Effects of Uniqueness,News Valence,and Liking on Personalization of Company News

Abstract

Many online information systems are delivering personalized news to users today. The essence of this personalization process is to match a news article to the reader’s self-identity. However, prior studies mostly focus on matching a positive news story to a person’s desired identity. No known research has discussed the possibility of matching a negative news story to a person’s undesired identity. This study aims to fill this theoretical gap by testing a three-way interaction effect among news valence, identity desirability, and uniqueness on attitude toward the news story. Through a 2 × 2 × 2 between-subjects experiment, it is shown that a positive news story tends to generate a more favorable attitude when matched to the reader’s desired self-identity, whereas a negative news story will generate a more favorable attitude when matched to the person’s undesired self-identity, and such an effect is especially pronounced when the identity is unique. Perceived news credibility is found to partially mediate these effects.

Keywords

personalization personalized news self-identity news valence identity desirability identity uniqueness

Algorithmic recommendation systems are being widely used by search engines and social media platforms today to personalize news for users. People in general prefer personalized news stories in alignment with their preferences than standardized news articles picked by traditional media gatekeepers, but the main challenge of presenting such personalized news content to them is the ability to predict their particular news interests (Sela et al., 2015). If users’ preferences over different news content areas can be accurately assessed and modeled, it can help guide newspaper space allocation to enhance circulation and revenue (Kanuri et al., 2014). From this perspective, understanding news readers’ preferences is extremely important in developing effective personalized news recommendation systems.

The fundamental idea of news personalization process is to present “relevant” news articles to consumers and filter out “irrelevant” ones. A news article being relevant or irrelevant is primarily dependent on whether it matches the reader’s self-identity (Kalyanaraman & Sundar, 2006). Theoretically speaking, a person’s self-identity consists of both personal and social identities (Brewer, 1991). For example, a student at Harvard University may define himself or herself as a sophomore majoring in sociology (i.e., a personal identity), or as a member of Harvard chess club (i.e., a social identity). These two types of identities are fundamentally interrelated because personal identity is defined in part by social groups, and social identity is infused with personal meanings (Deaux, 1993). Therefore, the concept of “self” is a collection of identities that reflects various roles that a person occupies in the social structure (Stokburger-Sauer et al., 2012).

Prior studies have shown a variety of ways of using information related to a person’s self-identity to create a personalized message, such as name (Dijkstra, 2014; Li, 2016), personality schema (York et al., 2012), travel location preference (Kim & Sundar, 2012; Li et al., 2019), news interest (Lavie et al., 2010), and online browsing (van Doorn & Hoekstra, 2013). All these aspects of the “self” are essentially category labels that can invoke a person’s mental presentation of his or her identity (Oyserman, 2009; Reed et al., 2012). Noticeably, however, the existing literature of personalization mostly focuses on how to create a personalized message based on an individual’s desired identity (i.e., a category label that the person wants to be associated with) but neglects the possibility of his or her undesired identity (i.e., a category label that the person does not want to be associated with). According to the theory of possible selves, one’s self-concept involves both desired and undesired selves (Banister & Hogg, 2004; Dunkel & Kerpelman, 2006). People tend to approach their desired selves and avoid undesired ones. Based on this logic, a news consumer should form a positive attitude toward a news story not only when it “praises” something desirable, but also when it “criticizes” something undesirable. The former is a way for the person to approach his or her desired self, whereas the latter is a way for him or her to stay away from the undesired self. This research argues that personalized news stories can be either positive or negative; thus, matching a news story to either desired or undesired self-identity of the message recipient is possible and meaningful. As such, a complete theoretical framework of personalization should incorporate both approaches.

Using college sports rivalries (e.g., Alabama vs. Auburn, Duke vs. UNC, Michigan vs. Ohio State) as an example to illustrate this point as being a college sports team fan is a common way for people to construct their social identities, a student at Duke University may construct his or her identity in a way that “I like Duke but I do not like UNC” while a student at the University of North Carolina may have the opposite identity that “I like UNC but I do not like Duke.” Suppose that there are two news articles, one reporting “UNC failed to recruit the most talented high school basketball player this year due to a budget cut” and the other reporting “Duke failed to recruit the most talented high school basketball player this year due to a budget cut.” The first article “matches” the student at Duke and the second article “matches” the student at UNC; thus, both of them may be regarded as personalized news stories and generate favorable attitudes although the factual information is negative (i.e., budget cut, losing talent). Different from what was demonstrated in many prior personalization studies, this match is achieved by linking the message to an undesired part of a person’s self-identity instead of a desired part.

Another gap in the personalization literature is that it is ambiguous whether personalized information should always match a unique component of a person’s self-identity. As an individual can define his or her self-identity in many different ways, the identity can be somewhat common (e.g., “I like football and all my friends like it”) or unique (e.g., “I like table tennis but none of my friends likes it”). According to distinctiveness theory, uniqueness can increase the salience of one’s identity (Berger & Heath, 2007); thus, it may amplify the effect of personalization.

Summarizing the above reasoning, this study is designed to test how two identity-related attributes (i.e., desired vs. undesired; common vs. unique) may influence the effectiveness of personalized news. Using an online experiment, this study will test how people respond to positively or negatively framed news stories based on their desired or undesired and common or unique self-identities. The study findings contribute to the literature by presenting empirical evidence to explain why the current conceptualization of personalization is incomplete and needs to be refined.

Literature Review

News Personalization and Self-Identity

Before discussing the effect of personalized news, the basic concept of personalization needs to be clarified. In general, personalization refers to a communication process in which individualized messages are designed and delivered to message recipients based on their characteristics and preferences (Li & Kalyanaraman, 2013). Other terms such as customization and tailoring have also been used in prior research to carry similar meanings (Maslowska et al., 2013). Although some scholars argue there are conceptual differences among these terms (e.g., Beam, 2014; Sundar & Marathe, 2010), others do not differentiate them strictly (e.g., Benlian, 2015; Dijkstra, 2014; Maslowska et al., 2013). This study focuses on how news stories can be personalized for users, but it is worth pointing out that the idea of creating individualized information for separate message receivers can be applied to many other areas such as health communication, advertising, marketing, and so on.

A personalized message is a message that has something to do with one or more aspects of the message recipient’s self-identity. It may be a personal feature such as name (Dijkstra, 2014; Li, 2016), personality schema (i.e., responsible, adventurous, relationship-oriented, and inquisitive; York et al., 2012), or personality trait (i.e., extraversion, agreeableness, conscientiousness, neuroticism, and openness; Hirsh et al., 2012). It may also be a personal preference such as favorite news sources and topics (Beam & Kosicki, 2014), favorite singers (Ho & Tam, 2005), favorite actors/actresses and musical groups (Kalyanaraman & Sundar, 2006), favorite movies and college sports (Li & Kalyanaraman, 2013), and so on. In general, when a message matches a person’s self-identity, it tends to generate favorable attitudinal and behavioral outcomes (e.g., Kalyanaraman & Sundar, 2006; Noar et al., 2007; Petty et al., 2002). Such positive personalization effects have been found in various contexts such as advertisements (e.g., Hirsh et al., 2012; Wan, 2008), web portals (e.g., Kalyanaraman & Sundar, 2006), and health communication for cancer screening (e.g., Noar et al., 2007).

