Intimate Relationships and Attitudes Toward Celebrities

Previous research indicates that persons who self-report a high level of preoccupation with celebrities tend to have lower levels of well-being. We administered the “Romantic Partner Conflict Scale”, the “Love Attitudes Scale”, the soulmate subscale from the “Relationship Theories Questionnaire”, and the anxiety subscale from the “Experiences in Close Relationships Scale” to 330 students from four universities to see how well scores on these measures would predict scores on each of the three subscales from the “Celebrity Attitude Scale” (CAS). We predicted that persons whose scores on these measures of intimate relationships indicated a troubled, anxious, or poor quality relationship would have higher scores on the CAS, especially on its two problematic subscales. In three multiple regressions, specific measures of behavior during conflict with a romantic partner and certain love styles significantly predicted scores on all three of the CAS subscales. We discuss the implications of being a celebrity worshiper on one’s relationship with an intimate partner.

A small percentage of celebrity worshipers develop an Intense Personal connection with their favorite celebrity (Level 2), reporting a strong identification with and an obsession about their favorite celebrity. An even smaller number of persons from Level 2 eventually become increasingly addicted and move to Level 3, which was termed Borderline Pathological because Level 3 persons report a willingness to engage in irrational and illegal behaviors on behalf of their favorite celebrity . These two higher levels appear to be more problematic, inasmuch as those who reach one or both of them appear to be prone to narcissism (Ashe et al., 2005), are less likely to be cognitively flexible and less able to cope with life's challenges (Maltby, Day, McCutcheon, Martin, & Cayanus, 2004;Martin, Cayanus, McCutcheon, & Maltby, 2003), more likely to be mentally unhealthy (Maltby, Day, McCutcheon, Gillett, et al., 2004), and more likely to have neurotic and psychotic tendencies . In other words, celebrity worshipers, especially those who are absorbed in or addicted to the personal lives of their favorite celebrities, seem to have a low level of well-being. Further evidence for this hypothesis stems from McCutcheon et al.'s (2004) findings in which favorable attitudes toward celebrities were significantly associated with difficulties in finding meaningfulness in life. In a more recent study, high scores on the Borderline Pathological level were associated with tendencies to have difficulty paying attention, disinhibition, and a focus on immediate concerns rather than the future . Although the link between the two problematic levels of celebrity worship and a variety of troubling attitudes and behaviors seems to be well documented, none of the above-mentioned studies targeted intimate romantic relationships.
There are several studies suggesting that mentally unhealthy persons are so preoccupied with their own problems that they have difficulty maintaining satisfactory relationships with romantic partners (Braithwaite, Delevi, & Fincham, 2010). For example, there is evidence that married persons who are addicted to video games ignore their spouses, fail to do household chores, and become withdrawn and irritable (Northrup & Shumway, 2014). In a study of Dutch couples, higher neuroticism and lower self-esteem in both partners were associated with lower quality relationships (Luteijn, 1994). A study of married veterans of the Vietnam war showed that those with posttraumatic stress disorder (PTSD) were more than twice as likely as non-PTSD veterans to report clinically significant levels of relationship distress (Riggs, Byrne, Weathers, & Litz, 1998). Persons deemed to be emotionally unstable were less likely to have had stable marriages over a 50-year period (Kelly & Conley, 1987). Impulsive persons were less accommodating to their partners as compared to those with high self-control (Finkel & Campbell, relationships. In this study, a game-playing style, which is characterized by minimal commitment to one's partner, was positively correlated with Borderline Pathological scores on the CAS in general, and especially for individuals whose favorite celebrity was a beautiful person of the opposite sex. Additionally, a dependent love style, which is characterized by possessiveness and jealousy, was positively correlated with the Intense Personal subscale on the CAS (McCutcheon, 2002).
The purpose of the present study is to explore the pattern of correlations between the Celebrity Attitude Scale and its subscales and several measures of the quality and type of intimate relationships. In addition to the Love Attitudes Scale we also measured the extent to which one feels anxiety about intimate relationships, the value one places on an intimate partner, the way in which disagreements are resolved with a partner, and the extent to which one's partner is regarded as a "soulmate. " Franiuk, Cohen, and Pomerantz (2002) found that a "soulmate theorist" is one who believes that there is just one person (or a very few) with whom they are destined to have a satisfying intimate relationship. Soulmate theorists are plentiful, in large part due to American culture and its obsession with songs, movies and books that promote the soulmate theory (Franiuk et al., 2002). This same culture also promotes the worship of celebrities.
