Parasocial Relationships With President Trump as a Predictor of COVID-19 Information Seeking

As the number of COVID-19 cases climbed in the United States, President Trump came under fire for dismissing its severity despite evidence to the contrary. Those most impressionable by the President's claims might be those who have established a parasocial relationship with him. Results of this survey study indicate that each of the three parasocial variables assessed (i.e., friends, understanding, and brand), were negatively related to information seeking for COVID-19. Moreover, initial evidence was found that identification with the Trump brand was an indirect influence of information seeking. Consistent with the supported model, women and Democrats reported higher information seeking for COVID-19 than men and other political groups.

Introduction

On January 23, 2020, the World Health Organization warned all governments that they must prepare for the emergence of the novel coronavirus, COVID-19, within their own borders (Yamey and Gonsalves, 2020). Between January 22 and April 27, 2020, 981,246 cases of COVID-19 were confirmed in the United States (CDC, 2020). President Trump has been harshly criticized by scholars for his management of this pandemic, with speculation that his communication about COVID-19 countered many efforts to keep people healthy and halt the spread of the virus. For example, Gadarian et al. (2020) wrote:

…the U.S. crisis response suffered both from a lack of coordinated information provided to the public and from President Trump's deliberate undercutting of expert messaging about the escalating pandemic. Downplaying COVID-19 by comparing it to a mild flu, decrying it as a “hoax by the Democrats”, and dismissing its seriousness as late as March 10 in saying “it will go away” each undermined the public health effort” (p. 1).

Indeed, President Trump classified COVID-19 not simply as a non-threat to the well-being of U.S. residents, but actively negated the claims that COVID-19 was a danger, claiming that these concerned voices were part of a conspiracy theory concocted by Democrats in an attempt to undermine his credibility (Dyer, 2020b).

This paper seeks to explore how perceived relationships with President Trump affected COVID-19 information seeking. Parasocial relationships, which are perceived relationships with individuals we are exposed to through media, have been identified to predict health information behavior (e.g., Brown and Basil, 1995; Williamson and Hocken, 2010; Moe, 2012; Tiger, 2013). This paper seeks to identify the ways in which parasocial relationships with President Trump predict information seeking among United States residents in regard to COVID-19.

Information Behavior

Information behavior pertains to “the accidental encountering, need, finding, choosing, using, and sometimes avoiding” of information (Case, 2012, p. 1). Information acquisition behaviors can be demarcated between the actions of intentionally seeking out new information and the more passive actions of cognitively processing information that is delivered to people without request (Kelly et al., 2014). As Case (2012) explains, just because information is placed before a person does not mean they will process any of the information; in fact, they may make an active decision to ignore the information. As such, passive information behavior is the decision to digest information that finds its way to a person while active information behavior is the intentional seeking of additional information (Kelly et al., 2014).

Information behavior becomes complex when the information subject is health-related. Knowledge of health-related information can save one's life by teaching the behaviors necessary to avoid illness (Maciosek et al., 2010), yet this same information behavior can also induce anxiety as it makes one aware of possible health issues (Case et al., 2005; Kelly et al., 2014). Research consistently reports that until a health topic becomes salient through either personal experience or the experience of someone they care about, an individual will not engage in information seeking (e.g., Brashers et al., 2002; Rice, 2006; Ayers and Kronenfeld, 2007; Khoo et al., 2008; Hogan and Brashers, 2009). In the face of an illness, information seeking can be stress-relieving rather than stress-inducing as it reduces uncertainty about the health condition (Tian and Robinson, 2008).

Information about the threat of COVID-19 was communicated inconsistently to residents of the United States. While it was considered to be a worldwide emergency outside of the United States and by the Center for Disease Control (CDC) within the United States, President Trump and his administration continued to downplay the risk of COVID-19, indicating it was a non-threat until mid-March 2020 (Dyer, 2020a). In other words, President Trump informed individuals engaged in passive information behavior that active information seeking for COVID-19 was unnecessary. This raises the question of how individuals who looked to President Trump as the authority to communicate information during the crisis would be affected by his downplaying of COVID-19 during a time of global health salience. A potential explanatory variable may be parasocial relationships.

Parasocial Relationships

Parasocial relationships are perceived interpersonal relationships between real people and media personas, which may be real people or fictional characters (Grant et al., 1991). These relationships are developed through the one-way communication that takes place when these personas communicate directly to consumers or are captured communicating with others through various media channels (Horton and Wohl, 1956). Horton and Wohl (1956) first proposed parasocial relationships when they noticed the high level of attachment and perceived sense of understanding some viewers had for news anchors whom they had never personally met. Although parasocial relationships have been historically studied among television and film personas, recent scholarship has identified the formation of parasocial relationships with social media personas (e.g., Lee, 2013; Labrecque, 2014; Ledbetter and Redd, 2016).

Similar to interpersonal relationships, parasocial relationships develop and strengthen over time (Perse and Rubin, 1990; Auter and Palmgreen, 2000) and provide a sense of social security (Stever, 2013). Time allows media consumers more opportunities to perceived similarity of personality (Rubin et al., 1985; Auter and Palmgreen, 2000; Rubin and Step, 2000; Rubin and Rubin, 2001; Tian and Hoffner, 2010), attitude and problem solving skill similarity (Turner, 1993; Auter and Palmgreen, 2000), and identification with the individual (Auter and Palmgreen, 2000; Tian and Hoffner, 2010), which strengthens the parasocial relationship. This means that identifying with the persona, understanding of the persona's way of thinking, and identifying with persona's goals or brand are all components of parasocial relationships. Identification with a persona's gender and attitudes are particularly influential (Hoffner and Cantor, 1991; McDonald and Kim, 2001; Eyal and Rubin, 2003; Hoffner and Buchanan, 2005).