The principle of personalization is also widely adopted in the news industry. Nowadays, people have access to news messages that are specifically created based on their preferences (Beam, 2014; Lavie et al., 2010). Research has shown that personalized news portals can impact individuals’ news reading habits (Beam & Kosicki, 2014). News consumers tend to selectively expose themselves to and spend more time on attitude-consistent news stories than counter-attitudinal news stories (Knobloch-Westerwick & Meng, 2009). The personalization cues contained in news stories (i.e., personalization features that allow users to reveal their preferences) tend to increase news consumers’ perceived preference fit and enjoyment in news reading (Benlian, 2015). All these findings suggest that when a news story matches the reader’s self-identity, it tends to be considered as personalized and generates a favorable attitude because it is an identity confirmation and reinforcement.

Identity Desirability

Defined as “any category label to which a consumer self-associates either by choice or endowment,” identity is enacted through mental representations of labels as people incorporate them into their senses of selves (Reed et al., 2012, p. 312). Thanks to computer technologies, such an identity enactment process becomes more readily available on the internet through personalized communication. Essentially, the process of personalization entails a phase of “learning” the message recipient’s characteristics and preferences and a follow-up phase of “matching” the personalized information to those characteristics and preferences (Kalyanaraman & Sundar, 2006; Murthi & Sakar, 2003; Petty et al., 2002). An assumption undertaken by prior personalization studies, however, seems to be that the message receiver’s identity is positive and should be measured in a positive fashion. In the first phase of personalization—“learning” a person’s identity—his or her characteristics and preferences are typically measured by some positively phrased questions such as “Which destination do you wish to travel to the most?” A potential problem of this approach is that a person’s identity may not always be positive, and a personalized message may not necessarily be positive either. For example, if a person says his or her favorite travel destination is Paris, a news story about Paris will be regarded as personalized information and thus generate a favorable attitude, according to the current personalization theory. But what if the news story is about a terrorist attack in Paris? Conceptually speaking, this news story should be considered personalized, but it is likely to generate an unfavorable attitude (i.e., the person may still hold a favorable attitude toward Paris but his or her attitude toward this particular news story is likely to be negative).

News stories often carry a valence. Both positive and negative news are not atypical. So is a person’s identity. A person’s self-identity or self-concept involves a multiplicity of selves (Banister & Hogg, 2004), including desired and undesired selves (Hogg & Banister, 2001), or positive and negative selves (Dunkel & Kerpelman, 2006). These are two important dimensions of one’s possible selves, which can be defined as “individuals’ ideas of what they might become, what they would like to become, and what they are afraid of becoming” (Markus & Nurius, 1986, p. 954). Accordingly, people tend to approach their desired possible selves and avoid possible undesired selves (Markus & Nurius, 1986).

For self-identity construction purpose (i.e., answering the question, “Who am I?”), people often identify with certain companies and brands because those companies and brands offer attractive and meaningful social identities to help them satisfy their self-definitional needs (Bhattacharya & Sen, 2003; Stokburger-Sauer et al., 2012). This identification can explain many consumption behaviors, reflecting a person’s tendency to approach his or her desired selves and avoid undesired selves. In general, people are inclined to consume products or brands that have positive symbolic meanings, and reject products that evoke negative symbolic meanings (Banister & Hogg, 2004; Hogg & Banister, 2001). For example, people ride Harley-Davidson motorcycles to signal an identity of “biker,” which is associated with such values as personal freedom, patriotism and American heritage, and machismo (Schouten & McAlexander, 1995).

As a person’s identification with a company can reflect his or her self-identity, a possible way to measure his or her desired and undesired selves is to examine the meaning he or she attaches to this target company. If he or she likes the company, it suggests a positive meaning (i.e., desired); if he or she dislikes the company, it implies a negative meaning (i.e., undesired). In other words, a person’s preference position about a company (i.e., like vs. dislike) will manifest his or her self-identity. Based on this logic, a positive message about a liked company should produce a favorable attitude because it helps people approach their desired selves. A negative message about a disliked company should also generate a favorable attitude as it helps people avoid their undesired selves. As noted earlier, prior personalization studies seem to focus on matching a positive message to a person’s desired self-identity (e.g., Beam, 2014; Beam & Kosicki, 2014; Knobloch-Westerwick & Meng, 2009; Lavie et al., 2010), but neglects whether a negative message can be matched to a person’s undesired self-identity. The current study attempts to fill this theoretical gap by testing the following hypothesis:

H1: There will be a matching effect between the valence of a personalized news story and identity desirability on attitude toward the news story. A positive news story will generate a more favorable attitude when it is about a liked company, whereas a negative news story will generate a more favorable attitude when it is about a disliked company.

Both types of news stories that match news valence and identity desirability (i.e., a positive news story about a liked company and a negative news story about a disliked company) are expected to generate a favorable attitude toward the news story based on the above theoretical reasoning. However, it is unknown whether there will be a significant difference between these two types of news stories in terms of the attitudinal outcome. To fill this gap in the existing literature, a research question is proposed as follows:

RQ: Will a positive news story about a liked company generate a more favorable attitude toward the news story than a negative story about a disliked company?

Identity Uniqueness

In addition to identity desirability, the effect of personalization may also be influenced by identity uniqueness. According to the distinctiveness theory (McGuire, 1984; McGuire et al., 1978, 1979; McGuire & McGuire, 1981), when a person’s certain trait or identity is the minority in relation to others in the immediate environment, it will become more salient through the numerical distinctiveness. The identity salience principle posits that the more salient a particular identity is, the larger influence it will have on a person’s subsequent attitude and behavior (Reed et al., 2012).

Empirical evidence in the literature supports the distinctiveness theory and identity salience principle. In McGuire et al.’s (1978) initial examination, it was found that members of minority groups were more aware of their ethnicity in an ethnically mixed group as opposed to members of majority groups. In a similar vein, Cota and Dion (1986) found that female participants who were minorities in terms of gender composition in a group were more conscious of their gender identity. Furthermore, it was shown that targeted advertisements were more effective when the targeted group was socially distinctive (Forehand et al., 2002; Grier & Deshpande, 2001). The salience of identity would lead to a higher level of bond and trust in the targeted information, thereby increasing people’s favorability toward the information (Aaker et al., 2000; Deshpandé & Stayman, 1994; Forehand & Deshpande, 2001).

Given that the essence of personalization is to match a message to a person’s identity, uniqueness of the identity is expected to affect the persuasion effect of the message. Because a unique identity is more salient than a common identity, it is likely to amplify personalization effects. Several prior studies have discussed the impact of uniqueness on personalization and drawn the conclusion that uniqueness can at least partially explain why personalized messages generate more favorable attitudes than nonpersonalized messages (e.g., Franke & Schreier, 2008; Merle et al., 2008; Park et al., 2013). Following this logic, no matter whether a personalized news story helps the reader approach his or her desired self-identity or avoid undesired self-identity, the effect will be more pronounced when the identity is unique because uniqueness enhances the signaling of this identity (Berger & Heath, 2007). Therefore, it is hypothesized:

H2: The matching effect between news valence and identity desirability on attitude toward the news story described in H1 will be moderated by identity uniqueness, such that the matching effect is more pronounced when the identity is unique.

