Sage Journals: Discover world-class research

Abstract

Public mood is a key concept in explaining collective activity, but the way social media shapes an audience’s public mood is still not fully understood. This study aims to explore how social media posts with various characteristics change public mood. The authors asked 351 participants to read 30 microblog newsletters with a 2 × 2 between-subject design (public × private sphere; positive × negative value). The results showed that (a) positive private information decreased negative public mood, (b) positive public information decreased positive public mood rather than increasing it, and (c) negative private information reduced the positive public mood of individuals who were high in justice sensitivity. The discussion focuses on the adverse effect of overexposure to positive public information and how individuals’ means of information processing vary.

Keywords

public sphere public opinion public mood information valence

On March 23, 2012, a man broke into a hospital in China, fatally stabbing one doctor and injuring three others. This tragic event was followed by a “public carnival” on the Internet, which included an online poll asking people how they felt about this news, and 4,018 out of 6,161 (65%) people chose “happy” (People’s Daily Online, 2012). Why did the public react positively to such tragic news? It makes sense that some netizens chose the happy response to express their negative perceptions of the doctor–patient relationship in China. Furthermore, the social media environment, in which official and nonofficial discourses fuse and news valence is sharpened, intensifies such a mood.

Emotion has been widely regarded as a fundamental component in mediating input information and political outcomes, including attitudes, judgments on political efficiency, the government’s reputation, voting intentions, and election success (Gross, 2008; Lecheler et al., 2013; Nabi, 1999). However, scholars lack knowledge of how the characteristics of information prime one’s emotional opinion of the political community—namely, public mood—and how a person’s emotional sensitivity to the information acts in between.

In this study, we intend to test the effects of two information characteristics. The first is the valence, with positive and negative dimensions. The second is the sphere, with public and private confines. From the perspective of information susceptibility, we take justice sensitivity into consideration. Using the experimental manipulation method, we can control the amount of exposure to information and determine its effect on public mood before and after exposure. Specifically, we aim to explore how valence and sphere in social media posts, along with individuals’ justice sensitivity, would change one’s public mood.

Private Mood Versus Public Mood

A crucial concept in the realm of emotional political opinion, public mood is people’s emotional reactions that stem from membership in political communities such as a nation or a political party (Rahn et al., 1996). It contains both positive and negative valences.

Public mood corresponds to the concept of private mood, or the broad sense of mood, which describes unspecified daily feelings. Conceptually, they share the same mental process. Although public mood is also affected by private mood, it is not just an aggregation of daily moods or a projection of personal moods onto a public target. The difference between public and private mood lies in public mood being specifically directed toward political communities (Rahn et al., 1996).

Rahn et al. (1996) summarized four clusters of causes that can affect public mood, taking stability (short term or long term) and collectivity (personal or collective) into consideration. Additional empirical analysis showed that positive public mood related to national identity and presidential approval ratings, whereas negative public mood related to gender, life satisfaction (a reflection of private moods), government trust, and collective economic expectations.

Public mood plays an important role in shaping public opinion because it immediately enables people to evaluate current situations and internalize political information, and influences how people make judgments and choices (Marcus, 2000). It also guides individual political behaviors and forms the culture. A systematic study involving experiments and a 20-year national survey revealed that emotions such as anger and fear could mobilize people to participate in elections (Valentino et al., 2011).

Public Mood and Social Media

Social media’s self-organization, plebification, and direct democratization features form a strong public sphere (Fuchs, 2014). A growing number of people are accustomed to consuming news through social media (Nielsen & Schrøder, 2014), and such consumption can influence individual political efficacy, knowledge, and offline participation (Kenski & Stroud, 2006). Therefore, social media has become an important medium for promoting campaigns and monitoring and predicting public sentiment.

News posted on social media shows distinct characteristics compared with traditional media. First, it provides a wide flow of information from positive to negative valences and from the public and private spheres. Some topics that are rare in traditional newspapers can be seen in social media (Zhao & Jiang, 2011). At the same time, when traditional news has discrete topic boundaries, the distinction among topics in online political discussion is blurred (Kim et al., 2016). Politicians also strategically use personal pages for their political campaigns and advertise their personal lives for election purposes. These factors combined add to the complexity of political news consumption in social media.

Second, in recent decades, news coverage has adopted more explicit emotional appeal strategies (Peters, 2011). The situation is further exacerbated online, where news needs to attract attention from fragmented audiences within limited word counts. Further influenced by their audiences’ preferences for more negative and cynical news content (Trussler & Soroka, 2014), news tweets are more often found to be biased in comparison with traditional news articles (Bjoergum, 2014), and to contain significantly different sentiment (Soroka et al., 2018). These factors lead to questions concerning how exposure to social media messages shapes public mood.

However, studies on media effects have mainly been conducted in natural environments or by using self-rating surveys with retrospective memories, which could not filter out media exposure effect or capture mood change precisely. To study the information effect on public mood, experimental manipulation is required. We will screen social media news messages into various information dimensions, use them as experimental stimuli, and simultaneously separate participants into groups that contain only one information feature to magnify the information effect.

Factors Influencing Public Mood

Public mood is influenced by short-term causes, such as private mood and personal experience, and long-term causes, such as demographics and national identity (Rahn et al., 1996). Considering the characteristics of social media, information valence and sphere should be carefully considered. At the same time, it is necessary to consider people’s emotional sensitivity toward public information, especially their justice sensitivity. In the following sections, we will separately discuss the influences of information valence, sphere, and justice sensitivity on public mood.

Information Valence

Valence refers to information’s emotional appeals. In news coverage and political advertisements, information valence, especially negative information, is frequently used in relation to the content. Specifically, information valence either attacks an opponent’s image or promotes trust-related information. At the same time, it can be regarded as how news and advertisements are framed, or whether the issue is attached to positive or detrimental meanings.

Psychological researchers have included emotion as an integral part of cognition processing and explanations of behavior. One group of researchers regarded mood as information (Schwarz et al., 1991). In the domain of heuristic processing, which deals with highly global but little personally relevant information, experienced feelings are taken as a direct source of judgments. Other researchers have introduced the affect infusion model (Forgas, 1995), which regards affect as a moderator in cognitive processing. Aside from heuristic processing, one’s mood could mediate the substantive process that deals with specific and highly relevant personal information by affect-priming mechanisms. Positive moods are also related to heuristic processing, and negative moods are related to systematic processing. These theories mainly focus on dimensions including positive and negative perspectives. With the development of appraisal and functional models, researchers have considered more discrete emotions like enthusiasm, anger, and fear as influence factors. Typically, the cognitive functional model (Nabi, 1999) focuses on negative emotional message effects in information processing. It assumes that emotions are connected with different goals and behavior intentions. For example, anger is related to offense and goal violation, leading to retributive behavior. If a message contains an emotion’s theme, then a corresponding emotional response is raised and people further decide whether to conduct systematic or heuristic processing based on levels of motivation and expectation.