Although parasocial relationships have been found to follow the predictions of numerous interpersonal theories, there is no overarching theory of parasocial relationships. To address this, Tukachinsky and Stever (2019) proposed a theoretical model of parasocial relationship development based upon a review of 60 years of parasocial research and Knapp's (1978) model of interpersonal relationships. Tukachinsky and Stever's (2019) model proposes four stages of parasocial relationship development. Importantly, Tukachinsky and Stever (2019) explain that not all consumers will experience each stage of a parasocial relationship. Rather, it is possible to plateau at any of the stages if social attraction wanes or expectations are violated. The four stages are:

1. Initiation: In the initiation stage, media consumers find themselves attracted to the persona physically and/or psychologically but know little about them. Therefore, they give abundant attention and critical evaluation to the persona in an attempt to satisfy their uncertainty.

2. Experimentation: In the experimentation stage, consumers seek more interactions with the persona (i.e., more places to view their communicative behaviors) to evaluate the extent to which the persona fits well with the consumers' beliefs and values.

3. Intensification: If a consumer reaches the intensification stage, they have cemented a parasocial relationship. At this phase, consumers develop affective connections to the persona and begin thinking about them when not consuming their media.

4. Integration/Bonding: At integration/bonding, a parasocial relationship has escalated to the point in which the relationship has become part of their identity that is socially recognizable, meaning they appreciate their fandom being recognized by others and feel that their bond with the persona is special.

As such, individuals who reach a high level of parasocial relationship feel that the behaviors and communication of the persona they have bonded with are reflections of themselves and therefore are prone to find their behaviors both justifiable and reasonable (Tukachinsky and Stever, 2019).

Residents of the United States have had decades to develop parasocial relationships with the current Republican President, Donald J. Trump. He has been a notable businessman for over 40 years (Koffler, 2015), hosted his own reality television show The Apprentice from 2004–2015 (IMDB, 2020), and has had an active Twitter account for over a decade (Twitter, 2020). In terms of potential parasocial relationship building, this means that individuals have had a chance to observe President Trump through multiple platforms to see if they identify with him as someone who could be a friend, someone whose problem solving skills make sense, and form an attachment to the Trump brand that preexisted long before he ran for President. Those most likely to have identified with him are those who can connect to him through gender and attitudes (Hoffner and Cantor, 1991; McDonald and Kim, 2001; Eyal and Rubin, 2003; Hoffner and Buchanan, 2005); therefore the following hypotheses are proposed:

H1:  Men will have stronger parasocial relationships with President Trump than women.

H2:  Republicans will have stronger parasocial relationships with President Trump than individuals who affiliate with other parties.

To reiterate, individuals are unlikely to seek out information about a health topic until they themselves or someone they know make that health topic salient (Johnson and Case, 2012), and people feel as though they know individuals whom they have formed parasocial relationships with (Auter and Palmgreen, 2000). Numerous research examples have demonstrated that having a parasocial relationship with a celebrity or character increases information behavior and/or knowledge about a health topic when that persona reveals that they are struggling with the health topic or endorse awareness of it. Examples of celebrity parasocial relationships that have resulted in increased health information seeking include Michael J. Fox and Parkinson's disease (Moe, 2012), Whitney Houston and alcohol addiction (Tiger, 2013), Magic Johnson and HIV (Brown and Basil, 1995), Mark McGuire and steroids (Brown et al., 2003), and Patrick Swayze and pancreatic cancer (Williamson and Hocken, 2010). In terms of characters, suicide awareness increased when character Lawrence Kutner committed suicide on the television show House (DeGroot and Leith, 2018), fans of Monk became more knowledgeable about obsessive compulsive disorder after observing the show's protagonist struggle with it (Hoffner and Cohen, 2012), and children have broadly been shown to care about eating healthy when their favorite television character does (Kraak and Story, 2015).

What research has not yet had reason to demonstrate, is whether someone in a position of fame and influence can temper health information seeking by negating the salience. Preliminary findings on the influence of President's Trump rhetoric regarding COVID-19 indicates that it is possible. Barrios and Hochberg (2020) found that the population of Trump voters in a county predicted perceived risk of COVID-19 and information seeking. Further, residents of Republican counties have been found to be less likely to comply with stay at home orders compared to Democrats (Painter and Qiu, 2020).

As mentioned, Tukachinsky and Stever's (2019) theoretical model of parasocial relationships predicts that at the third stage of parasocial relationship development, intensification, individuals' parasocial relationships begin to affect attitudes and beliefs. By the intensification stage, individuals have already determined that they typically agree with the persona's attitudes and behaviors. Thus, at the intensification stage and beyond, individuals are likely to become increasingly supportive of the persona's behaviors, supporting them even when they perceive those behaviors to be illogical (Tukachinsky and Stever, 2019). As such, individuals who are in an intensification or integration/bonding stage of parasocial relationships with a persona are likely to consistently believe that the persona's behaviors are reasonable and their decisions are justified. This means that individuals at higher stages of parasocial relationship with President Trump are likely to believe that his methods for handling a problem are correct. As such, the higher parasocial relationship with President Trump, the more likely individuals are to believe that his attitudes and behaviors regarding COVID-19 are correct. Because of this, degree of parasocial relationship with President Trump should directly affect someone's attitude about how he has handled COVID-19.