Self-Confirmation Bias and Perceived Message Credibility

A personalized news story aims to match the reader’s self-identity. When this match occurs, the reader’s identity gets confirmed and reinforced (Kalyanaraman & Sundar, 2006). Consequently, the person tends to spend more time reading the news story and forms a more positive attitude toward it (Garrett, 2009; Knobloch-Westerwick & Meng, 2009). An important question here is, “Why does confirmation of one’s identity lead to these favorable consequences?” The underlying psychological mechanism may be people’s self-confirmation bias.

Self-confirmation bias refers to the tendency to place greater weight on information that is consistent with preexisting beliefs, while undervaluing the information that does not align with those beliefs (Klayman & Ha, 1987; Metzger & Flanagin, 2013). Previous studies have shown that people tend to evaluate a message as credible as long as it confirms their preexisting beliefs, even when the information is not based on sound logic and/or is not appropriately sourced (Evans et al., 1983; Fischer et al., 2005; Metzger et al., 2010; Metzger & Flanagin, 2013). When a message is deemed credible, it leads to a more favorable attitude and a stronger behavioral change (Appelman & Sundar, 2016; Pornpitakpan, 2004).

Such biased information processing is in accordance with the notion of confirming and reinforcing self-identity—either approaching desired selves or avoiding undesired selves (Dunkel & Kerpelman, 2006). When one’s self-identity gets confirmed by a message, the person is likely to process the message in a biased way and judges it to be highly credible, and subsequently forms a favorable attitude toward it (Klayman & Ha, 1987; Metzger & Flanagin, 2013). In other words, perceived message credibility will mediate the effect of personalization when a personalized news story confirms the message receiver’s self-identity. As discussed earlier, this personalized news story can be either a positive news story about a commonly or uniquely liked company or a negative news story about a commonly or uniquely disliked company. Thus, the last hypothesis is proposed:

H3: Perceived news credibility will mediate the effect of news valence, identity desirability, and identity uniqueness on attitude toward the news story.

A conceptual model summarizing all three hypotheses is drawn in Figure 1.

Figure 1.

The conceptual model of this study.

Method

Study Overview

This study was designed as a 2 × 2 × 2 between-subjects experiment. The three factors included news valence (positive vs. negative), identity desirability (desired vs. undesired), and identity uniqueness (common vs. unique). The dependent variable was attitude toward the news story. The mediator was perceived news credibility.

Study Participants

A total of 206 adult participants residing in the U.S. were recruited through Survey Sampling International (SSI), a market research company. These participants voluntarily opted in to participate in this experiment and received small financial incentives for completing the study. In recruiting these participants through SSI, a gender- and age-balanced sample was requested. As a result, the mean age of the participants was 43.82 (SD = 16.42), and 52.4% were male. The majority of them were non-Hispanic Caucasians (78.2%), followed by Hispanics (8.3%), African Americans (5.8%), Asian Americans (5.8%), and others (1.9%). Forty-eight percent of the participants had obtained either a college or more advanced degree. Regarding their occupations, more than half of the participants were working professionals (58.8%), followed by retirees (16.0%), house makers (9.7%), currently unemployed (8.7%), and full-time students (6.8%).

Experimental Stimuli

Because this study aimed to test three hypotheses and a research question about the effects of different types of company news, a news template was created for the experiment based on a real Forbes news article. The news story was about a recently published company ranking according to the data collected from thousands of employee surveys. The basic content (370 words) and layout of this news story were identical in all experimental conditions, but the news title and a couple of keywords in the article were manipulated to differ across conditions. For the positive news conditions, the news title was “Most Desirable Place to Work,” and the article said this newly released ranking included 50 most desirable companies to work for and explained why “xxx” company ranked number one on the list (as described in the following space, this “xxx” company was a randomly selected company from the short list of companies the participant had listed from previous questions to reflect his or her self-identity). In contrast, for the negative news conditions, the news focused on “xxx” company that was ranked number one on a recently released list of “Least Desirable Place to Work” (again, this “xxx” company was a randomly selected company that reflected the participant’s self-identity). The positive news and negative news only differed in two places in the text (i.e., “top” companies vs. “bottom” companies; most desirable place vs. least desirable place).

As a major objective of this study was to examine the effect of matching news stories (either positive or negative) to readers’ self-identities (either desired or undesired, and either common or unique), measuring participants’ self-identities from both desirability and uniqueness perspectives was necessary. For this purpose, each participant was asked to respond to the following four questions: (a) Name one company that you think most people like but you dislike (unique undesired identity), (b) Name one company that you think most people like and you also like (common desired identity), (c) Name one company that you think most people dislike but you like (unique desired identity), and (d) Name one company that you think most people dislike and you also dislike (common undesired identity). These measures reflected the operational definitions of identity desirability (i.e., whether a person attaches positive or negative meanings to an object) and identity uniqueness (i.e., whether a person shares an identity with others). As each participant listed four companies in his or her responses to these questions, one company was randomly selected from the short list of these four companies he or she had listed from the questions and incorporated into the stimulus news story for him or her.

Study Measures

Attitude toward the news article

Participants’ attitudes toward the news article were measured by four pairs of adjectives on a 7-point semantic differential scale, including bad/good, dislike/like, unfavorable/favorable, and boring/interesting with 1 representing bad, dislike, unfavorable, and boring (adapted from MacKenzie et al., 1986, α = .95).

Attitude toward the company

Participants’ preexisting attitudes toward each of the four companies that they named in response to the identity-related questions were measured by three pairs of adjectives on a 7-point semantic differential scale, including dislike/like, unfavorable/favorable, and negative/positive with 1 representing dislike, unfavorable, and negative (adapted from Nan & Heo, 2007, α ranged from .95 to .98).

Perceived credibility of the news article

Perceived news credibility was measured with two statements on a 7-point Likert-type scale (1 = strongly disagree and 7 = strongly agree), including “I find this article convincing” and “I find this article believable” (adapted from Rifon et al., 2004, α = .85; the third statement, “I find this article unbiased” in the original measure was dropped as it significantly lowered the measurement reliability).

Manipulation check of perceived news article valence

Participants’ perceived news valence was measured with one statement: “The article is written in a negative tone” on a 7-point Likert-type scale (1 = strongly disagree and 7 = strongly agree).