In political research practices, Ridout and Franz (2011) compared four models of tone and emotional appeals in 2000 and 2004 campaign advertising: the intended effects model (positive advertisements raise the voting intention of the sponsor and negative advertisements decrease evaluations of the attacked); the backlash effects model (negative advertisements decrease the evaluations of the sponsor); the affect transfer model (emotional advertisements of the same valence exert the same effects); and the discrete emotions model (different emotional advertisements have different effects). The results showed that the intended effects model fully explained voting choices, where positive advertisements triggered positive evaluations of the advertisements’ sponsor and led to increased voting intention, and negative advertisements triggered negative evaluations of the attacked candidate and led to decreased voting intention. However, emotional advertisements did not directly motivate individuals’ voting intention, but raised emotions such as fear and enthusiasm, which stimulated evaluations and promoted the desirable behavior (Lecheler et al., 2013).

Framing studies have shown a similar picture. Positive and negative news framing was mediated by enthusiasm and anger in predicting intention of support for political issues (Lecheler et al., 2013). Putrevu (2014) induced moods in participants before having them read a positively or negatively attribute-framed appeal. The results showed that positive frames and moods induced relational processing with more category thoughts and higher clustered recalls, whereas negative frames and moods engendered item-specific processing with more attribute thoughts and fewer clustered recalls.

In general, cognitive models and emotional appeal and framing studies have mainly regarded mood or emotion as a moderator or factor in influencing the cognition and behavioral outcomes of information processing, but have not focused on what contributed to the change in public mood. However, they have all indicated a matching relationship between emotion-related information and emotional reactions. Therefore, we expected that information valence has a corresponding effect on public mood and formed the following hypotheses:

Hypothesis 1.1. Positive information increases positive public mood and decreases negative public mood.

Hypothesis 1.2. Negative information decreases positive public mood and increases negative public mood.

Furthermore, in view of the information valence research (Lecheler et al., 2013; Pratto & John, 1991; Putrevu, 2014; Schwarz et al., 1991), a fundamental asymmetry emerges in people’s evaluations of positive and negative events. Empirical studies have also shown that negative information is weighted more heavily than positive information in impression formation (Fiske, 1980) and public opinion toward the entity (Soroka, 2006). Therefore, positive events in news might not be able to increase the audience’s positive public emotion to the same extent as negative events. Therefore, we developed the following hypothesis:

Hypothesis 1.3. Compared with positive information, negative information has greater influence on public mood change by decreasing public positive mood and increasing public negative mood.

Information Sphere and Public or Private Relevance

In Habermas’s (1991) concept of spheres, the public sphere is a place for people to discuss societal problems and initiate political actions, and the private sphere is a place to discuss closed affairs. In the contemporary world, the public plays a vital role in mass communication, acting as a communicator and audience. Simultaneously, political issues have become increasingly entertaining. As a result, the boundaries between spheres have been blurred. Specifically, social media has become a public place in which a broad range of topics can be discussed (Edgerly et al., 2009). Issues on social media can be categorized as public-relevant or private-relevant information based on their content. Public-relevant information can include national, political, and macroeconomic issues, whereas private-relevant information is about intimacy, livelihood, and personal morals. Therefore, we still need to differentiate between the effect of discussion about public-relevant information and private-relevant information.

The mundane political culture, which is discursive, multivalent, inclusionary, and direct, has emerged in communication. It regards emotions and arguments as fundamental aspects because they enhance the reproduction of political culture (Merelman, 1998). In this kind of culture, private-relevant issues are linked with public opinion via emotions. China’s political ecology demonstrates unique features. Rather than a party, the subject of political communication is the central government, which aims to preserve its image and maintain stable public opinion. As a result, the official and nonofficial discourse universe runs parallel to Habermas’s (1991) public and private sphere definitions. Because nonofficial discourse takes short messages on the Internet as an important means of presentation, Chan and Zhou (2011) compared factors influencing Chinese online expression of political and livelihood issues. They found that the effective factors were similar and so were their effect sizes. At the same time, political and livelihood issue expression was affected by internal political attitudes. Moreover, individual and positive events were included as causes of public mood. In brief, public mood was directly related to public-relevant information. On the other hand, however, the current media ecology and existing studies have shown a possible link between private-relevant information and public mood (Chan & Zhou, 2011; Gross, 2008; Roddy & Garramone, 1988).

Although private-relevant information may be as important as public-relevant information, it tends to trigger different cognitive processes. Systematic processing might be activated when people read private-relevant information because this kind of information is closer to their personal lives. People will be more cautious in comprehending the content. Conversely, public-relevant information might have a closer relation to heuristic processing because it is separate from daily content. A study of news frames revealed that people were more emotionally engaged in episodic frames that focused on a specific example and event, but were more easily persuaded by thematic frames that presented the issue in a broader context (Gross, 2008). From another perspective, the arrogation of the boundary may antagonize people, decreasing their trust in the information provider and content. For example, Roddy and Garramone (1988) found that compared with attacking a specific political issue, political advertisements that attacked a candidate’s personal character increased the voting intention for the target and decreased the evaluation of the attacker’s commercials and character. However, in general, the private sphere effect is so diverse and implicit that few survey studies have been able to investigate it.

Considering the situation in China where the main divide of issues is between politics and livelihood rather than between two political ideologies, and where private-related information is more likely to induce systematic processing and emotional reactions, we formed the following hypothesis:

Hypothesis 2. Compared with public-relevant information, private-relevant information will increase positive public mood and decrease negative public mood more.

Justice Sensitivity

Considering the news about the murdered doctor, the netizens’ comments reflected justice-based judgments. Justice sensitivity, especially observer sensitivity, must be considered. Justice sensitivity is a kind of emotional sensitivity toward unjust information, which is highly correlated with the belief in an unjust world and anger (Schmitt et al., 2005). The perception of unfair events could be regarded as a personality trait for its stability. It is considered an important factor in examining the cognition–emotion processing in social contact. After exposure to unjust information, participants who are high in justice sensitivity will pay more attention to unjust words, interpret an ambiguous situation as more unjust, and have a higher recognition accuracy regarding unjust information (Baumert et al., 2011). Justice sensitivity will also enhance a source memory in a cheating context but consume the source memory for irrelevant and positive contexts (Bell & Buchner, 2010). Mikula et al. (1998) investigated the relationship between injustice evaluation and antecedent emotional events. They found that perceived injustice elicited anger and was strongly linked with all negative emotions. It also increases the possibility of political engagement (Rothmund et al., 2014) and altruistic sharing (Fetchenhauer & Huang, 2004).