That attitude, in turn, is likely to affect information seeking for COVID-19. Because COVID-19 is a pandemic, it is possible that when seeking information about COVID-19 initially, individuals who had high parasocial relationships with President Trump heard his dismissal, had positive attitudes about their parasocial relation's abilities to manage the virus, and as a result felt no need to seek additional information. Likewise, those who had low parasocial relationships with President Trump, who are those who do not identify with his problem solving skills (Auter and Palmgreen, 2000), would hear his dismissive rhetoric and because of their associated poor attitude toward him, expect him to handle COVID-19 inappropriately, prompting further information seeking so that they can prepare for the virus. As such, the following hypothesis is proposed:

H3:  Parasocial relationships with Donald Trump will indirectly predict COVID-19 information seeking through the mediation of attitude.

The hypothesis can be depicted in the mediated model shown in Figure 1. Because men and Republicans should have higher parasocial relationships with President Trump than other demographics, and it is expected that gender and political affiliation will moderate information seeking behavior scores. Therefore, the following hypotheses are proposed:

H4:  Men will have lower information seeking regarding COVID-19 than women.

H5:  Republicans will have lower information seeking regarding COVID-19 than those affiliated with another party.

FIGURE 1

Figure 1. Proposed model.

Method

Participants

In total, 529 individuals completed the online questionnaire for this study. On average, participants were 36.0 (SD = 14.47) years old. There were 150 participants who identified as male, 379 that identified as female, and none that identified with an alternative gender option. Political affiliation broke down as follows: 246 Democrats, 58 Independents, 32 Libertarians, 131 Republicans, and 62 selected “Other” as their party affiliation. (Notably, two participants contacted the lead researcher after completing the questionnaire to confess that though they were registered as a Republican, they no longer felt connected to the party, and had chosen Other as their political affiliation. It is possible that additional respondents also felt this way, accounting for the abnormally high Other sample.) Education of participants broke down as follows: four some high school, 28 high school diploma or GED, 132 some college, 40 Associate's degree, 134 Bachelor's degree, and 191 graduate or professional degree. Occupational fields broke down as follows: 15 clerical, 137 education, 7 entertainment, 7 factory, 65 health, 10 legal, 24 managerial, 3 military, 1 non-farm labor, 15 public service, 17 skilled labor, 33 sales or retail, 38 technology, 6 transportation, 89 unemployed, 6 retired, 55 “other,” and one individual chose not to identify their occupational field. It should be noted that at the time this questionnaire was disseminated, the stay at home orders and subsequent effect on the economy resulted in high unemployment across the country, so for many the unemployed status was potentially new.

Procedure

By April 2020, most states within the United States were in some mode of lockdown with stay at home orders issued by local governments in an attempt to curve the spread of COVID-19 with the exception of North Dakota, South Dakota, Nebraska, Iowa, and Arkansas (Secon, 2020). The four researchers shared a brief, IRB approved, recruitment message and link to the online questionnaire hosted in Qualtrics via their social media accounts and personal email contacts during April 2020 when most of the United States was under a stay at home order due to COVID-19.

Although using this non-probability sampling method was not optimal, it was necessary given the unique circumstances of the global climate. By the time IRB approval for this study was granted, President Trump's narrative about COVID-19 had started changing, meaning that the longer data collection took, the more convoluted the narrative prompting the questionnaire would become. The fastest options for data collection available to the researchers were convenience sampling and using a crowdsource site. Using a crowdfunded site like MTurk was decided against given recent discussion of the risk of bot participants (c.f., McEwan, 2020). This left convenience sampling as an unpreferable, yet necessarily fast sampling method.

Once participants followed the link, the study was introduced as an effort to understand information behaviors regarding COVID-19 and purported to require no more than 5 minutes of time for completion. Potential participants were informed that they must be at least 18 years old and reside in the United States to participate in the questionnaire.

Instrumentation

Information Seeking

Information seeking was measured through the Kelly et al. (2014) active information behavior assessment. The assessment was composed of 6 Likert-type items, each with a 7-point response scale ranging from Disagree Strongly to Agree Strongly. Kelly et al. (2014) found that the measure showed evidence of content and concurrent validity.

Parasocial Relationships

Parasocial relationships were assessed through three measures as refined by Chung and Cho (2017). These measures were chosen because of President Trump's unique position of having a strong social media presence, a brand prior to becoming president, and a strong presence in traditional media. These assessments focused on how well respondents understood President Trump's behaviors and decisions (i.e., understanding), whether they identified characteristics from President Trump that made them believe they could develop a friendship if given the opportunity (i.e., friends), and identification and trust with the Trump brand (i.e., brand). Chung and Cho's (2017) initial item development adapted items from existing parasocial measures to better suit parasocial relationships created through social media, which removed language in older measures that constrained media consumption to television viewing at designated times. The dimension of understanding was assessed through a 6-item assessment, friendship was assessed through a 3-item assessment, and brand was assessed through a 5-item assessment. Chung and Cho's (2017) brand measure was adapted to match identification with the Donald Trump brand rather than a product brand promoted by the persona on social media; because this measure was adapted, the utilized items are presented in the Appendix. Each of these items were Likert-type with a 7-point response scale ranging from Disagree Strongly to Agree Strongly. Chung and Cho (2017) found that the measures yielded evidence of convergent validity.

Attitude

McCroskey and Richmond's (1996) generalized attitude measure was utilized to assess attitudes toward President Trump's management of COVID-19. The measure is composed of six semantic differential items with 7-point response options. McCroskey and Richmond (1996) report that the measure had good face validity.

Results

Unidimensional Measurement Models

Before testing hypotheses, four of the unidimensional measurement models were subjected to confirmatory factor analysis (CFA) to confirm the predicted factor structure using the AMOS maximum likelihood parameter estimation algorithm. (The friendship measure could not be assessed through CFA because with only three items, the measure is just-identified.) The information seeking measure had acceptable fit statistics. The attitude measure had acceptable fit statistics except for a slightly elevated root mean square error approximation (RMSEA); however, RMSEA is greatly influenced by minor misfit, so it is not concerning if RMSEA is the only fit statistics indicating slight misfit (Chen et al., 2008).