Experimental Procedure

The experiment was conducted online via Qualtrics. Each participant was provided with a hyperlink to the experimental stimulus and questionnaire. After providing their consent to take part in the study, the participants were asked to name four companies that reflected their undesired/desired and common/unique self-identities in a prequestionnaire. Their attitudes toward each of these four companies were measured afterward. Their demographic information was also collected in this prequestionnaire, including gender, age, ethnicity, educational background, and occupation. Then, each participant was randomly assigned to one of the eight experimental conditions, with the stimulus news article being either positive or negative, and with the company mentioned in the article reflecting either desired or undesired, and common or unique self-identity. After reading the news article, participants’ attitudes toward the article were measured in a postquestionnaire. Their perceived news credibility and news valence were also measured in the postquestionnaire. Participants’ viewing of stimulus materials and responding to questions were self-paced. It took approximately 10 to 15 minutes to complete the experiment.

Results

Manipulation Checks

An independent-samples t test showed that the participants in the negative news conditions (M = 4.14) perceived the news article to be written in a significantly more negative tone than those in the positive news conditions (M = 2.84), t (204) = 5.86, p < .001, confirming the success of news valence manipulation.

The participants reported their attitudes toward four identity-related companies in the prequestionnaire. To test whether their attitudes toward the desired-identity-related companies were more favorable than the undesired-identity-related companies, a repeated measures analysis of variance (ANOVA) test was conducted. The results confirmed that participants’ attitudes toward both common (M = 6.38) and unique (M = 5.77) desired-identity-related companies were significantly more favorable than toward both common (M = 2.03) and unique (M = 2.24) undesired-identity-related companies, based on pairwise comparisons using the Bonferroni method (all p values < .001).

Tests of Hypotheses

To test H1 and H2, a three-way ANOVA test was performed with news article valence (negative news = −1; positive news = 1), identity desirability (undesired identity = −1; desired identity = 1), and identity uniqueness (common identity = −1; unique identity = 1) being the three fixed factors and attitude toward the news article being the dependent variable. A significant two-way interaction effect was found between news article valence and identity desirability, F(1, 198) = 10.48, p < .001, η² = .05, supporting H1. This interaction effect is visually displayed in Figure 2. Post hoc analysis showed that a positive news story matched to a desired identity (M = 5.04) generated a significantly more favorable attitude than a negative news story matched to a desired identity (M = 3.87), p = .003, whereas a positive news story matched to an undesired identity (M = 3.98) did not significantly differ from a negative news story matched to an undesired identity (M = 4.35), p = .688. To address RQ, attitude toward a positive news story matched to a desired identity (M = 5.04) was compared with attitude toward a negative news story matched to an undesired identity (M = 4.35); the difference was found to be not significant, p = .195.

Figure 2.

Two-way interaction between news article valence and identity desirability on attitude.

Furthermore, there was a significant three-way interaction effect among news article valence, identity desirability, and identity uniqueness on attitude, F(1, 198) = 4.06, p = .045, η² = .02, supporting H2. This interaction was depicted in Figure 3. Post hoc analysis showed that when identity was unique, a positive news story matched to a desired identity generated a significantly more favorable attitude (M = 5.02) than a negative news story matched to a desired identity (M = 3.70), p = .022, but the difference between positive (M = 3.67) and negative (M = 4.84) news stories matched to an undesired identity was not significant, p = .073. When identity was common, positive and negative news stories did not significantly differ from each other in attitude (all p values > .05). It was also found that attitude toward a positive news story matched to a desired identity did not significantly differ from attitude toward a negative news story matched to an undesired identity, no matter whether the identity was common or unique (all p values > .05).

Figure 3.

Three-way interaction between news article valence, identity desirability, and identity uniqueness on attitude.

To test H3, a mediation analysis was conducted by using Model 11 of the PROCESS macro in SPSS (Hayes, 2013). The significance of the mediating effect was judged based on the 95th percentile confidence intervals (bootstrap = 5,000 samples). The independent variable was news article valence (negative news = −1; positive news = 1) and the dependent variable was attitude toward the news article. The mediator was perceived news credibility. The primary moderator was identity desirability (undesired identity = −1; desired identity = 1) and the secondary moderator was identity uniqueness (common identity = −1; unique identity = 1). The detailed analysis results are presented in Tables 1 and 2. As seen in Table 1, a significant interaction effect between news valence and identity desirability was found on perceived news credibility (coefficient = .33; 95% confidence interval [CI] = [.12, .54]; t = 3.09, p = .002). The direct effect of news valence on attitude was not significant (coefficient = .11; 95% CI = [−.07, .29]; t = 1.18, p = .238), but that of perceived news credibility on attitude was significant (coefficient = .74; 95% CI = [.62, .85]; t = 12.61, p < .001). How news valence indirectly affected attitude through perceived news credibility was examined. As shown in Table 2, this mediation effect was significant in three conditions (common desired identity: coefficient = .34; 95% CI = [.05, .66]; unique desired identity: coefficient = .32; 95% CI = [.02, .65]; unique undesired identity: coefficient = −.32; 95% CI = [−.66, −.01]) because the confidence intervals did not contain zero. It was not significant in the condition of common undesired identity (coefficient = .00; 95% CI = [−.32, .32]) as the confidence interval included zero. The index of moderated mediation, as explained in Hayes (2018), was not significant (coefficient = .08; 95% CI = [−.08, .24]) as the confidence interval contained zero. The conditional moderated mediation by identity desirability was significant in the condition of unique identity (coefficient = .32; 95% CI = [.09, .56]), but not significant in the condition of common identity (coefficient = .17; 95% CI = [−.05, .40]). That was to say, identity desirability significantly moderated the indirect effect of news valence on attitude only when identity was unique. This moderated mediation effect was not significant when identity was common. Based on these results, it was concluded that H3 was partially supported.

Table 1.

Regression Results for Variables Predicting Perceived News Credibility and Attitude Toward the News Article.

Variable	B	t	p	95% CI (LL)	95% CI (UL)
Credibility regressed on
Valence	.12	1.11	.267	−.09	.33
Desirability	.00	0.01	.992	−.21	.21
Uniqueness	−.02	−0.23	.821	−.24	.19
Valence × Desirability	.33	3.09	.002	.12	.54
Valence × Uniqueness	−.12	−1.07	.284	−.33	.10
Desirability × Uniqueness	−.13	−1.19	.235	−.34	.08
Valence × Desirability × Uniqueness	.10	0.97	.336	−.11	.32
Attitude regressed on
Valence	.11	1.18	.238	−.07	.29
Credibility	.74	12.61	.000	.62	.85

Note. CI = confidence interval; LL = lower limit; UL = upper limit.

Table 2.

Conditional Indirect Effect of News Article Valence (Negative News = −1; Positive News = 1) on Attitude Toward the News Article (Mediator = Perceived News Credibility).

Type of effect	Identity desirability	Identity uniqueness	Effect (index)	SE	95% CI (LL)	95% CI (UL)
Indirect effect of valence on attitude	Undesired	Common	.00	.16	−.32	.32
	Undesired	Unique	−.32	.17	−.66	−.01
	Desired	Common	.34	.16	.05	.66
	Desired	Unique	.32	.16	.02	.65
Moderated mediation			.08	.08	−.08	.24
Conditional moderated mediation by desirability		Common	.17	.11	−.05	.40
Conditional moderated mediation by desirability		Unique	.32	.12	.09	.56

Note. CI = confidence interval; LL = lower limit; UL = upper limit.