Although scholars regard justice sensitivity as a vital trait that drives selective information processing and contributes to emotional and behavioral reaction, we still lack investigation into how it influences individuals in the processing of public mood change. Considering that unjust information is a type of negative information, it is reasonable to assume that negative rather than positive information pushes people with high levels of justice sensitivity to engage more heavily in the negative content and experience more negative emotions. Therefore, we developed the following hypothesis:

Hypothesis 3.1. When exposed to negative information, people with high levels of justice sensitivity will elicit a more intense response in public mood (i.e., have greater negative public mood and lower positive public mood) than people with low justice sensitivity.

In addition, the literature on justice sensitivity indicates its possible relationship with sphere-specific information. Although justice sensitivity has been found to diversify people’s reactions in response to both society-level (Schlösser et al., 2018) and person-specific (Baumert et al., 2011) inequality, there has been no comparison made between the spheres of the simulated situation. Justice sensitivity is generally closer to personal context. First, Edele et al. (2013) suggested that justice sensitivity includes the component of affective empathy, where people experience and emotionally respond to others’ distress. When a person’s distress is highlighted in the information, the sensitivity will be easily triggered. At the same time, moral value is one of the central issues in the private sphere. Second, the way justice sensitivity reacts to global inequality is contingent on people’s political ideologies and global identities (Reese et al., 2014), whereas private sphere information hides potential ideology and identity cues. Therefore, we expected to find a three-way interaction between valence, sphere, and justice sensitivity, leading to the following hypothesis:

Hypothesis 3.2. When exposed to negative private information, people with high levels of justice sensitivity will elicit a more intense response in public mood (i.e., have greater negative public mood and lower positive public mood) than when they are exposed to negative public information.

Methods

Design, Procedure, and Participants

To test the hypotheses, we conducted a survey experiment using a 2 (information valence: positive valence vs. negative valence) × 2 (information sphere: public sphere vs. private sphere) between-participant design.

The experiment included 378 undergraduate students from universities in Beijing, China. The survey was conducted in class. The participants were randomly assigned to one of the four conditions. The only difference among the conditions was the type of posts they read (corresponding to the information valence × information sphere). Before reading the posts, the participants were asked to report their sociodemographic information, current public and private moods, and justice sensitivity. Then they were exposed to 30 social media posts. All of the posts that each individual read were from one of the four conditions. To avoid a placebo effect, we used a cover story, requiring the participants to read a list of posts and rate their willingness to comment on and forward each of them on social media. The third step was the posttest. The participants were required to report their current public mood again to finish the experiment, and they then received a gift of 15 yuan. The researchers filtered out participants who rated the same option or did not complete eight consecutive items. The final sample contained 351 valid survey responses, 112 from males and 239 from females. Their mean age was 21.85 (SD = 2.74).

Material

All the real Morning Post news posts on Weibo (weibo.com) from September to November 2012 (2,387 posts) were collected as material. Weibo is one of the most popular social networking sites in China. In 2012, 56.4% of social networking site users in China used Weibo (China Internet Network Information Center, 2012). The Morning Post is an influential organization on Weibo. It is ranked as one of the top two newspaper accounts on Sina Weibo in 2012. We deleted emoticons and information regarding time stamps and hot topics to avoid the familiarity effect.

The next stage was the content coding process. The operational definition of the public sphere involves content related to public organizations, including government, education, commercial, and community issues. The operational definition of the private sphere refers to content related to personal topics, including individual and family issues. The material evaluation was conducted in two rounds. In the first round, we invited four participants to rate all the material according to the information sphere and valence. The evaluation options were: 1 (private sphere), 2 (partially private), 3 (mixed sphere), 4 (partially public), 5 (public sphere), and information valence (positive, neutral, or negative). We selected items whose average sphere ratings were greater than 4 as public-relevant information and below 2 as private-relevant information. Then we selected the items that received the same valence ratings from more than half the raters in terms of positive and negative valences. This sorting resulted in 120 distinctive messages from the positive private, positive public, negative private, and negative public spheres (each with 30 messages; for examples, see Appendix 1). The interrater consistency kappa is .79 for sphere and .60 for valence, which indicates a moderate level of agreement (McHugh, 2012). In the second round, we invited 60 raters to one of the four formal experimental conditions. They were asked to rate information sphere and valence on a 5-point Likert scale for each of the 30 messages. An independent t test showed that significant differences exist in the average evaluation of the messages the people read (t_sphere = −14.20, p < .001; t_valence = −18.18, p < .001). This indicates that the materials are distinctive and valid in terms of sphere (M_private = 1.79, SD = 0.67; M_public = 4.09, SD = 0.61) and valence (M_negative = 1.68, SD = 0.56; M_positive = 4.19, SD = 0.54).

In the analysis, private-relevant information was recoded as −1, public-relevant information was recoded as 1, negative information was recoded as −1, and positive information was recoded as 1.

Measures

Public Mood

We adopted the public mood scale constructed by Rahn et al. (1996), which includes positive mood (hopeful, happy, proud, and secure) and negative mood (angry, afraid, frustrated, and sad) indicators, asking participants how strongly they experienced these moods when they thought about Chinese society on a 5-point scale from 1 (never) to 5 (always). The internal consistency was 0.83 for positive public mood and 0.82 for negative public mood. The mean pre-reading public positive mood was 3.00 (SD = 0.83) and the mean public negative mood was 1.92 (SD = 0.70). The mean post-reading public positive mood was 2.74 (SD = 0.90) and the mean public negative mood was 2.00 (SD = 0.83).

Observer Justice Sensitivity

Using a 6-point scale, we employed 10 observer perspective items derived from the justice sensitivity scale developed by Schmitt et al. (2005). The participants were required to respond to the following prompt: “Now consider situations in which you observe someone else is disadvantaged or being exploited.” Such situations included: “It bothers me when someone gets something they do not deserve” and “I am upset when someone does not get a reward they have earned.” They rated their agreement from 0 to 5 (Cronbach’s α = 0.87, M = 2.88, SD = 0.71).

Control Variable

Rahn et al. (1996) mentioned the intercorrelation between private mood and public mood. Therefore, in this study, private mood was controlled. We used Thompson’s (2007) 5-point Likert scale, asking the extent to which the participants felt positive moods (active, determined, attentive, inspired, and alert) and negative moods (afraid, nervous, upset, hostile, and ashamed). The responses ranged from 1 (never) to 5 (always), with Cronbach’s α = 0.92 for positive private mood and 0.90 for negative private mood.