Based on our review of the relationship literature we hypothesized that individuals who score high on the Intense Personal and Borderline Pathological levels, will tend to report lower quality intimate relationships on the measures described below. Specifically, we predict that high CAS scores will be associated with greater argumentation and less compromise with one's intimate partner, and greater anxiety about the relationship. We also predict that high CAS scores will be associated with dependent and game-playing styles of loving, and with greater endorsement of the "soulmate" theory.

Method Participants
The sample consisted of N = 330 undergraduate students (74% female) from four universities in the eastern and central United States, ranging in age from 18 to 60 years (M = 20.97; SD = 4.82). The majority of participants reported being White (62%), or African American (26%), with others identifying as Asian (4%), Hispanic (3%), two or more races (5%), and other (1%); 1% of the participants did not indicate their ethnicity.
The Romantic Partner Conflict Scale (RPCS) consists of 39 items divided into 6 subscales through factor analysis.
Respondents are asked to "think about a significant conflict issue that you and your partner have disagreed about recently. Using the scale below, fill in which response is most like how you handled conflict. If you do not have a romantic partner, respond with your most current partner in mind. If you have never been in a romantic relationship, answer in terms of what you think your responses would most likely be." The Compromise subscale has 14 items and is typified by items like "We try to find solutions that are acceptable to both of us." A typical item from the 6item Domination subscale is "When we argue or fight, I try to win." The Submission subscale has 5 items and contains items like "When we have conflict, I usually give in to my partner." The Separation subscale has 5 items including items such as "When we have conflict, we separate but expect to deal with it later." The Avoidance subscale has 3 items including items like "My partner and I try to avoid arguments." The Interactional Reactivity subscale has 6 items including items such as "When my partner and I disagree, we argue loudly." All items are answered on a Likert-type scale with anchor points being "strongly agree with statement" equal to 4 and "strongly disagree with statement" equal to 0. High scores on each subscale indicate endorsement of that strategy of dealing with partner conflict. However, due to a coding error in the survey, 78% of the sample completed the measure using a 1 to 5 scale. Because the majority of participants completed the measure using this scale, the response scale was changed to 1 to 5 for the remaining participants for the purpose of data analysis. The subscales appear to have very good validity. Test-retest reliabilities for the subscales are strong, and alpha reliabilities have ranged from .82 to .96 (Zacchilli, Hendrick, & Hendrick, 2009). In the present study the alphas were .93, .91, .88, .88, .82, and .85 respectively.
The anxiety subscale from the Experiences in Close Relationship Scale (ECR-anxiety) consists of six Likert-type items with anchor points being "strongly disagree" equal to 1 and "strongly agree" equal to 7. Respondents are asked "how you feel in romantic relationships… how you generally experience relationships, not just in what is happening in a current relationship." A sample item reads "My desire to be very close sometimes scares people away." High scores suggest a person who experiences considerable anxiety in intimate relationships. This subscale has been shown to have adequate test-retest and internal reliability, and it correlated .45 and .42 with measures of excessive reassurance seeking and depression, respectively (Wei, Russell, Mallinckrodt, & Vogel, 2007). In the present study the alpha reliability was .70.
The version of the Love Attitudes Scale (LAS) that we used was the 4-item per subscale revision used by Hendrick and colleagues (1998). Respondents are requested to "answer the questions with your current partner in mind.
If you are not currently dating anyone, answer the questions with your most recent partner in mind. If you have never been in love, answer in terms of what you think your responses would most likely be." Strong agreement with any of the 24 items results in a score of 5, and strong disagreement yields a score of 1. High scores indicate the endorsement of a particular love style. The six love styles include Passionate (e.g., "My partner and I have the right physical chemistry between us"), Game-playing (e.g., "I have sometimes had to keep my partner from finding out about other lovers"), Friendship (e.g., "Our friendship merged gradually into love over time"), Practical (e.g., "An important factor in choosing my partner was whether or not he/she would be a good parent"), Dependent (e.g., "When my partner doesn't pay attention to me, I feel sick all over"), and Selfless (e.g., "I would rather suffer myself than let my partner suffer"). This version of the LAS has been described as having "even stronger psychometric properties than the original scale" (1998, p. 147 The Soulmate subscale from the Relationship Theories Questionnaire (RTQ-Soulmate) consists of 11 items. Respondents are simply requested to "circle the number that best indicates how you feel." Strong agreement with any of the items results in a score of 7, and strong disagreement yields a score of 1, except two items that are reverse-scored. High scores indicate the endorsement of the idea that there is only one, or very few persons with whom they can have a satisfactory intimate relationship, and they are destined to meet that individual. A sample item is "There is a person out there who is perfect (or close to perfect) for me." This subscale has been shown to have good reliability and validity (Burnette & Franiuk, 2010;Franiuk et al., 2002;Franiuk, Shain, Bieritz, & Murray, 2012). Cronbach alpha in the current study was .75.