The brand and understanding measures yielded multiple signs of misfit. As such, the measures were examined for individual items that caused a statistically significant amount of residual error on other items. These problematic items were identified and removed. Two items were removed from the understanding measure (understand quite well and can feel emotions) and one was removed from the brand measure (delivers promises). The fit statistics for original and finalized measures are provided in Table 1. Descriptive statistics for the final measures, which were used in hypothesis testing, are provided in Table 2.

TABLE 1

Table 1. Fit statistics.

TABLE 2

Table 2. Descriptive statistics.

Because the friendship measure could not be assessed via CFA, alternative evidence of validity was sought. Because each of the parasocial relationship measures assess evidence of the same relationship, they should be highly correlated. The three finalized parasocial measures, which include the original friends measure, respecified understanding measure, and adapted respecified brand measure are strongly correlated as expected, yielding evidence of convergent validity. The correlation matrix of finalized measures is provided in Table 3.

TABLE 3

Table 3. Correlation matrix.

Hypothesis Testing

Hypothesis one predicted that men would have a stronger parasocial relationship with Donald Trump than women. Three independent samples t-tests were used to test this hypothesis, comparing the scores of friends, understanding, and brand by sex. In each instance, the data supported hypothesis one, indicating that men scored higher than women across friends [t(527) = 6.88, p < 0.001, ${\bar{\text{x}}}_{\text{men}}$ = 3.13, ${\bar{\text{x}}}_{\text{women}}$ = 2.00], understanding [t(527) = 5.41, p < 0.001, ${\bar{\text{x}}}_{\text{men}}$ = 2.93, ${\bar{\text{x}}}_{\text{women}}$ = 2.17], and brand [t(527) = 4.64, p < 0.001, ${\bar{\text{x}}}_{\text{men}}$ = 3.19, ${\bar{\text{x}}}_{\text{women}}$ = 2.34].

Hypothesis two predicted that Republicans would have stronger parasocial relationships with Donald Trump than individuals who affiliated with other parties. One-way ANOVAs with Tukey post-hoc tests were used to test this hypothesis. Each hypothesis was confirmed, but with more nuance than expected. For the parasocial dimension of friends, Republicans scored statistically significantly higher than any other group and Democrats scored statistically significantly lower than any group [F_{(4, 524)} = 91.82, p < 0.001]. In the parasocial dimension of understanding, Republicans scored statistically significantly higher than any other political group, and Democrats also scored statistically significantly lower than any group except for Independents [F_{(4, 524)} = 73.93, p < 0.001]. Finally, in the parasocial dimension of brand, Republicans scored statistically significantly higher than any group, Libertarians scored statistically significantly higher than any group except Republicans, and Democrats scored statistically significantly lower than any group [F_{(4, 524)} = 12.26, p < 0.001]. Table 4 provides a summary of the Tukey post-hoc results.

TABLE 4

Table 4. Tukey summary hypothesis 2.

Hypothesis three predicted that parasocial relationships with Donald Trump would affect information seeking through the mediation of attitude. The predicted model is depicted in Figure 1. To test this hypothesis, structural equation modeling was used with the AMOS maximum likelihood parameter estimation algorithm. The global fit statistics for the model were as follows: GFI = 1.00, CFI = 1.00, RMSEA = 0.00, SRMR = 0.01, χ²(3, N = 529) = 2.66, p = 0.45. Indirect effects were verified through bootstrapping. The indirect effect for brand (−0.15 < ρ < −0.06) was statistically significant, but not for understanding (0.00 < ρ < 0.02) nor friends (−0.01 < ρ < 0.01). As such, only the simple causal chain from brand to attitude to information seeking was supported. The supported chain is depicted in Figure 2. Fit statistics for the supported chain were as follows: GFI = 1.00, CFI = 1.00, RMSEA = 0.02, SRMR = 0.01, χ²(1, N = 529) = 1.25, p = 0.26.

FIGURE 2

Figure 2. Supported model.

Hypothesis four predicted that men would report lower information seeking for COVID-19 than women. The data were consistent with this hypothesis [t(527) = −3.70, p = 0.01, $\bar{\text{x}}$_men = 5.25, $\bar{\text{x}}$_women = 5.58].

Hypothesis five predicted that Republicans would report statistically significantly lower information seeking for COVID-19 than individuals who identified with other parties. This hypothesis was not supported by the data. Although Republicans did exhibit the lowest mean information seeking scores, they only reported information seeking statistically significantly lower than Democrats [F_{(4, 524)} = 4.55, p < 0.001]. Table 5 displayed the post-hoc results.

TABLE 5

Table 5. Tukey summary hypothesis 5.

Discussion

COVID-19, the first pandemic in over a century, has created unique research circumstances in that though it was already recognized as a pandemic by the rest of the world, United States residents were told for months by President Trump that it was going to be a non-issue. As such, this was a novel instance in which a parasocial relationship could be used to diminish salience of a health topic rather than build it.

Hypotheses one and two predicted that men and Republicans, those individuals who could identify likeness with President Trump most easily, would have higher parasocial relationships with him. The data supported both hypotheses. Notably, the measures utilized in this study assess parasocial relationship strength through three different assessments: identification with President Trump as a person, perceived ability to be a friend with President Trump, and trust/investment in the Trump brand. It is particularly notable that though some political groups had statistically significantly higher parasocial relationships with President Trump than others, the average strength of parasocial relationship was only above neutral (4 on a 7-point scale) for one group: Republicans on the dimensions of friendship ($\bar{\text{x}}$ = 4.19) and brand ($\bar{\text{x}}$ = 4.79). As such, across the sample, neither men nor women, Republicans nor Democrats nor third-party members, found President Trump to be a highly understandable person, someone they could envision being friends with, nor someone whose brand they could trust.