Analysis Strategies

The data analysis was conducted in two steps. First, to explore the variance of the public emotion (before and after priming) under exposure to different information contexts, the paired sample t test was applied. The second step was to explore the three-way interaction of the information value, public and private sphere relevance, and justice sensitivity of the participants. Because justice sensitivity is a continuous variable, we adopted linear regression on the post-reading public mood, which could explore not only the two-way interactions but also the three-way interactions in the same model. We conducted four-level regression on positive and negative public moods, respectively. Model 1 contained pretests of public mood and private mood as control variables. Information valence, sphere, and justice sensitivity were included in Model 2 as the main predictors. Model 3 included a two-step interaction between sphere, valence, and justice sensitivity, and in Model 4, a three-step interaction was included for analysis. The interaction term was created by multiplying the centered value of the three variables. Centering was used to avoid nonessential multicollinearity and for clearer interpretations (Cohen et al., 2013).

Results

We first conducted a paired-sample t test to compare the mood before and after the information presentation in the four conditions (see Table 1 and Figure 1). The results showed that negative private-relevant information, negative public-relevant information, and positive public-relevant information reduced the public positive mood (t₁ = −3.72, p < .001; t₂ = −6.01, p < .001; t₃ = −2.33, p < .05). But positive mood was not significantly affected by positive private information, which partially rejects Hypothesis 1.2. On the other hand, negative information increased public negative mood (t_private = 3.27, p < .01; t_public = 4.02, p < .001); positive private information reduced public negative mood (t = −3.20, p < .01); and individuals maintained their negative mood levels before and after reading the positive public information. These results support Hypotheses 1.1 and 1.3, where negative information exerts a consistent and strong influence on decreasing positive public mood and increasing negative public mood.

Table 1.

Paired Sample t Test of Information Sphere, Valance, and Public Mood.

Between-subject conditions		Public positive mood		Cohen’s d	t test	p (95% CI)	Public negative mood		Cohen’s d	t test	p (95% CI)
Information sphere	Information valance	Before M (SD)	After M (SD)	Cohen’s d	t test	p (95% CI)	Before M (SD)	After M (SD)	Cohen’s d	t test	p (95% CI)
Private	Negative	3.00 (0.85)	2.70 (0.94)	0.40	−3.72***	< .001 (−.47, −.14)	1.93 (0.72)	2.18 (0.86)	−0.36	3.27**	.002 (.10, .41)
Private	Positive	3.07 (0.82)	3.01 (0.83)	0.10	−.953	.343 (−.16, .06)	1.81 (0.68)	1.64 (0.63)	0.35	−3.20**	.002 (−.30, −.07)
Public	Negative	2.95 (0.85)	2.41 (0.84)	0.62	−6.01***	< .001 (−.72, −.36)	2.00 (0.70)	2.35 (0.86)	−0.42	4.02***	<.001 (.18, .52)
Public	Positive	2.97 (0.81)	2.85 (0.89)	0.25	−2.33*	.022 (−.22, −.02)	1.92 (0.71)	1.80 (0.76)	0.19	1.73	.088 (−.25, .02)

Note. CI = confidence interval.

*p < .05. **p < .01. ***p < .001.

Figure 1.

Mean Public Mood in Four Conditions. (a) Public positive mood and (b) Public negative mood.

We then built regression models to predict post-reading public mood. The results are shown in Tables 2 and 3 for negative and positive moods, respectively.

Table 2.

Multi-Level Regression on Public Negative Mood.

	Model 1			Model 2			Model 3			Model 4
	β	t	p (95% CI)	β	t	p (95% CI)	β	t	p (95% CI)	β	t	p (95% CI)
Constant	.70***	3.60	<.001 (.32, 1.09)	.75***	4.16	<.001 (.40, 1.11)	.72***	3.97	<.001 (.36, 1.07)	.74***	4.10	<.001 (.39, 1.09)
Public positive mood before	−.05	−1.00	.317 (−.15, .05)	−.05	−1.14	.255 (−.15, .04)	−.05	−1.14	.255 (−.15, .04)	−.06	−1.16	.247 (−.15, .04)
Public negative mood before	.63***	10.51	<.001 (.51, .74)	.57***	10.10	<.001 (.46, .68)	.59***	10.44	<.001 (.48, .70)	.58***	10.43	<.001 (.47, .69)
Private positive mood	.07	1.42	.157 (−.03, .17)	.09	1.80	.073 (−.01, .18)	.08	1.53	.128 (−.021, .17)	.07	1.43	.153 (−.03, .17)
Private negative mood	.04	0.70	0.51 (−.08, .17)	.06	0.93	.351 (−.06, .18)	.07	1.18	0.239 (−.05, .19)	.07	1.16	.248 (−.05, .19)
Information valence				−.23***	−6.76	<.001 (−.30, −.17)	−.23***	−6.66	<.001 (−.30, −.16)	−.23***	−6.75	<.001 (−.30, −.16)
Information sphere				.06	1.80	.073 (−.01, .13)	.06	1.69	.092 (−.01, .13)	.05	1.49	.138 (−.02, .12)
Justice sensitivity				.13**	2.61	.010 (.03, .23)	.11*	2.05	.042 (.01, .21)	.09	1.75	.082 (−.01, .19)
Sphere × justice sensitivity							.03	0.54	.589 (−.07, .13)	.03	0.49	.624 (−.07, .12)
Valence × justice sensitivity							−.14**	−2.72	.007 (−.23, −.04)	−.12*	−2.30	.022 (−.22, −.02)
Sphere × valence							.01	0.12	.905 (−.06, .07)	.01	0.21	.838 (−.06, .08)
Sphere × valence × justice sensitivity										−.09	−1.73	.086 (−.19, .01)
Δ R²				.102			.014			.005
R ²	.323			.425			.439			.444

Note. Two-tailed tests. CI = confidence interval.*p < .05. **p < .01. ***p < .001.

Table 3.

Multi-Level Regression on Public Positive Mood.