Procedure
We administered a questionnaire on "Intimate Relationships" that included four instruments and demographic questions. The measures were administered in accordance with the IRBs from the four universities. Participants from three institutions completed the survey online in exchange for course credit toward their psychology course.
Students at one university completed a paper and pencil version of the questionnaire in a classroom setting, and in some cases received a small amount of extra-credit for volunteering to participate. To reduce the likelihood of a systematic order effect the six measures were arranged and presented in 10 different orders that were selected randomly.
Some cases (N = 18 total) were eliminated from the data before analyses were performed. Some participants (N = 17) were eliminated because they did not follow directions on a question about their favorite celebrity, defined as "a famous living person (or one who died during your lifetime) that you greatly admire" (e.g., "Jane Austen," "N/A"). One case was removed because the pattern of responses indicated that the participant was not effectively reading or responding to questions sincerely. In cases where 20% or less of data were missing within a measure, mean substitutions were performed. If more than 20% of data were missing in that (sub)scale, then the total was not calculated.

Results
The descriptive statistics for all the measures are shown in Table 1. These values are consistent with scores obtained in previous studies. For example, the mean scores per item on each of the CAS subscales typically are highest for ES (2.8 in our study), and somewhat lower for IP (1.99 in our study) and BP (2.11 in our study) (Mc-Cutcheon, Griffith, Aruguete, & Haight, 2012;McCutcheon, Scott, Aruguete, & Parker, 2006), and coefficient alphas for ES and IP that are about equal and somewhat higher than those for the four-item BP McCutcheon, 2002;McCutcheon et al., 2012;McCutcheon et al., 2006). Not shown in Table 1 is the finding that the favorite celebrities that were chosen followed a pattern that was similar to that found in previous studies, namely that approximately 75% of the favorites were either actors, singers, or athletes (Green, Griffith, Aruguete, Edman, & McCutcheon, 2014;McCutcheon et al., 2004).   To test our hypotheses, we first performed correlations to illustrate bivariate associations between the celebrity attitude measure and the relationship measures (see Table 2). Because of the large number of correlation coefficients that we computed we used .01 as the cutoff for significance in order to lessen the probability of making a Type 1 error. Individuals with higher scores on all three celebrity attitude measures reported higher dominance, submission, separation, and interactional reactivity during conflict with a romantic partner; higher anxiety in intimate relationships; a stronger tendency to endorse higher practical, and dependent love styles; and they were more likely to endorse soulmate beliefs. Other relationship measures were correlated with one of the three celebrity attitude scales. Specifically, individuals with higher scores on the borderline pathological subscale reported a lower tendency to compromise during conflict with romantic partners, and individuals with higher scores on the entertainment-social subscale reported higher avoidance during conflict with romantic partners. Passionate, Friendship, and Selfless love styles were not significantly associated with any of the celebrity attitude scales.   As a next test of our hypotheses, we computed three multiple stepwise regressions, one for each CAS subscale in an effort to determine the relative contribution of the combined predictive power of the relationship variables (see Table 3). We eliminated the following predictors from the models due to not being correlated with the celebrity scales and/or because they are less pathological: RPCS-Compromise, LAS-Passionate, LAS-Friendship, and LAS-Selfless. Given our hypotheses, we are primarily interested in the relative contributions of the more negative relationship variables. In Step 1 of the models, we entered whether or not the survey was completed online or in person, race (White, non-White), and age as controls. In Step 2, we entered the relationship variables (except the four noted above).
Step 2 in all three models was significant, explaining 26 -28% of the variance in the outcomes. In the CAS-ES model the significant predictors were race, RP SEP, RP AVO, and LAS-D. Nonwhite students, individuals who were higher on separation and avoidance during conflict with a romantic partner, and individuals who had more practical love styles tended to value celebrities more highly for entertainment and social reasons. In the CAS-IP model, race, RP SUB, RP SEP, RP AVO, LAS-P, and LAS-D were significant predictors. Non-white students, individuals who were higher on submission, separation, and avoidance during conflict with a romantic partner, and individuals who had more practical and dependent love styles had more intense personal feelings for celebrities. In the CAS-BP model, individuals who completed the survey online, individuals who were higher on submission and separation during conflict with a romantic partner, and individuals with more dependent love styles had more borderline pathological attitudes toward their favorite celebrities.