Discussion

With the increasing popularity of personalized news recommendation systems, people are frequently exposed to information that aligns with their interests and preferences. This study examined the effect of personalized news based on identity theories. Specifically, it investigated how the two attributes of one’s self-identity, as being desired or undesired, and common or unique, impacted news consumers’ attitudes toward positively or negatively framed news articles. In addition, the mediation role of perceived news credibility was tested.

Desired and Undesired Selves

Striving for a sense of “self” is a fundamental human action (Stokburger-Sauer et al., 2012). This concept of “self” is important to personalization because the essence of personalizing a message is to create a match between the message and one or more aspects of the message receiver’s self-identity (Kalyanaraman & Sundar, 2006; Petty et al., 2002). Although prior studies have tested the effect of personalization from various perspectives, they mostly focused on a desired aspect of the message receiver’s self-identity but neglected a possible undesired aspect. Conceptually speaking, one’s identity consists not only desired selves but also undesired selves (Hogg & Banister, 2001). Thus, both matching a positive message to a person’s desired self-identity and matching a negative message to a person’s undesired self-identity should be considered personalization and generate a favorable attitude.

This study empirically tested this proposition, using positively and negatively framed company news stories as the experimental stimuli. One of the ways people often use to construct their identities is to identify with a company, and this is “an active, selective, and volitional act motivated by the satisfaction of one or more self-definitional (i.e., ‘Who am I?’) needs” (Bhattacharya & Sen, 2003, p. 77). Reflecting their identification positions with different companies (i.e., like vs. dislike), people tend to consume products or brands that have positive meanings and avoid those with negative meanings (Banister & Hogg, 2004; Hogg & Banister, 2001). Based on this logic, it was hypothesized that a positive new story about a liked company would be more favorable than that about a disliked company, whereas a negative news story about a disliked company would be more pleasant than that about a liked company. This hypothesis was supported by the results reported in this study, suggesting that people tend to approach their desired identities and avoid their undesired identities.

Another interesting finding of this study was that a negative news story about a disliked company yields a favorable attitude similar to that of a positive news story about a liked company (i.e., the difference between the two was not statistically significant). This finding demonstrated the importance of undesired selves in constructing a person’s identity. In addressing the “Who am I?” question, people pay attention to not only what they want to become but also what they do not wish to become (Markus & Nurius, 1986). Given that personalization has been widely discussed in the literature and is considered a premier communication strategy (Li & Kalyanaraman, 2013), an accurate and complete conceptualization of personalization is much needed. Based on the findings of this study, it is suggested that personalization should be defined from two perspectives, including matching a message to either desired or undesired self-identity of the message recipient.

It is worth pointing out here that a person can define his or her self-identity in many different ways. A person’s identification with a certain company is a component of his or her social identity, but apparently this identification is not the only way that the person can use to define himself or herself. It is also worth noting that the effect of personalized news was measured by attitude only in the current study. In theory, liking a news story (i.e., favorable attitude) and perceiving it to be useful (i.e., high utility value) are two different concepts. When people perceive a news article to contain high informational utility, they tend to share the news story and retain more information from it (Bobkowski, 2015; Carpentier, 2008), but this does not necessarily mean that they like the news story. It is not wise to build personalized news recommendation systems based on the sole objective of giving news readers what they want rather than what they need, as this may potentially jeopardize social functions of journalism such as informing the public and bridging communities (Lee & Tandoc, 2017). From this perspective, measuring the effect of personalized news should include not only attitude but also other outcomes.

Common and Unique Selves

Besides identity desirability, this study also examined the impact of identity uniqueness on personalization. Although previous studies somewhat agreed on the “matching” process of personalization, they were inconsistent regarding whether a personalized message should always match a unique aspect of the message recipient’s self-identity. For example, Beam (2014) defined personalization as a process in which information is modified to align with a user’s preferences but did not specify whether those preferences need to be unique. Similarly, several personalization studies (e.g., Arora et al., 2008; Dijkstra, 2014; Hawkins et al., 2008; Hirsh et al., 2012; Ho & Tam, 2005; Tam & Ho, 2006; York et al., 2012) focused on the congruence between a personalized message and the message recipient without specifying commonness/uniqueness of this congruence. In contrast, some other prior studies incorporated the concept of uniqueness or distinctiveness when examining the effect of personalization (e.g., Kalyanaraman & Sundar, 2006; Kim & Sundar, 2012; Kreuter et al., 2000).

While whether uniqueness should be included in the definition of personalization is certainly worth a debate, what the current experiment showed was that identity uniqueness significantly affected the effect of personalization. Specifically, a significant three-way interaction effect was found among news valence, identity desirability, and identity uniqueness on attitude. When the news valence “matched” identity desirability, no matter whether it was a positive match or a negative match, a favorable attitudinal effect tended to occur when the identity was unique.

These findings indicate the importance of constructing personalized messages based on people’s unique self-identities instead of common self-identities. As argued by Bhattacharya and Sen (2003), people tend to be attracted to companies that help satisfy their self-definitional needs such as self-distinctiveness. In other words, people are likely to identify with a company that has distinct and unique traits that they value. When a person defines himself or herself in a self-inclusive category such as a social group, he or she perceptually recognizes differences between in-group and out-group and similarities among in-group members, and behaviorally favors in-group over out-group (Terry et al., 1999). A unique identity is valuable for a person, likely because it facilitates this categorization process.

It is worth pointing out that whether an identity is common or unique is quite subjective because it is at the discretion of the person. Whether a person perceives his or her identity to be common or unique depends on how he or she compares the identity with the immediate environment. Future research needs to further examine how subjectivity of identity uniqueness may affect the effectiveness of personalized information.

Self-Confirmation Bias and Perceived Message Credibility

Most previous personalization studies argued that perceived message relevance is the mediator of personalization effects (Noar et al., 2007). This study offered a different perspective by testing the mediation effect of perceived message credibility based on the theory of self-confirmation bias. When a message meets the message receiver’s need for approaching his or her desired self-identity or avoiding his or her undesired self-identity, this person’s self-identity gets confirmed or reinforced. Thus, he or she may place more weight on this message and forms a more positive attitude toward it (Evans et al., 1983; Fischer et al., 2005; Metzger et al., 2010; Metzger & Flanagin, 2013). This proposition was tested in the current experiment and gained partial support. Perceived news credibility was found to mediate the effect of news valence on attitude toward the news article in the condition of unique identity, but not in that of common identity. A positive news story matched to a unique desired identity or a negative news story matched to a unique undesired identity tends to be perceived as more credible, and consequently generates a more favorable attitude. A potential explanation for these results is that people value information in relation to their unique identities more than that associated with their common identities; thus, they are likely to have more exposure to such information in everyday life. Due to frequent selective exposure, they may feel a positive news story about their unique desired identity or a negative news story about their unique undesired identity to be natural, familiar, and believable.