	Model 1			Model 2			Model 3			Model 4
	β	t	p (95% CI)	β	t	p (95% CI)	β	t	p (95% CI)	β	t	p (95% CI)
Constant	.42*	2.28	.024 (.06, .78)	.39*	2.20	.029 (.04, .73)	.38*	2.14	.034 (.03, .72)	.40*	2.29	.023 (.06, .75)
Public positive mood before	.68***	14.04	<.001 (.58, .77)	.68***	14.65	<.001 (.59, .77)	.68***	14.67	<.001 (.59, .77)	.68***	14.72	<.001 (.59, .77)
Public negative mood before	−.10	−1.76	.079 (−.21, .01)	−.05	−0.98	.328 (−.16, .05)	−.05	−0.99	.323 (−.16, .05)	−.06	−1.05	.296 (−.16, .05)
Private positive mood	.09	1.88	.061 (−.01, .19)	.08	1.75	.081 (−.01, .18)	.10*	2.03	.044 (.01, .19)	.09	1.91	.057 (−.01, .19)
Private negative mood	.16*	2.53	.012 (.04, .28)	.14*	2.37	.019 (.02, .26)	.13*	2.17	.032 (.01, .25)	.13*	2.14	.033 (.01, .24)
Information valence				.17***	5.04	<.001 (.10, .24)	.17***	5.00	<.001 (.10, .23)	.16***	4.92	<.001 (.10, .23)
Information sphere				−.08*	−2.33	.021 (−.14, −.01)	−.08*	−2.43	.016 (−.15, −.02)	−.09**	−2.66	.009(−.16, −.02)
Justice sensitivity				−.09	−1.90	.062 (−.19, .01)	−.11*	−2.13	.034 (−.20, −.01)	−.12*	−2.46	.015 (−.22, −.03)
Sphere × justice sensitivity							.10	1.95	.053 (−.01, .19)	.09	1.90	.059 (−.01, .19)
Valence × justice sensitivity							.03	0.60	.547 (−.07, .13)	.05	1.04	.299 (−.05, .15)
Sphere × valence							.05	1.53	.127 (−.02, .12)	.06	1.64	.102 (−.01, .12)
Sphere × valence × justice sensitivity										−.10*	−2.12	.035 (−.20, −.01)
Δ R²				.051			.008			.006
R ²	.479			.530			.538			.544

Note. Two-tailed tests. CI = confidence interval.*p < .05. **p < .01. ***p < .001.

Considering negative public mood, information valence had a main effect (β = −.23, p < .001), which indicated that individuals who read positive information had lower levels of negative mood than individuals who read negative information. In Models 3 and 4, the effect of justice sensitivity and its interaction with valence were robust (β₁ = −.14, p < .01; β₂ = −.12, p < .05). We then conducted simple slope analysis using the PROCESS method with 1,000 bootstrap replications (Hayes, 2017). The bootstrap results also support the interaction effect (R² change = .014, F = 8.26, p < .01). As Figure 2 shows, individuals with high levels of justice sensitivity had a stronger negative mood increase after reading negative information (β_{low justice sensitivity} = −.13, p < .01; β_{high justice sensitivity} = −.34, p < .001). This supports Hypothesis 3.1.

Figure 2.

Information Valence and Justice Sensitivity Predicting Negative Mood.

Information valence, information sphere, and justice sensitivity each had a main effect on public positive mood (β₁ = .16, p < .001; β₂ = −.09, p < .01; β₃ = −.12, p < .05). Positive information increased positive mood more than negative information, and individuals exposed to public-relevant information had a lower positive mood than individuals exposed to private-relevant information, supporting Hypothesis 2. Moreover, highly justice-sensitive individuals had lower positive moods regardless of information type.

Model 4 showed a significant three-way effect between valence, sphere, and justice sensitivity on public positive mood (β = −.10, p < .05). The bootstrap results show that the three-way interaction is salient (R² change = .006, F = 4.49, p < .05). The detailed interplay pattern is illustrated in Figure 3. When participants were exposed to positive information (Figure 3(a)), the valence effect was not significant under justice sensitivity (β_{low justice sensitivity} = −.03, p > .05; β_{high justice sensitivity} = −.04, p > .05). When participants were exposed to negative information (Figure 3(b)), individuals with high justice sensitivity reported a medium level of positive mood regardless of the information valence (β = .01, p > .05). The positive mood of individuals with low justice sensitivity decreased after reading negative private-relevant information, which partially supports Hypothesis 3.2.

Figure 3.

Information Valence, Sphere, and Justice Sensitivity Predicting Post-Positive Mood. (a) Exposed to positive information and (b) Exposed to negative information.

Discussion

Our study examined how information sphere, valence, and justice sensitivity affected public mood. Information valence significantly influenced public mood. Negative valence information is more influential in decreasing the public’s positive mood and increasing the negative mood. These results support our hypothesis 1.3 and Ridout and Franz’s (2011) intended effects model. Negative valence is widely used for political advertising and news framing, which has a consistent and strong influence on people’s cognition, emotions, and behavior (Lecheler et al., 2013; Pratto & John, 1991; Putrevu, 2014; Schwarz et al., 1991).

The positive valence information effect is interesting. It did not simply decrease negative mood and raise positive public mood, as we predicted. The results showed that positive public information stabilized the negative mood but decreased the positive mood, indicating that overdecorating behavior in political persuasion might result in a backlash against what it intended to shape. This unintended result of strategic messages, which could also be regarded as a boomerang effect, is caused by either the processing part of information or resistance to the intended outcome (Byrne & Hart, 2016). Empirically, the reinforcing effect of partisan news on existing voter preferences is small and modest (Dilliplane, 2014). Although the news valence effect is predictable in voting propensity, it is specialized for less sophisticated and nonpartisan individuals. Positive tones tend to have an adverse but insignificant influence on partisans (Eberl et al., 2015). On the other hand, exposure to positive private information could help relieve negative public mood. This finding indicates that, compared with public-relevant information, private-relevant information may ease a frustrated mood toward society but could not increase positive perceptions such as pride and security. Despite the features of the sphere effect where negative public mood changes beyond a sphere, private-relevant information may activate low motivation and peripheral processing, where individuals increase their focus on emotion-inducing environments and thus alleviate the negative effect.

Regarding the information sphere, positive mood decreased more when participants were exposed to public-relevant information than when they were exposed to private-relevant information. The information sphere had no effect on the changes in negative mood, which may indicate that private sphere information is as important as public sphere information with regard to negative public mood. Considering that public mood is improved based on individual mental processing, it is reasonable that it correlates with private moods and is affected by information that evokes a private mood (i.e., private-relevant information; Rahn et al., 1996). In addition, the ways in which members of the public lead their lives may reflect governmental efficiency on a micro level and thereby influence the evaluation of society. The reason why the evaluation effect of private-relevant information is similar to public-relevant information on negative moods but not positive moods is possibly that negative mood is more diffusive than positive mood (Nabi, 1999). The results also highlight that positive public mood is more sphere-specific, and little influenced by private information.