Borderline Pathological Intense-Personal Entertainment-Social Predictor
Step 1

Discussion
There is a substantial body of literature that correlates excessive celebrity admiration with poorer mental health Maltby, Day, McCutcheon, Gillett, et al., 2004) difficulty in coping with life's challenges (Maltby, Day, McCutcheon, Martin, & Cayanus, 2004;Maltby, McCutcheon, & Lowinger, 2011;, and the tendency to exploit others (Ashe et al., 2005). However, the literature generally does not address relationships between celebrity worship and intimate relationships.
The present study extends this previous research by showing that excessive celebrity admiration is associated with a combination of previous poor quality intimate relationships, current ones, and the expectation of poor quality intimate relationships in the future. As predicted, those persons who scored high on the two problematic subscales of the CAS also tended to resolve intimate relationship conflicts by arguing rather than compromising (at least for CAS BP). We found that the tendency to argue was negatively related to the tendency to compromise (see Table 2). Gottman (1998) found that healthy marriages tended to be ones in which couples were able to Interpersona 2016, Vol. 10(1), 77-89 doi:10.5964/ijpr.v10i1.208 reconcile conflict. Other researchers have found that when unhappy couples argue, they tend to say hurtful things that might tend to weaken a relationship (Markman, Floyd, Stanley, & Storaasli, 1988;Yovetich & Rusbult, 1994).
Thus excessive arguing rather than compromising, is one indicator of an unhealthy intimate relationship. Individuals who are obsessed with celebrities tend to be cognitively inflexible (Maltby, Day, McCutcheon, Martin, & Cayanus, 2004;Martin et al., 2003), which suggests an inability or unwillingness to compromise.
We predicted and found a positive correlation between ECR-Anxiety and CAS-IP and CAS-BP. Thus, persons who experience a great deal of anxiety in close relationships also tend to have high scores on the two problematic subscales of the CAS. It is tempting to suggest that persons who experience a great deal of anxiety in the course of forming and maintaining satisfactory intimate relationships become strongly absorbed and even addicted to their favorite celebrity, as a less threatening parasocial substitute for a satisfactory intimate relationship. However, it is also possible that some individuals first become absorbed or addicted to their favorite celebrity, then find that their absorption interferes with attempts to establish and maintain a satisfactory intimate relationship with a noncelebrity.
A previous study showed that those who endorsed a game-playing love style tended to score high on CAS-borderline pathological if their favorite celebrity was a beautiful person of the opposite sex, and those who endorsed a dependent love style tended to score high on CAS-Intense Personal (McCutcheon, 2002). Our attempt to replicate and extend this earlier finding was successful in that game-playing was found to be correlated with both CAS Intense Personal and CAS Borderline Pathological. Furthermore, the correlation between scores on the measure of dependent love style and scores on CAS Intense Personal and CAS Borderline Pathological were both significant and relatively high (.36 & .40 respectively). It strikes us that a person who endorses a dependent love style might be somewhat maladjusted. Consider the sample item "When my partner doesn't pay attention to me, I feel sick all over," from the LAS-D subscale. Jealousy and extreme anxiety are seemingly implied by endorsing such a statement. Indeed, we found that LAS-D scores correlated .47 with scores on ECR-Anxiety (See Table 2). Perhaps high scorers on LAS-D are attracted to celebrities because they can fantasize that their favorite celebrities are fulfilling the needs that they feel are being neglected by their intimate partners. Indeed, one study found a significant link between fantasy proneness and CAS Intense Personal scores, and between both fantasy proneness and dissociation, and scores on CAS Borderline Pathological (Maltby, Day, McCutcheon, Houran, & Ashe, 2006).