Practical Implications

The findings of this study provide practical implications for media professionals in the news industry. News consumers are exposed to an enormous amount of information on a daily basis. Personalized news recommendation systems help people save time and energy because news stories that match their preferences and interests can be automatically recommended and displayed. It is worth noting that very few news consumers will spend much time and effort updating their online profiles manually (Liu et al., 2010); thus, most personalized news systems will track users’ online behaviors (e.g., history of clicking news) and use algorithms to recommend news stories that match those behaviors. Based on the results of this study, such algorithms need to take news valence or tone into consideration. Presenting a news article (whether positive or negative) based on a user’s past online behaviors may be a match or mismatch, depending on whether those behaviors represent the person’s desired or undesired identity. Apparently, making an inference about a person’s desired or undesired identity based on his or her online clicking behaviors is a challenging task. A machine learning program in this regard needs to take both news content valence (e.g., the tone of a news story being automatically coded based on negative or positive words used) and user feedback (e.g., the star rating a user gives to a news article being automatically recorded) into consideration in the calculation. Zihayat et al. (2019) offered such an example in building personalized news recommendation systems, in which both article-related attributes (e.g., popularity of the news article) and user-article interaction attributes (e.g., readers’ social activities including “liking” the news article and “sharing” the news article) are incorporated in the data modeling.

From the perspective of news consumers, it is fairly easy for them to offer feedback to news articles on social media such as “liking,” “sharing,” and “commenting.” Such publicly visible audience feedback is becoming a powerful force to influence the news production and dissemination process, including news topic selection, news story placement, and even journalists’ performance evaluations (Lee & Tandoc, 2017). How to balance the need to provide news consumers with what they want to read (e.g., news stories consistent with their preexisting beliefs) and what they need to read (e.g., news stories challenging their preexisting beliefs) is a critical issue to consider in building effective and beneficial news recommendation systems. As an example, Sela et al. (2015) argued that online news systems should include both personalized news stories and general news stories. Although users may not explicitly express preferences for general and standardized news stories, those stories may still interest them and be beneficial to read.

Limitations and Future Research

Despite the theoretical and practical implications of this study, there are several limitations and future research directions that need to be addressed. First, this study suggested that personalization could be realized by either matching a positive message to a person’s desired self-identity or matching a negative message to a person’s undesired self-identity. However, whether a neutral message can be matched to a person’s desired or undesired identity is unclear. Moreover, whether a single personalized message can reflect both the message recipient’s desired and undesired self-identities is unknown. These theoretical questions deserve more attention in future research.

Second, this study used attitude toward the news story as the dependent variable, but neglected other attitudinal outcomes. An interesting question waiting for future research to examine is, “When a news story mismatches the reader’s identity, will it result in not only the person’s negative attitude toward the news story, but also a negative feeling toward the online system that automatically generates the news story?”

Third, by asking them to name four companies related to their identities in a questionnaire, this experiment measured the participants’ desired and undesired identities from a very narrowly defined angle. Although it is reasonable to assume that these companies reflect a part of the participants’ self-identities, how important or relevant they are to the participants is unknown. Future research may replicate the experimental design of the current study but use another method to measure people’s desired and undesired self-identities, examining whether identity importance/relevance will influence the study results.

On a related note, the participants in this experiment were asked to identify their desirable and undesirable companies before exposure to the stimulus news article. It was possible that they realized the company in the article was one they provided in the questionnaire; thus, the personalization effect observed in this study might be stronger than what it would be in a natural setting. In other words, there might be a possibility of “demand effects” that participants picked up clues in the task and behaved accordingly (Zizzo, 2010). Although this possibility is concerning, in a recent study, Mummolo and Peterson (2019) showed that participants are generally unlikely to aid researchers in confirming their hypotheses even if the purpose of the experiment is revealed to them.

It is unclear whether people retrieve information related to their desired identity and undesired identity in a similar way. If one type of identity information is more accessible than the other, whether it will positively affect the effect of a message based on this information needs to be tested in future research. Also, some prior research argued that a common self-identity represents a person’s collectivist tendency and a unique self-identity represents his or her individualist tendency (Li & Kalyanaraman, 2013); thus, a person’s cultural orientation may determine how he or she perceives his or her identity. Whether the culture factor may influence the interaction effects presented in this experiment needs future studies to explore.

Finally, both positive and negative news stories tested in this experiment were considered generic news (i.e., news produced by an impartial source). However, what if a news article is created by a company to attack its competitor? Such a news article is likely to be perceived as biased. Future research may examine how this type of negative news influences the effect of personalization. Furthermore, future studies should test the possibility of replicating the findings of this experiment by using other types of stimulus messages instead of news stories (e.g., positive vs. negative health-related messages).

Footnotes

Declaration of Conflicting Interests

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

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Cong Li

References

Aaker

Brumbaugh

Grier

S. A.

(2000). Nontarget market effects and viewer distinctiveness: The impact of target marketing on attitudes. Journal of Consumer Psychology, 9, 127–140. https://doi.org/10.1207/S15327663JCP0903_1

Appelman

Sundar

(2016). Measuring message credibility: Construction and validation of an exclusive scale. Journalism & Mass Communication Quarterly, 93, 59–79. https://doi.org/10.1177/1077699015606057

Arora

Dreze

Ghose

Hess

J. D.

Iyengar

Jing

… John Zhang

(2008). Putting one-to-one marketing to work: Personalization, customization, and choice. Marketing Letters, 19, 305–321. https://doi.org/10.1007/s11002-008-9056-z

Banister

E. N.

Hogg

M. K.

(2004). Negative symbolic consumption and consumers’ drive for self-esteem: The case of the fashion industry. European Journal of Marketing, 38, 850–868. https://doi.org/10.1108/03090560410539285

Beam

M. A.

(2014). Automating the news: How personalized news recommender system design choices impact news reception. Communication Research, 41, 1019–1041. https://doi.org/10.1177/0093650213497979

Beam

M. A.

Kosicki

G. M.

(2014). Personalized news portals: Filtering systems and increased news exposure. Journalism & Mass Communication Quarterly, 91, 59–77. https://doi.org/10.1177/1077699013514411

Benlian

(2015). Web personalization cues and their differential effects on user assessments of website value. Journal of Management Information System, 32, 225–260. https://doi.org/10.1080/07421222/2015.1029394

Berger

Heath

(2007). Where consumers diverge from others: Identity signaling and product domains. Journal of Consumer Research, 34, 121–134. https://doi.org/10.1086/519142

Bhattacharya

C. B.

Sen

(2003). Consumer–company identification: A framework for understanding consumers’ relationships with companies. Journal of Marketing, 67, 76–88. https://doi.org/10.1509/jmkg.67.2.76.18609

10.

Bobkowski

P. S.

(2015). Sharing the news: Effects of informational utility and opinion leadership on online news sharing. Journalism & Mass Communication Quarterly, 92, 320–345. https://doi.org/10.1177/1077699015573194

11.

Brewer

M. B.

(1991). The social self: On being the same and different at the same time. Personality and Social Psychology Bulletin, 17, 475–482. https://doi.org/10.1177/0146167291175001

12.