Individuals with high levels of justice sensitivity experienced decreased positive mood and increased negative mood to a larger extent than individuals with low levels of justice sensitivity, regardless of information type. Justice sensitivity is related to the connotation of injustice tolerance in observed information and preference for punishment, which may enhance the availability and accessibility of injustice concerns about the information gained. As a result, negative mood increases. This negative emotion toward sensitivity might, on the other hand, inactivate or impair the cognitive processing in a positive mood.

The three-way interaction indicates the complexity between content and individual characteristics with regard to public mood. In general, sensitivity is related to the increase of negative public mood and the decrease of positive public mood when people are exposed to private negative information. This finding is in line with prior research which showed that people who had high levels of justice sensitivity had strong emotional reactions to negative information (Bell & Buchner, 2010). In addition, because the nature of justice sensitivity is individualistic and private-relevant information arouses more empathy, which is the core concept of observer justice sensitivity, private negative information could certainly raise the reaction of people with high justice sensitivity at large. At the same time, the moderation effect of justice sensitivity on negative mood is sphere-free. This also reveals the close connection between justice sensitivity and negative mood. Negative empathy is the vital element of observer justice sensitivity (Edele et al., 2013). Highly justice-sensitive people have an easier time falling into a negative mood when seeing unfairness. In addition, the moral outrage produced by justice sensitivity directs further political engagement (Rothmund et al., 2014).

Implications

Government and political parties have the motivation to monitor and guide public opinion. Without a benign interaction between government and the public, public opinion would negatively affect the governing foundations and impede social stability. However, in China, attempts to promote “positive energy” (e.g., illustrating role models and delivering positive policy and livelihood news) usually do not achieve what the government and the media expect, while negative current affairs sometimes elicit a “public carnival.” Chen (1999) categorized public opinion in contemporary China into three characteristics: contradictive, emotional, and diversified. Our results provide evidence of the existence of this phenomenon and offer possible mechanisms and explanations for its formation. Facing negative and private-relevant information such as the news at the beginning of this article, people still have a tendency to decrease their positive mood toward their political communities. This influence is especially intensified when people interpret injustice in a story.

As for ways of guiding public opinion, apart from conveying the mainstream ideology, focusing on private-relevant and positive news in agenda-setting may be beneficial for increasing the positive public mood and decreasing the negative public mood. Additionally, justice could be embedded in the news frame to offer readers an affirmative perspective.

Limitations and Future Research

Our study has some limitations. First, our research sample only included students. Future research could expand the sample to make it more representative, and investigate the sociodemographic influence on public mood change. Second, the identification of mood could be more specific, because the type of mood will bring about differences in appraisal processing. Anger and fear in negative emotions and enthusiasm and nurturing love in positive emotions are of the same valence but differ in appraisal, therefore leading to varying attributions, behavioral intentions, and information processing toward events (Lerner & Keltner, 2000). Third, the positive and negative information could be sorted into specific categories, and the difference in the public mood and its cause could be studied. Fourth, in our study, the participants were asked to rate their public mood just after information exposure. It is still unknown whether this information had a long-lasting effect on emotion. Although peripheral processing can be triggered by emotion-inducing environments, its effect on participants' future behavior is less stable and predictable. Studying the long-tern effect can add to our understanding of media influence and public opinion.

Appendix 1
Examples of Experiment Materials.

Information sphere Information valance Examples

Private Positive 情侣同年同月同日同医院生, 还曾同班同桌同赴美]高苗和殷虎, 两人都出生于 1983 年11 月 12 日, 都在西安市第四医院出生。高中时, 他们从同桌发展到挚友再到情侣。大二时, 两人又成功取得全额奖学金, 一同赴美留学。昨天, 相恋 11 年的他们在西安完婚[A couple born on the same day, in the same month and year, in the same hospital, who were also classmates, went to the United States together] Gao Miao and Yin Hu were both born on November 12, 1983, in Xi’an Fourth Hospital. In high school, they developed from being classmates to close friends and then lovers. In their sophomore year, they successfully obtained a full scholarship to study in the United States together. Yesterday, they married in Xi’an after falling in love 11 years earlier.

Private Negative 男子不忍母亲去世离开给遗体抹盐防腐伴尸近一年]台湾屏东市邱某不忍相依为命的母亲过世, 以盐巴涂抹尸体“防腐”, 天天相伴长达半年, 出嫁的姊姊多次电话向母亲问安, 邱都推说母亲外出, 姊姊起疑返家将邱骗开, 冲进母亲卧室, 才发现母亲已成干尸。[A man cannot bear his mother's death. He wipes the body with salt, preventing corrosion of the corpse for nearly a year]Qiu, from Pingtung City, Taiwan, cannot bear the death of his mother; they kept each other alive. He wiped the body with a salt embalming “preservative” for up to half a year. His sister made many phone calls to their mother. Qiu said their mother was out. The sister came back home and tricked Qiu into going out; she rushed into their mother’s bedroom, only to find that their mother had become a mummy.

Public Positive 五一恢复有望？全国假日办建议对五一假期调整并适度延长]针对今年十一黄金周出现的问题, 全国假日办建议加快推进带薪休假制度落实, 适时启动黄金周假日旅行机制研究。在不增加原有法定节假日的前提下, 可通过协调和腾挪其他假期和周末, 对五一假期进行调整并适度延长。[Hope of recovering the May Day holiday—National Holiday Office recommended adjusting and appropriately extending the May Day holiday]For problems emerging in this year’s National Day Golden Week, the National Holiday Office recommended improving the implementation of the paid vacation system, and launching the study of Golden Week holiday travel behavior. On the premise of not increasing the original statutory holidays, the May Day holiday can be adjusted and appropriately extended through coordination and transfer of other holidays and weekends.

Public Negative 江西赣州在建隧道塌方16人被困]昨晚11点左右, 江西省赣州市龙南县发生在建隧道塌方, 导致16人被困。据了解, 塌方隧道内有80米左右的活动空间, 但由于塌方泥土较多, 隧道口被封堵。[A tunnel under construction in Ganzhou, Jiangxi Province, collapsed and trapped 16 people] At around 11 o’clock last night, a tunnel collapse occurred in Longnan County, Ganzhou City, Jiangxi Province, causing 16 people to be trapped. There is about 80 meters of movement space in the collapsed tunnel, but due to the large amount of mud in the collapsed tunnel, the tunnel entrance is blocked.