We predicted and found a positive correlation between RTQ-Soulmate and both CAS-Intense Personal and CAS-Borderline Pathological. Those participants who tended to endorse the soulmate belief also tended to be more absorbed and addicted to their favorite celebrity. One interpretation of this finding is that American popular culture is a mediator that promotes both the soulmate theory (songs with themes like "you're the only one for me," & "I can't live without you") and the obsession with celebrities (see the magazine rack at a supermarket, or TV shows about celebrities). Another possibility is that a decline in both the soulmate belief and celebrity adulation comes with the wisdom about relationships that accrues gradually through adulthood. There are at least six studies indicating that CAS scores decline slightly through the adult years (see McCutcheon et al., 2004 for a discussion of this), and the soulmate belief seems unrealistic in light of the success of many arranged marriages and the failure of so many love ("soulmate") marriages (Gupta & Singh, 1982). Some adults gradually come to realize that celebrities are dispensable, serving little purpose other than the promotion of products and services. Some adults also become disillusioned with the soulmate they married, and come to realize that there are several other suitable persons with whom a happy long-term relationship is possible. Although our hypotheses about relationships between the two more problematic subscales of the CAS and anxiety, arguing with one's partner, game-playing as a love style, and endorsement of the soul mate belief were all confirmed with bivariate correlations, the result of our three multiple regressions changed the landscape somewhat. Our contention that LAS-Dependent is a very strong predictor of celebrity worship was supported, but the other four predictors (anxiety, game playing, arguing, soul mate belief) were supplanted by RPCS-Separation, which was significant in all three regressions, and RPCS-avoidance, which was significant in two of the three multiple regressions. Could it be that persons who score high on RPCS-Separation and RPCS-Avoidance have a difficult time dealing with conflict, so they physically separate themselves from their partners and avoid discussing their disagreements? Do they also need an escape from the conflicts of everyday life, and their excessive interest in celebrities provides an escape from those conflicts? We recommend that future research on the link between celebrity adulation and intimate relationships explore the reasons why the increased usage of separation and avoidance as strategies for resolving conflicts is positively correlated with celebrity adulation.
We were somewhat surprised that problematic intimate relationships predicted CAS-ES about as well as CAS-IP and CAS-BP. Several studies show that problematic attitudes and behaviors are usually associated only with the Intense Personal and/or Borderline Pathological levels, as the Absorption-Addiction Model predicts (Ashe et al., 2005;Maltby et al., 2003;Maltby, Day, McCutcheon, Gillett, et al., 2004;Maltby, Day, McCutcheon, Martin, & Cayanus, 2004;McCutcheon et al., 2014). On the other hand, there are at least two studies in which the Entertainment Social subscale also correlated significantly with problematic behavior (Maltby, McCutcheon, Ashe, & Houran, 2001;McCutcheon et al., 2006). Also, the three CAS subscales usually correlate highly with each other, as they did in the present study (see Table 2). Perhaps valuing celebrities only for their entertainment/social value may be problematic for intimate relationships in particular because the time spent engaging with celebrity media may take time away from interacting with one's romantic partner.
Another unexpected finding from the multiple regressions is that race was a significant predictor of scores on two of the three CAS subscales. One previous study, albeit with a smaller sample, did not show that Black students scored significantly higher on the CAS than White students . A very recent study, however, has shown that Black students scored higher than White students on all three of the CAS subscales (Mc-Cutcheon, Aruguete, Jenkins, McCarley, & Yockey, submitted). The authors speculated that celebrity worship provides temporary relief from the stress of racism by fantasizing about a celebrity, who because of money and status, presumably faces fewer race-related problems. We recommend research that attempts to replicate and further explain this recent finding.
A limitation of our study is that respondents were not specifically asked if they were currently in an intimate relationship, or if they had ever been in one. The instructions that accompany each of the four measures of intimate relationships that we used vary. Future research should ask if respondents are currently in an intimate relationship, or if they had ever been in one. Other limitations relate to our sample. Although we sampled from four institutions, our participants were college students, thus they might not have been representative of the population at large. Furthermore, with a mean age of about 21 and a standard deviation of less than five years, it seems likely that the number and the nature of their intimate relationships would differ from older samples. We did not choose to measure sex differences, partly because Franiuk and her colleagues (2002) did not find sex differences on the RTQ-Soulmate Scale, but primarily because our main focus was on the Celebrity Attitude Scale, an instrument which does not usually yield sex differences (see Maltby, McCutcheon, & Lowinger, 2011, for example the possibility of sex differences in intimate relationships is a real one (see Shackelford & Goetz, 2012, for example), so we recommend that future research on this topic include a discussion of sex differences.
Our research supports and extends the findings of previous studies that have shown correlations between excessive celebrity admiration and a variety of problematic relationship attitudes and behaviors. The present study extends the previous work by showing that celebrity worship may also be related to problematic attitudes and behaviors when the relationships in question are with intimate partners. Of particular concern is the relationship between excessive celebrity admiration and the tendency to favor a dependent (needy, submissive, anxious) love style.
Furthermore, unlike the findings of some previous studies, this one implicated the Entertainment Social level of celebrity admiration, a level generally regarded as relatively benign.

Funding
The authors have no funding to report.