Carpentier

F. R. D.

(2008). Applicability of the informational utility model for radio news. Journalism & Mass Communication Quarterly, 85, 577–590. https://doi.org/10.1177/107769900808500306

13.

Cota

A. A.

Dion

K. L.

(1986). Salience of gender and sex composition of ad hoc groups: An experimental test of distinctiveness theory. Journal of Personality and Social Psychology, 50, 770–776. https://doi.org/10.1037/0022-3514.50.4.770

14.

Deaux

(1993). Reconstructing social identity. Personality and Social Psychology Bulletin, 19, 4–12. https://doi.org/10.1177/0146167293191001

15.

Deshpandé

Stayman

D. M.

(1994). A tale of two cities: Distinctiveness theory and advertising effectiveness. Journal of Marketing Research, 31, 57–64. https://doi.org/10.2307/3151946

16.

Dijkstra

(2014). The persuasive effects of personalization through: Name mentioning in a smoking cessation message. User Modeling and User-Adapted Interaction, 24, 393–411. https://doi.org/10.1007/s11257-014-9147-x

17.

Dunkel

Kerpelman

(2006). Possible selves: Theory, research and applications. Nova Science Publishers.

18.

Evans

J. S. B. T.

Barston

J. L.

Pollard

(1983). On the conflict between logic and belief in syllogistic reasoning. Memory &Cognition, 11, 295–306. https://doi.org/10.3758/BF03196976

19.

Fischer

Jonas

Frey

Schulz-Hardt

(2005). Selective exposure to information: The impact of information limits. European Journal of Social Psychology, 35, 469–492. https://doi.org/10.1002/ejsp.264

20.

Forehand

Deshpande

(2001). What we see makes us who we are: Priming ethnic self-awareness and advertising response. Journal of Marketing Research, 38, 336–348. https://doi.org/10.1509/jmkr.38.3.336.18871

21.

Forehand

M. R.

Deshpande

Reed

II. (2002). Identity salience and the influence of differential activation of the social self-schema on advertising response. Journal of Applied Psychology, 87, 1086–1099. https://doi.org/10.1037/0021-9010.87.6.1086

22.

Franke

Schreier

(2008). Product uniqueness as a driver of customer utility in mass customization. Marketing Letters, 19, 93–107. https://doi.org/10.1007/s11002-007-9029-7

23.

Garrett

R. K.

(2009). Echo chambers online?: Politically motivated selective exposure among Internet news users. Journal of Computer-Mediated Communication, 14, 265–285. https://doi.org/10.1111/j.1083-6101.2009.01440.x

24.

Grier

S. A.

Deshpande

(2001). Social dimensions of consumer distinctiveness: The influence of group social status and identity on advertising persuasion. Journal of Marketing Research, 38, 216–224. https://doi.org/10.1509/jmkr.38.2.216.18843

25.

Hawkins

R. P.

Kreuter

Resnicow

Fishbein

Dijkstra

(2008). Understanding tailoring in communication about health. Health Education Research, 23, 454–466. https://doi.org/10.1093/her/cyn004

26.

Hayes

A. F.

(2013). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. Guilford Press.

27.

Hayes

A. F.

(2018). Partial, conditional, and moderated moderated mediation: Quantification, inference, and interpretation. Communication Monographs, 85, 4–40. https://doi.org/10.1080/03637751.2017.1352100

28.

Hirsh

J. B.

Kang

S. K.

Bodenhausen

G. V.

(2012). Personalized persuasion: Tailoring persuasive appeals to recipients’ personality traits. Psychological Science, 23, 578–581. https://doi.org/10.1177/0956797611436349

29.

S. Y.

Tam

K. Y.

(2005). An empirical examination of the effects of web personalization at different stages of decision making. International Journal of Human-Computer Interaction, 19, 95–112. https://doi.org/10.1207/s15327590ijhc1901_7

30.

Hogg

M. K.

Banister

E. N.

(2001). Dislikes, distastes and the undesired self: Conceptualising and exploring the role of the undesired end state in consumer experience. Journal of Marketing Management, 17, 73–104. https://doi.org/10.1362/0267257012571447

31.

Kalyanaraman

Sundar

S. S.

(2006). The psychological appeal of personalized content in web portals: Does customization affect attitudes and behavior? Journal of Communication, 56, 110–132. https://doi.org/10.1111/j.1460-2466.2006.00006.x

32.

Kanuri

V. K.

Thorson

Mantrala

M. K.

(2014). Using reader preferences to optimize news content: A method and a case study. International Journal on Media Management, 16, 55–75. https://doi.org/10.1080/14241277.2014.943898

33.

Kim

N. Y.

Sundar

S. S.

(2012). Personal relevance versus contextual relevance: The role of relevant ads in personalized websites. Journal of Media Psychology, 24, 89–101. https://doi.org/10.1027/1864-1105/a000067

34.

Klayman

Y. W.

(1987). Confirmation, disconfirmation and information in hypothesis testing. Psychological Review, 94, 211–228. https://doi.org/10.1037/0033-295X.94.2.211

35.

Knobloch-Westerwick

Meng

(2009). Looking the other way: Selective exposure to attitude-consistent and counterattitudinal political information. Communication Research, 36, 426–448. https://doi.org/10.1177/0093650209333030

36.

Kreuter

M. W.

Farrell

Olevitch

Brennan

(2000). Tailoring health messages: Customizing communication with computer technology. Erlbaum.

37.

Lavie

Sela

Oppenheim

Inbar

Meyer

(2010). User attitudes towards news content personalization. International Journal of Human-Computer Studies, 68, 483–495. https://doi.org/10.1016/j.ijhcs.2009.09.011

38.

Lee

E.-J.

Tandoc

E. C.

Jr (2017). When news meets the audience: How audience feedback online affects news production and consumption. Human Communication Research, 43, 436–449. https://doi.org/10.1111/hcre.12123

39.

(2016). When does web-based personalization really work? The distinction between actual personalization and perceived personalization. Computers in Human Behavior, 54, 25–33. https://doi.org/10.1016/j.chb.2015.07.049

40.

Kalyanaraman

(2013). “I, me, mine” or “Us, we, ours?”: The influence of cultural psychology on web-based customization. Media Psychology, 16, 272–294. https://doi.org/10.1080/15213269.2013.815049

41.

Liu

Hong

(2019). The effect of preference stability and extremity on personalized advertising. Journalism & Mass Communication Quarterly, 96, 406–427. https://doi.org/10.1177/1077699018782203

42.

Liu

Dolan

Pedersen

E. R.

(2010). Personalized news recommendation based on click behavior [Conference session]. IUI’10, Hong Kong, China. https://www.researchgate.net/profile/Peter_Dolan3/publication/221607614_Personalized_news_recommendation_based_on_click_behavior/links/0deec53481fd037f15000000.pdf

43.

MacKenzie

S. B.

Lutz

R. J.

Belch

G. E.

(1986). The role of attitude toward the ad as a mediator of advertising effectiveness: A test of competing explanations. Journal of Marketing Research, 23, 130–143. https://doi.org/10.2307/3151660

44.