Information sphere	Information valance	Examples
Private	Positive	情侣同年同月同日同医院生, 还曾同班同桌同赴美]高苗和殷虎, 两人都出生于 1983 年11 月 12 日, 都在西安市第四医院出生。高中时, 他们从同桌发展到挚友再到情侣。大二时, 两人又成功取得全额奖学金, 一同赴美留学。昨天, 相恋 11 年的他们在西安完婚[A couple born on the same day, in the same month and year, in the same hospital, who were also classmates, went to the United States together] Gao Miao and Yin Hu were both born on November 12, 1983, in Xi’an Fourth Hospital. In high school, they developed from being classmates to close friends and then lovers. In their sophomore year, they successfully obtained a full scholarship to study in the United States together. Yesterday, they married in Xi’an after falling in love 11 years earlier.
Private	Negative	男子不忍母亲去世离开给遗体抹盐防腐伴尸近一年]台湾屏东市邱某不忍相依为命的母亲过世, 以盐巴涂抹尸体“防腐”, 天天相伴长达半年, 出嫁的姊姊多次电话向母亲问安, 邱都推说母亲外出, 姊姊起疑返家将邱骗开, 冲进母亲卧室, 才发现母亲已成干尸。[A man cannot bear his mother's death. He wipes the body with salt, preventing corrosion of the corpse for nearly a year]Qiu, from Pingtung City, Taiwan, cannot bear the death of his mother; they kept each other alive. He wiped the body with a salt embalming “preservative” for up to half a year. His sister made many phone calls to their mother. Qiu said their mother was out. The sister came back home and tricked Qiu into going out; she rushed into their mother’s bedroom, only to find that their mother had become a mummy.
Public	Positive	五一恢复有望？全国假日办建议对五一假期调整并适度延长]针对今年十一黄金周出现的问题, 全国假日办建议加快推进带薪休假制度落实, 适时启动黄金周假日旅行机制研究。在不增加原有法定节假日的前提下, 可通过协调和腾挪其他假期和周末, 对五一假期进行调整并适度延长。[Hope of recovering the May Day holiday—National Holiday Office recommended adjusting and appropriately extending the May Day holiday]For problems emerging in this year’s National Day Golden Week, the National Holiday Office recommended improving the implementation of the paid vacation system, and launching the study of Golden Week holiday travel behavior. On the premise of not increasing the original statutory holidays, the May Day holiday can be adjusted and appropriately extended through coordination and transfer of other holidays and weekends.
Public	Negative	江西赣州在建隧道塌方16人被困]昨晚11点左右, 江西省赣州市龙南县发生在建隧道塌方, 导致16人被困。据了解, 塌方隧道内有80米左右的活动空间, 但由于塌方泥土较多, 隧道口被封堵。[A tunnel under construction in Ganzhou, Jiangxi Province, collapsed and trapped 16 people] At around 11 o’clock last night, a tunnel collapse occurred in Longnan County, Ganzhou City, Jiangxi Province, causing 16 people to be trapped. There is about 80 meters of movement space in the collapsed tunnel, but due to the large amount of mud in the collapsed tunnel, the tunnel entrance is blocked.

Footnotes

Acknowledgement

We thank Ms. Zhang Guannan for her contributions in scale development and data collection.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research,authorship,and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research,authorship and/or publication of this article: This work was supported by Key Project of National Social Science Foundation of China (Grant No. 19ZDA358).

References

Baumert

Gollwitzer

Staubach

Schmitt

(2011). Justice sensitivity and the processing of justice-related information. European Journal of Personality, 25(5), 386–397. https://doi.org/10.1002/per.800

Bell

Buchner

(2010). Justice sensitivity and source memory for cheaters. Journal of Research in Personality, 44(6), 677–683. https://doi.org/10.1016/j.jrp.2010.08.011

Bjoergum

M. C. H.

(2014). The credibility of news media: The difference in framing between traditional media and Twitter after the Boston Marathon bombing [Unpublished doctoral dissertation]. Hawaii Pacific University.

Byrne

Hart

P. S.

(2016). The boomerang effect: A synthesis of findings and a preliminary theoretical framework. Annals of the International Communication Association, 33(1), 3–37. https://doi.org/10.1080/23808985.2009.11679083

Chan

J. M.

Zhou

(2011). Expressive behaviors across discursive spaces and issue types. Asian Journal of Communication, 21(2), 150–166. https://doi.org/10.1080/01292986.2010.543558

Chen

(1999). Public opinion. China Radio and Television Press.

China Internet Network Information Center. (2012). Research report of 2012 Chinese netizens social networking site application use. http://www.cnnic.net.cn/hlwfzyj/hlwxzbg/sqbg/201302/P020130219611651054576.pdf

Cohen

West

S. G.

Aiken

L. S.

(2013). Applied multiple regression/correlation analysis for the behavioral sciences. Taylor & Francis.

Dilliplane

(2014). Activation, conversion, or reinforcement? The impact of partisan news exposure on vote choice. American Journal of Political Science, 58(1), 79–94. https://doi.org/10.1111/ajps.12046

10.

Edgerly, S., Vraga, E., Fung, T., Moon, T. J., Yoo, W. H., & Veenstra, A. (2009, October). YouTube as a public sphere: The proposition 8 debate. In Association of Internet Researchers Conference, Milwaukee, WI.

11.

Eberl

J. M.

Boomgaarden

H. G.

Wagner

(2015). One bias fits all? Three types of media bias and their effects on party preferences. Communication Research, 44(8), 1125–1148. https://doi.org/10.1177/0093650215614364

12.

Edele

Dziobek

Keller

(2013). Explaining altruistic sharing in the dictator game: The role of affective empathy, cognitive empathy, and justice sensitivity. Learning and Individual Differences, 24, 96–102. https://doi.org/10.1016/j.lindif.2012.12.020

13.

Fetchenhauer

Huang

(2004). Justice sensitivity and distributive decisions in experimental games. Personality and Individual Differences, 36(5), 1015–1029. https://doi.org/10.1016/s0191-8869(03)00197-1

14.

Fiske, S. T. (1980). Attention and weight in person perception: The impact of negative and extreme behavior. Journal of personality and Social Psychology, 38(6), 889. https://doi.org/10.1037/0022-3514.38.6.889

15.

Forgas

J. P.

(1995). Mood and judgment: The affect infusion model (AIM). Psychological Bulletin, 117(1), 39–66. https://doi.org/10.1037/0033-2909.117.1.39

16.

Fuchs, C. (2014). Social media and the public sphere. tripleC: Communication, Capitalism & Critique. Open Access Journal for a Global Sustainable Information Society, 12(1), 57--101. https://doi.org/10.31269/triplec.v12i1.552

17.

Gross

(2008). Framing persuasive appeals: Episodic and thematic framing, emotional response, and policy opinion. Political Psychology, 29(2), 169–192. https://doi.org/10.1111/j.1467-9221.2008.00622.x

18.