Hypothesis three predicted a model in which the strength of parasocial relationships with President Trump across the dimensions of understanding, friendship, and brand would indirectly predict information seeking for COVID-19 through the mediation of attitude. The data supported only the path from brand to information seeking where the stronger the parasocial relationship, the lower the information seeking. It is interesting that only trust in and identification with the Trump brand truly influenced how much information seeking participants reported engaging in for COVID-19. This implies that it is not President Trump's personality or understanding of his decision making that the influenced health information seeking, but rather the enduring Trump brand that has been built across decades. In fact, the relationship between attitude regarding how well President Trump was managing COVID-19 and the Trump brand (0.90 < ρ < 0.92) was statistically significantly higher than the relationships between attitude and friendship (0.79 < ρ < 0.85) or understanding (0.75 < ρ < 0.81).

Scholars in marketing and advertising have theorized for decades that a brand can be conceptualized such that it has a personality (c.f., Aaker and Fournier, 1995). This is why individuals can become parasocially attached to a brand. When individuals form such a relationship with a brand, they exhibit consistent behaviors toward the brand, including forgiving mishaps, maintaining loyalty over time, and endorsing the brand even when aware that it does not fit the needs of those it is endorsed to (Fedorikhin et al., 2008; Patwardhan and Balasubramanian, 2011). This means that those who have become parasocially attached to the Trump brand have a loyalty to it that would override evidence that calls the integrity of the brand into question. This loyalty to brand is evident in the results of this study with the r² = 0.83 for the relationship between the Trump brand and attitude, indicating that 83% of the variance in one's attitude regarding how well President Trump was managing COVID-19 was explained by their attitude toward the Trump brand alone. It is also noteworthy that while it may seem that overall the Trump brand was able to account for only 5% of the variance in information seeking, this is a mediated relationship. As such, r² = 0.05 is only the square root of the impact it has on the variance in participants' COVID-19 information seeking because it is an indirect effect.

Hypotheses four and five predicted that males and Republicans would report the lowest information seeking out of their demographic groups. Men did indeed report statistically significantly lower information seeking than women. Although Republicans reported the lowest information seeking, they only reported behaviors statistically significantly lower than Democrats. The hypotheses and data, however, were consistent with the predictions of Figure 2. Although the differences were moderate, these superficial demographic predictors of parasocial relationships with President Trump do predict lower information seeking for COVID-19. This is particularly notable in that men have been found to die of COVID-19 more frequently than women (Jin et al., 2020). The current data indicate that a small part of this variance in death rates may be attributable to men being less informed about the virus than women.

This paper constitutes one of the first studies to use Tukachinsky and Stever's (2019) theoretical model of parasocial relationship enhancement to guide a study. It was consistent with both their model and the present dataset that individuals with the strongest parasocial relationships with President Trump would have the highest attitudes toward his management of COVID-19 and that Republicans and men would be those most likely to have those high parasocial relationships. Ironically, in their paper explicating the model, Tukachinsky and Stever (2019) specifically propose that an individual allowing their parasocial relationship with Donald Trump formed through watching The Apprentice to override their political judgement about him (c.f., Gabriel et al., 2018) as an example of the integration phase of their model. In this first use of the theoretical model, the data were consistent with Tukachinsky and Stever's (2019) predictions indicating that future studies on parasocial relationships and health should consider using this model as guiding framework.

Implications

In times of stress, individuals are more likely to engage in information seeking to reduce anxiety and increase their ability to predict the future (Johnson and Case, 2012). In the present dataset, information seeking was reported to be considerably above neutral for each demographic group represented in this study. These high means may be explained by the need to reduce anxiety during a pandemic, state lockdowns, and a time of high unemployment, especially in the wake of more threatening narratives about COVID-19. As such, anxiety management may have been a much stronger predictor of information seeking than the parasocial variables focused upon in this study. Scholars are encouraged to assess anxiety management as well as information seeking behaviors if conditions for a replication study ever arise.

Though anxiety management may have explained more variance, the brand of President Trump was indeed an indirect predictor of participants' COVID-19 information seeking. Despite President Trump changing his stance on COVID-19 in mid-March of 2020, designating it as a viable threat, parasocial relationships through his brand still managed to negatively affect information seeking. This highlights the pervasive and long-lasting effect the Trump brand established over time through parasocial relationships. Though the detrimental effect was small, this data collection happened 1 month after his narrative changed. Had data been collected earlier, while he still maintained that COVID-19 was not problematic, it is possible that a more robust variance in behaviors would have been found. That parasocial relationships with President Trump's brand still negatively affected information seeking for COVID-19 a month after his narrative changed is a substantive finding. That brand held the strongest relationship with attitude about how President Trump was managing COVID-19 is not truly surprising. Identifying with a brand means to have faith in and loyalty to that brand (Chung and Cho, 2017). Having faith in and loyalty to the Trump brand means having faith that the brand will endure, meaning faith that President Trump can overcome the obstacles he is faced with. As such, identification with the Trump brand constitutes a level of skepticism when the brand is threatened by other voices, even if they are voices of authority like the World Health Organization or the Center for Disease Control. What is concerning is that the brand identification was found to be powerful enough to discourage individuals from seeking new information, be it information that confirms or refutes their assumed truths.

Herein lies the implications of this work for those whose role is to set health narratives. Because the decision to not seek information was instigated by a parasocial relationship, there is an innate tendency for relationship holders to defend the opinions of that parasocial source (Tukachinsky and Stever, 2019). This means that expert opinions are unlikely to counter the narrative (unless individuals happen to have a stronger parasocial relationship with the expert than the person downplaying the illness). As such, the narrative must change at its source to have an impact on those negatively affected.

