research at Seoul National University, the Korea News Agency Commission, and the National Research Foundation of Korea [to IJ Lim: NRF-2021S1A5C2A02088387].

Notes on contributors

In-Jae Lim

In-jae Lim (PhD, Sogang University, 2020) is a senior researcher in the Global Convergence Content Research Center at Sungkyunkwan University. Her research interests include digital journalism, environment communication, science communication, and health communication in mediated contexts.

Minsun Shim

Minsun Shim (PhD, University of Pennsylvania, 2008) is a professor in the Department of Media and Communication at Inha University. Her research interests include health and risk communication, and cognitive and affective processes in online communication and social media.

Chul-Joo Lee

Chul-joo Lee (PhD, University of Pennsylvania, 2009) is a professor in the Department of Communication at Seoul National University. His research interests include interplay between communication behaviors and social environment, health communication in mediated and interpersonal contexts, social marketing, and science communication.

Se-Hoon Jeong

Se-Hoon Jeong (PhD., University of Pennsylvania, 2008) is a full professor at the School of Media and Communication at Korea University. His research interests include persuasive, strategic and marketing communication.

Hyojin Lee

Hyojin Lee (M.S, Seoul National University) is a graduate student in the Department of Communication at Seoul National University. Her research interests include health communication in mediated contexts, media effects, and media psychology.