Markus

Nurius

(1986). Possible selves. American Psychologist, 41, 954–969. https://doi.org/10.1037/0003-066X.41.9.954

45.

Maslowska

Smit

E. G.

van den Putte

(2013). Assessing the cross-cultural applicability of tailored advertising: A comparative study between the Netherlands and Poland. International Journal of Advertising, 32, 487–511. https://doi.org/10.2501/IJA-32-4-487-511

46.

McGuire

W. J.

(1984). Search for the self: Going beyond self-esteem and the reactive self. In Zucker

R. A.

Aronoff

Rabin

A. I.

(Eds.), Personality and the prediction of behavior (pp. 73–120). Academic Press.

47.

McGuire

W. J.

McGuire

C. V.

(1981). The spontaneous self-concept as affected by personal distinctiveness. In Lynch

M. D.

Gergen

(Eds.), Self-concept: Advances in theory and research (pp. 147–171). Ballinger.

48.

McGuire

W. J.

McGuire

C. V.

Child

Fujioka

(1978). Salience of ethnicity in the spontaneous self-concept as a function of one’s ethnic distinctiveness in the social environment. Journal of Personality and Social Psychology, 36, 511–520. https://doi.org/10.1037/0022-3514.36.5.511

49.

McGuire

W. J.

McGuire

C. V.

Winton

(1979). Effects of household sex composition on the salience of one’s gender in the spontaneous self-concept. Journal of Experimental Social Psychology, 15, 77–90. https://doi.org/10.1016/0022-1031(79)90020-9

50.

Merle

Chandon

J.-L.

Roux

(2008). Understanding the perceived value of mass customization: The distinction between product value and experiential value of co-design. Recherche et Applications en Marketing, 23, 27–50. https://doi.org/10.1177/205157070802300303

51.

Metzger

M. J.

Flanagin

A. J.

(2013). Credibility and trust of information in online environments: The use of cognitive heuristics. Journal of Pragmatics, 59, 210–220. https://doi.org/10.1016/j.pragma.2013.07.012

52.

Metzger

M. J.

Flanagin

A. J.

Medders

(2010). Social and heuristic approaches to credibility evaluation online. Journal of Communication, 60, 413–439. https://doi.org/10.1111/j.1460-2466.2010.01488.x

53.

Mummolo

Peterson

(2019). Demand effects in survey experiments: An empirical assessment. American Political Science Review, 113, 517–529. https://doi.org/10.1017/S0003055418000837

54.

Murthi

B. P. S.

Sakar

(2003). The role of the management sciences in research on personalization. Management Science, 49, 1344–1362. https://doi.org/10.1287/mnsc.49.10.1344.17313

55.

Nan

Heo

(2007). Consumer responses to corporate social responsibility (CSR) initiatives: Examining the role of brand-cause fit in cause-related marketing. Journal of Advertising, 36, 63–74. https://doi.org/10.2753/JOA0091-3367360204

56.

Noar

S. M.

Benac

C. N.

Harris

M. S.

(2007). Does tailoring matter? Meta-analytic review of tailored print health behavior change interventions. Psychological Bulletin, 133, 673–693. https://doi.org/10.1037/0033-2909.133.4.673

57.

Oyserman

(2009). Identity-based motivation: Implications for action-readiness, procedural readiness, and consumer behavior. Journal of Consumer Psychology, 19, 250–260. https://doi.org/10.1016/j.jcps.2009.05.008

58.

Park

Han

Park

(2013). Psychological antecedents and risk on attitudes toward e-customization. Journal of Business Research, 66, 2552–2559. https://doi.org/10.1016/j.jbusres.2013.05.048

59.

Petty

R. E.

Barden

Wheeler

S. C.

(2002). The elaboration likelihood model of persuasion: Health promotions that yield sustained behavioral change. In DiClemente

R. J.

Crosby

R. A.

Kegler

(Eds.), Emerging theories in health promotion practice and research (pp. 71–99). Jossey-Bass.

60.

Pornpitakpan

(2004). The persuasiveness of source credibility: A critical review of five decades’ evidence. Journal of Applied Social Psychology, 34, 243–281. https://doi.org/10.1111/j.1559-1816.2004.tb02547.x

61.

Reed

II Forehand

Puntoni

Warlop

(2012). Identity-based consumer behavior. International Journal of Research in Marketing, 29, 310–321. https://doi.org/10.1016/j.ijresmar.2012.08.002

62.

Rifon

N. J.

Choi

S. M.

Trimble

C. S.

(2004). Congruence effects in sponsorship: The mediating role of sponsor credibility and consumer attributions of sponsor motive. Journal of Advertising, 33, 30–42. https://doi.org/10.1080/00913367.2004.10639151

63.

Schouten

J. W.

McAlexander

J. H.

(1995). Subcultures of consumption: An ethnography of the new bikers. Journal of Consumer Research, 22, 43–61. https://doi.org/10.1086/209434

64.

Sela

Lavie

Inbar

Oppenheim

Meyer

(2015). Personalizing news content: An experimental study. Journal of the Association for Information Science & Technology, 66, 1–12. https://doi.org/10.1002/asi.23167

65.

Stokburger-Sauer

Ratneshwar

Sen

(2012). Drivers of consumer–brand identification. International Journal of Research in Marketing, 29, 406–418. https://doi.org/10.1016/j.ijresmar.2012.06.001

66.

Sundar

S. S.

Marathe

S. S.

(2010). Personalization versus customization: The importance of agency, privacy, and power usage. Human Communication Research, 36, 298–322. https://doi.org/10.1111/j.1468-2958.2010.01377.x

67.

Tam

K. Y.

S. Y.

(2006). Understanding the impact of web personalization on user information processing and decision outcomes. MIS Quarterly, 30, 865–890. https://doi.org/10.2307/25148757

68.

Terry

D. J.

Hogg

M. A.

White

K. M.

(1999). The theory of planned behaviour: Self-identity, social identity and group norms. British Journal of Social Psychology, 38, 225–244. https://doi.org/10.1348/014466699164149

69.

van Doorn

Hoekstra

J. C.

(2013). Customization of online advertising: The role of intrusiveness. Marketing Letters, 24, 339–351. https://doi.org/10.1007/s11002-012-9222-1

70.

Wan

(2008). Resonance as a mediating factor accounting for the message effect in tailored communication—Examining crisis communication in a tourism context. Journal of Communication, 58, 472–489. https://doi.org/10.1111/j.1460-2466.2008.00395.x

71.

York

V. K.

Brannon

L. A.

Miller

M. M.

(2012). Increasing the effectiveness of messages promoting responsible undergraduate drinking: Tailoring to personality and matching to context. Health Communication, 27, 302–309. https://doi.org/10.1080/10410236.2011.585450

72.

Zihayat

Ayanso

Zhao

Davoudi

(2019). A utility-based news recommendation system. Decision Support Systems, 117, 14–27. https://doi.org/10.1016/j.dss.2018.12.001

73.

Zizzo

D. J.

(2010). Experimenter demand effects in economic experiments. Experimental Economics, 13, 75–98. https://doi.org/10.1007/s10683-009-9230-z