Limitations

This study had a number of limitations. First, data collection took place during April 2020, yet President Trump ended his consistent declarations that COVID-19 was a non-issue in mid-March 2020. Effects for this study may have been stronger if data could have been collected sooner. Second, ethnicity was not collected for this study which could have been a very informative variable, particularly in terms of parasocial relationships. Third, the method of data collection resulted in a disproportionate convenience sample, weighted toward Democrats and individuals with graduate degrees. Fourth, cross sectional data was used in this study, which means that the mediated effect supported in hypothesis 3 can only be confirmed to the extent that it was demonstrated that effects were consistent with patterns that would be present in the case of mediation, not confirmation that a causal mediated relationship occurred. Finally, though the data were consistent with Figure 2, this supported model had only one degree of freedom, which carries the risk of inflated fit statistics. However, it is unlikely this was the case as the path represented in Figure 2 was also supported in the original model test of Figure 1, which had ample degrees of freedom.

Conclusion

Previous research in health information behaviors and parasocial relationships have focused on how personas can cause health topics to become salient, thereby increasing information seeking. This study is novel in that it studied attempts by a persona to negate salience of a health topic during a pandemic. Though the authors sincerely hope that there is never an opportunity to replicate such a study, future assessment of parasocial relationships as a demotivator of health information seeking would help understand this unique parasocial phenomenon.

Data Availability Statement

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

Ethics Statement

The studies involving human participants were reviewed and approved by IRB Committee at North Carolina A&T State University. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.

Author Contributions

SK, RG, MM, and SD confirmed that they collaborated from the data collection through composition of this manuscript. All authors contributed to the article and approved the submitted version.

Conflict of Interest

SD was extensively involved in politics with a Republican affiliation.

The remaining authors declare that the research was conducted in the absence of any relationships that could be construed as a potential conflict of interest.

References

Aaker, J., and Fournier, S. (1995). A brand as a character, a partner and a person: three perspectives on the question of brand personality. Adv. Consum. Res. 22, 391–395.

Google Scholar

Auter, P. J., and Palmgreen, P. (2000). Development and validation of a parasocial interaction measure: the audience–persona interaction scale. Commun. Res. Rep. 17, 79–89. doi: 10.1080/08824090009388753

CrossRef Full Text | Google Scholar

Ayers, S. L., and Kronenfeld, J. J. (2007). Chronic illness and health-seeking information on the Internet. Health 11, 327–347. doi: 10.1177/1363459307077547

PubMed Abstract | CrossRef Full Text | Google Scholar

Barrios, J. M., and Hochberg, Y. (2020). Risk Perception Through the Lens of Politics in the Time of the COVID-19 Pandemic, (No. w27008). Cambridge, MA: National Bureau of Economic Research. doi: 10.3386/w27008

CrossRef Full Text | Google Scholar

Brashers, D. E., Goldsmith, D. J., and Hsieh, E. (2002). Information seeking and avoiding in health contexts. Hum. Commun. Res. 28, 258–271. doi: 10.1111/j.1468-2958.2002.tb00807.x

CrossRef Full Text | Google Scholar

Brown, W., and Basil, M. (1995). Media celebrities and public health: responses to “Magic” Johnson's HIV disclosure and its impact on AIDS risk and high-risk behaviors. Health Commun. 7, 345–371. doi: 10.1207/s15327027hc0704_4

CrossRef Full Text | Google Scholar

Brown, W., Basil, M., and Bocarnea, M. (2003). The influence of famous athletes on health beliefs and practices: Mark McGwire, child abuse prevention, and androstenedione. J. Health Commun. 8, 41–57 doi: 10.1080/10810730305733

PubMed Abstract | CrossRef Full Text | Google Scholar

Case, D. O. (2012). Looking for Information: A Survey of Research on Information Seeking, Needs, and Behavior, 3rd Edn. London: Elsevier Academic Press.

Google Scholar

Case, D. O., Andrews, J. E., Johnson, J. D., and Allard, S. L. (2005). Avoiding versus seeking: The relationship of information seeking to avoidance, blunting, coping, dissonance, and related concepts. J. Med. Libr. Assoc. 93, 353–362.

PubMed Abstract | Google Scholar

CDC (2020). Previous U.S. COVID-19 Case Data. Centers for Disease Control and Prevention. Available online at: https://www.cdc.gov/coronavirus/2019-ncov/cases-updates/previouscases.html (accessed June 28, 2020).

Google Scholar

Chen, F., Curran, P. J., Bollen, K. A., Kirby, J., and Paxton, P. (2008). An empirical evaluation of the use of fixed cutoff points in RMSEA test statistic in structural equation models. Sociol. Methods Res. 36, 462–494. doi: 10.1177/0049124108314720

PubMed Abstract | CrossRef Full Text | Google Scholar

Chung, S., and Cho, H. (2017). Fostering parasocial relationships with celebrities on social media: implications for celebrity endorsement. Psychol. Market. 34, 481–495. doi: 10.1002/mar.21001

CrossRef Full Text | Google Scholar

DeGroot, J. M., and Leith, A. P. (2018). R.I.P. Kutner: parasocial grief following the death of a television character. OMEGA J. Death Dying 77, 199–216. doi: 10.1177/0030222815600450

PubMed Abstract | CrossRef Full Text | Google Scholar

Dyer, O. (2020a). COVID-19: US testing ramps up as early response draws harsh criticism. Br. Med. J. 368:m1167. doi: 10.1136/bmj.m1167