Habermas

(1991). The structural transformation of the public sphere: An inquiry into a category of bourgeois society. MIT Press.

19.

Hayes

A. F.

(2017). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. Guilford Press.

20.

Kenski

Stroud

N. J.

(2006). Connections between Internet use and political efficacy, knowledge, and participation. Journal of Broadcasting and Electronic Media, 50(2), 173–192. https://doi.org/10.1207/s15506878jobem5002_1

21.

Kim

E. H.

Jeong

Y. K.

Kim

Kang

K. Y.

Song

(2016). Topic-based content and sentiment analysis of Ebola virus on Twitter and in the news. Journal of Information Science, 42(6), 763–781. https://doi.org/10.1177/0165551515608733

22.

Lecheler

Schuck

A. R. T.

Vreese

C. H. de

(2013). Dealing with feelings: Positive and negative discrete emotions as mediators of news framing effects. Communications, 38(2), 189–209. https://doi.org/10.1515/commun-2013-0011

23.

Lerner

J. S.

Keltner

(2000). Beyond valence: Toward a model of emotion-specific influences on judgement and choice. Cognition and Emotion, 14(4), 473–493. https://doi.org/10.1080/026999300402763

24.

Marcus

G. E.

(2000). Emotions in politics. Annual Review of Political Science, 3(1), 221–250. https://doi.org/10.1146/annurev.polisci.3.1.221

25.

McHugh

M. L.

(2012). Interrater reliability: The kappa statistic. Biochemia Medica: Biochemia Medica, 22(3), 276–282.

26.

Merelman

R. M.

(1998). The mundane experience of political culture. Political Communication, 15(4), 515–535. https://doi.org/10.1080/105846098198876

27.

Mikula

Scherer

K. R.

Athenstaedt

(1998). The role of injustice in the elicitation of differential emotional reactions. Personality and Social Psychology Bulletin, 24(7), 769–783. https://doi.org/10.1177/0146167298247009

28.

Nabi

R. L.

(1999). A cognitive‐functional model for the effects of discrete negative emotions on information processing, attitude change, and recall. Communication Theory, 9(3), 292–320. https://doi.org/10.1111/j.1468-2885.1999.tb00172.x

29.

Nielsen

R. K.

Schrøder

K. C.

(2014). The relative importance of social media for accessing, finding, and engaging with news: An eight-country cross-media comparison. Digital Journalism, 2(4), 472–489. https://doi.org/10.1080/21670811.2013.872420

30.

People’s Daily Online. (2012, March 24). A malignant injury incident occurred in the First Affiliated Hospital of Harbin Medical University. https://baike.baidu.com/reference/7290482/a812wrP3R9nMOe9dA8ojpJjhDxBFzzQIR6rnmT2suOpLBnjS27mQtGUIXxIFe87vp3EqdHGJcUhxQ9LhQZDTZ IllNYdaA5GzgFk

31.

Peters

(2011). Emotion aside or emotional side? Crafting an “experience of involvement” in the news. Journalism, 12(3), 297–316. https://doi.org/10.1177/1464884910388224

32.

Pratto

John

O. P.

(1991). Automatic vigilance: The attention-grabbing power of negative social information. Journal of Personality and Social Psychology, 61(3), 380–391. https://doi.org/10.1037/0022-3514.61.3.380

33.

Putrevu

(2014). The influence of mood and attribute framing on consumer response toward persuasive appeals. Journal of Current Issues and Research in Advertising, 35(2), 107–125. https://doi.org/10.1080/10641734.2014.899528

34.

Rahn

W. M.

Kroeger

Kite

C. M.

(1996). A framework for the study of public mood. Political Psychology, 17(1), 29–58. https://doi.org/10.2307/3791942

35.

Reese

Proch

Cohrs

J. C.

(2014). Individual differences in responses to global inequality. Analyses of Social Issues and Public Policy, 14(1), 217–238. https://doi.org/10.1111/asap.12032

36.

Ridout

T. N.

Franz

M. M.

(2011). The persuasive power of campaign advertising. Temple University Press.

37.

Roddy

B. L.

Garramone

G. M.

(1988). Appeals and strategies of negative political advertising. Journal of Broadcasting and Electronic Media, 32(4), 415–427. https://doi.org/10.1080/08838158809386713

38.

Rothmund

Baumert

Zinkernagel

(2014). The German “Wutbürger”: How justice sensitivity accounts for individual differences in political engagement. Social Justice Research, 27(1), 24–44. https://doi.org/10.1007/s11211-014-0202-x

39.

Schlösser

Steiniger

Ehlebracht

Fetchenhauer

(2018). How justice sensitivity predicts equality preferences in simulated democratic systems. Journal of Research in Personality, 73, 75–81. https://doi.org/10.1016/j.jrp.2017.11.005

40.

Schmitt

Gollwitzer

Maes

Arbach

(2005). Justice sensitivity: Assessment and location in the personality space. European Journal of Psychological Assessment, 21(3), 202–211. https://doi.org/10.1027/1015-5759.21.3.202

41.

Schwarz

Bless

Bohner

(1991). Mood and persuasion: Affective states influence the processing of persuasive communications. Advances in Experimental Social Psychology, 24, 161–199. https://doi.org/10.1016/S0065-2601(08)60329-9

42.

Soroka

(2006). Good news and bad news: Asymmetric responses to economic information. Journal of Politics, 68(2), 372–385. https://doi.org/10.1111/j.1468-2508.2006.00413.x

43.

Soroka

Daku

Hiaeshutter-Rice

Guggenheim

Pasek

(2018). Negativity and positivity biases in economic news coverage: Traditional versus social media. Communication Research, 45(7), 1078–1098. https://doi.org/10.1177/0093650217725870

44.

Thompson

E. R.

(2007). Development and validation of an internationally reliable short-form of the Positive and Negative Affect Schedule (PANAS). Journal of Cross-Cultural Psychology, 38(2), 227–242. https://doi.org/10.1177/0022022106297301

45.

Trussler

Soroka

(2014). Consumer demand for cynical and negative news frames. International Journal of Press/Politics, 19(3), 360–379. https://doi.org/10.1177/1940161214524832

46.

Valentino

N. A.

Brader

Groenendyk

E. W.

Gregorowicz

Hutchings

V. L.

(2011). Election night’s alright for righting: The role of emotions in political participation. Journal of Politics, 73(1), 156–170. https://doi.org/10.1017/s0022381610000939

47.

Zhao

Jiang

(2011). An empirical comparison of topics in Twitter and traditional media (Singapore Management University School of Information Systems Technical Paper Series). http://www.mysmu.edu/faculty/jingjiang/papers/TechReport%28Zhao2011%29.pdf