PubMed Abstract | CrossRef Full Text | Google Scholar

Dyer, O. (2020b). Trump claims public health warnings on COVID-19 are a conspiracy against him. Br. Med. J. 368:m941. doi: 10.1136/bmj.m941

PubMed Abstract | CrossRef Full Text | Google Scholar

Eyal, K., and Rubin, A. M. (2003). Viewer aggression and homophily, identification, and parasocial relationships with television characters. J. Broadcast. Electron. Media 47, 77–98. doi: 10.1207/s15506878jobem4701_5

CrossRef Full Text | Google Scholar

Fedorikhin, A., Park, C. W., and Thomson, M. (2008). Beyond fit and attitude: the effect of emotional attachment on consumer responses to brand extensions. J. Consum. Psychol. 18, 281–291. doi: 10.1016/j.jcps.2008.09.006

CrossRef Full Text | Google Scholar

Gabriel, S., Paravati, E., Green, M. C., and Flomsbee, J. (2018). From apprentice to president: the role of parasocial connection in the election of Donald Trump. Soc. Psychol. Personal. Sci. 9, 299–307. doi: 10.1177/1948550617722835

CrossRef Full Text | Google Scholar

Gadarian, S. K., Goodman, S. W., and Pepinsky, T. B. (2020). Partisanship, health behavior, and policy attitudes in the early stages of the COVID-19 pandemic. SSRN. doi: 10.2139/ssrn.3562796

CrossRef Full Text

Grant, A. E., Guthrie, K., and Ball-Rokeach, S. J. (1991). Television shopping: a media system dependency perspective. Communic. Res. 18, 773–798. doi: 10.1177/009365091018006004

CrossRef Full Text | Google Scholar

Hoffner, C., and Buchanan, M. (2005). Young adults' wishful identification with television characters: the role of perceived similarity and character attributes. Media Psychol. 7, 323–349. doi: 10.1207/S1532785XMEP0704_2

CrossRef Full Text | Google Scholar

Hoffner, C., and Cantor, J. (1991). “Perceiving and responding to mass media characters,” in Responding to the Screen: Reception and Reaction Processes, eds J. Bryant, and D. Zillmann (Mahwah, NJ: Erlbaum), 63–101.

Google Scholar

Hoffner, C. A., and Cohen, E. L. (2012). Responses to obsessive compulsive disorder on Monk among series fans: parasocial relations, presumed media influence, and behavioral outcomes. J. Broadcast. Electron. Media 56, 650–668. doi: 10.1080/08838151.2012.732136

CrossRef Full Text | Google Scholar

Hogan, T. P., and Brashers, D. E. (2009). “Implications of information behavior in the theory of communication and uncertainty management,” in Uncertainty and Information Regulation, eds W. Afifi and T. Afifi (New York, NY: Routledge), 45–66.

Horton, D., and Wohl, R. R. (1956). Mass communication and parasocial interaction: observations on intimacy at a distance. Psychiatry 19, 215–299. doi: 10.1080/00332747.1956.11023049

PubMed Abstract | CrossRef Full Text | Google Scholar

IMDB (2020). The Apprentice. IMDB. Available online at: https://www.imdb.com/title/tt0364782/?ref_=fn_al_tt_1 (accessed June 28, 2020).

Jin, J.-M., Bai, P., He, W., Wu, F., Liu, X.-F., Han, D.-M., et al. (2020). Gender differences in patients with COVID-19: focus on severity and mortality. Front. Public Health 8:152. doi: 10.3389/fpubh.2020.00152

PubMed Abstract | CrossRef Full Text | Google Scholar

Johnson, J. D., and Case, D. O. (2012). Health Information. Peter Lang.

Google Scholar

Kelly, S., Eldredge, S., Dalton, E. D., and Miller, L. (2014). Health information behavior: an initial validity portfolio for active and passive measures. Commun. Res. Rep. (New York, NY) 31, 171–182. doi: 10.1080/08824096.2014.907145

CrossRef Full Text | Google Scholar

Khoo, K., Bolt, P., Babl, F. E., Jury, S., and Goldman, R. D. (2008). Health information seeking by parents in the internet age. J. Paediatr. Child Health 44, 419–423. doi: 10.1111/j.1440-1754.2008.01322.x

PubMed Abstract | CrossRef Full Text | Google Scholar

Knapp, M. L. (1978). Social Intercourse: From Greeting to Goodbye. Boston, MA: Allyn and Bacon Incorporated.

Koffler, J. (2015, July 30). Donald Trump's 16 biggest business failures and successes. Time. Available online at: https://time.com/3988970/donald-trump-business/ (accessed June 28, 2020).

Kraak, V. I., and Story, M. (2015). Influence of food companies' brand mascots and entertainment companies' cartoon media characters on children's diet and health: a systematic review and research needs. Obes. Rev. 16, 107–126. doi: 10.1111/obr.12237

PubMed Abstract | CrossRef Full Text | Google Scholar

Labrecque, L. I. (2014). Fostering consumer-brand relationships in social media environments: the role of parasocial interaction. J. Interact. Market. 28, 134–148. doi: 10.1016/j.intmar.2013.12.003

CrossRef Full Text | Google Scholar

Ledbetter, A. M., and Redd, S. M. (2016). Celebrity credibility on social media: a conditional process of analysis of online self-disclosure attitude as a moderator of posting frequency and parasocial interaction. Commun. Mass Media Complete 80, 601–618. doi: 10.1080/10570314.2016.1187286

CrossRef Full Text | Google Scholar

Lee, E. (2013). Effectiveness of politicians' soft campaign on twitter versus TV: cognitive and experiential routes. J. Commun. 63, 953–974. doi: 10.1111/jcom.12049

CrossRef Full Text | Google Scholar

Maciosek, M. V., Coffield, A. B., Flottemesch, T. J., Edwards, N. M., and Solberg, L. I. (2010). Greater use of preventative services in U.S. health care could save lives at little or no cost. Health Affairs 29, 1656–1660. doi: 10.1377/hlthaff.2008.0701

PubMed Abstract | CrossRef Full Text | Google Scholar

McCroskey, J. C., and Richmond, V. P. (1996). Fundamentals of Human Communication: An Interpersonal Perspective. Prospect Heights, IL: Waveland Press.

McDonald, D. G., and Kim, H. (2001). When I die, I feel small: electronic game characters and the social self. J. Broadcast. Electron. Media 45, 241–258. doi: 10.1207/s15506878jobem4502_3

CrossRef Full Text | Google Scholar

McEwan, B. (2020). Sampling and validity. Ann. Int. Commun. Assoc. 44, 235–247. doi: 10.1080/23808985.2020.1792793

CrossRef Full Text | Google Scholar

Moe, P. (2012). Revealing rather than concealing disability: the rhetoric of Parkinson's advocate Michael J. Fox. Rhetoric Rev. 31, 443–460. doi: 10.1080/07350198.2012.711200

CrossRef Full Text | Google Scholar

Painter, M., and Qiu, T. (2020). Political Beliefs Affect Compliance With Covid-19 Social Distancing Orders. SSRN. Available online at: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3569098 (accessed June 28, 2020).

Google Scholar

Patwardhan, H., and Balasubramanian, S. K. (2011). Brand romance: a complementary approach to explain emotional attachment toward brands. J. Prod. Brand. Manag. 20, 297–308. doi: 10.1108/10610421111148315

CrossRef Full Text | Google Scholar

Perse, E. M., and Rubin, A. M. (1990). Chronic loneliness and television use. J. Broadcast. Electron. Media 34, 37–53. doi: 10.1080/08838159009386724

CrossRef Full Text | Google Scholar

Rice, R. E. (2006). Influences, usage, and outcomes of Internet health information searching: multivariate results from the Pew surveys. Int. J. Med. Inform. 75, 8–28. doi: 10.1016/j.ijmedinf.2005.07.032

PubMed Abstract | CrossRef Full Text | Google Scholar

Rubin, A. M., Perse, E. M., and Powell, R. A. (1985). Loneliness, parasocial interaction, and local television news viewing. Hum. Commun. Res. 12, 155–180. doi: 10.1111/j.1468-2958.1985.tb00071.x

CrossRef Full Text | Google Scholar

Rubin, A. M., and Step, M. M. (2000). Impact of motivation, attraction, and parasocial interaction on talk radio listening. J. Broadcast. Electron. Media 44, 635–654. doi: 10.1207/s15506878jobem4404_7

CrossRef Full Text | Google Scholar

Rubin, R. B., and Rubin, A. M. (2001). “Attribution in social and parasocial relationships,” in Attribution, Communication Behavior, and Close Relationships, eds V. Manusov, and J. H. Harvey (Cambridge: Cambridge University Press), 320–337.

Google Scholar

Secon, H. (2020, June 3). An interactive map of the US cities and states still under lockdown - and those that are reopening. Business Insider. Available online at: https://www.businessinsider.com/us-map-stay-at-home-orders-lockdowns-2020--s-2023 (accessed June 28, 2020).

Stever, G. (2013). Mediated vs. parasocial relationships: an attachment perspective. J. Media Psychol. 17, 1–31.

Google Scholar

Tian, Q., and Hoffner, C. A. (2010). Parasocial interaction with liked, neutral, and disliked characters on a popular TV series. Mass Commun. Soc. 13, 250–269. doi: 10.1080/15205430903296051

CrossRef Full Text | Google Scholar

Tian, Y., and Robinson, J. D. (2008). Incidental health information use and media complementarity: a comparison of senior and non-senior cancer patients. Patient Educ. Couns. 71, 340–344. doi: 10.1016/j.pec.2008.02.006

PubMed Abstract | CrossRef Full Text | Google Scholar

Tiger, R. (2013). Celebrity drug scandals, media double standards. Contexts 12, 36–41. doi: 10.1177/1536504213511214

CrossRef Full Text | Google Scholar

Tukachinsky, R., and Stever, G. (2019). Theorizing development of parasocial engagement. Commun. Theory 29, 297–318. doi: 10.1093/ct/qty032

CrossRef Full Text | Google Scholar

Turner, J. R. (1993). Interpersonal and psychological predictors of parasocial interaction with different television performers. Commun. Q. 41, 443–453. doi: 10.1080/01463379309369904

CrossRef Full Text | Google Scholar

Twitter. (2020, July 18). Donald J. Trump, @therealDonaldTrump. Twitter. Available online at: https://twitter.com/realDonaldTrump?ref_src=twsrc%5Egoogle%7Ctwcamp%5Eserp%7Ctwgr%5Eauthor (accessed June 28, 2020).

Williamson, J., and Hocken, B. (2010). Pancreatic cancer in the media: the Swayze shift. Ann. R. Coll. Surg. Engl. 92, 537–538. doi: 10.1308/003588410X12771863936044

PubMed Abstract | CrossRef Full Text | Google Scholar

Yamey, G., and Gonsalves, G. (2020). Donald Trump: a political determinant of COVID-19. Br. Med. J. 369:m1643. doi: 10.1136/bmj.m1643

PubMed Abstract | CrossRef Full Text | Google Scholar

Appendix

1. President Trump reminds me of someone who is competent and knows what he is doing.

2. ^*President Trump has the ability to deliver what he promises.

3. President Trump delivers what he promises.

4. President Trump has believable claims about COVID-19.

5. President Trump can be trusted.

^* Removed after CFA.