This is an open access article distributed under the terms of the Creative Commons
Attribution License (In 2009, Sara Henry, a preschool director in California, was told by her employer to either marry her live-in boyfriend or be terminated from her position. Henry chose to continue cohabiting and filed a wrongful termination lawsuit against the Red Hill Lutheran Church of Tustin. The lawsuit, along with Henry’s 2011 appeal, was struck down, with the court citing legal and civil rights acts along with a ministerial exception to consequences for employment discrimination (Jolly, 2011). Similar sanctions have been levied against a physical education and baseball coach in Wisconsin (Borsuk, 2009) and an emergency dispatcher in North Carolina (Weigl, 2006) for cohabiting, whose employers cited contractual issues and ‘moral considerations’ as the rationale behind their terminations.
Although seldom enforced in criminal court, anti-cohabitation laws in several states (West Virginia, Florida, Michigan, Mississippi, and North Dakota) have been employed in other legal contexts to deny unmarried couples rights to child custody and visitation, unemployment financial assistance, and protection from housing and employment discrimination. The persistence of these laws reflects the reluctance of relationship traditionalists to grant legitimacy to any living arrangement other than heterosexual marriage. As a consequence, recent challenges to these laws have become flash points in the dispute over the morality and advisability of adults’ choice to live as an unmarried couple.
Despite a dramatic increase in the number of American couples engaged in unmarried cohabitation (UC hereafter) in recent decades (Casper, Cohen, & Simmons, 1999; see also Poponoe & Whitehead, 2007) social acceptance of the practice remains elusive in certain quarters (Poponoe & Whitehead, 2008; Thornton, 1989; Thornton & Young-DeMarco, 2001). This was recently illustrated in the state of Virginia’s 2013 repeal of a ban on unmarried cohabitation and sexual relations outside of marriage when 25 legislators voted to uphold the ban (Wilkins, 2013). Other research suggests UC may carry a social stigma, particularly for women. For example, in-depth interview and focus group research suggests that male participants infer social disapproval from their female partner’s family when they engage in a UC relationship and female participants perceive UC as having lower social status and legitimacy, as compared to marriage (Huang, Smock, Manning, & Bergstrom-Lynch, 2011).
Central to proponents’ and opponents’ perspectives are differing viewpoints regarding religion, the significance of marriage and family, and commitment to historical family roles and institutions (Feltey & Poloma, 1991; Lottes & Kuriloff, 1992; Lye & Waldron, 1997; Thornton & Young-DeMarco, 2001). UC proponents frequently cite pragmatic arguments (e.g., financial considerations) or attach a practical consideration to their support of the practice (e.g., as a test for compatibility; Popenoe & Whitehead, 1999; Surra, Gray, Cottle, & Boettcher, 2004). In addition, proponents tend to have less traditional views towards family and marriage than opponents (Lye & Waldron, 1997).
In contrast to the utilitarian arguments of proponents, UC opponents frequently cite religious or moral arguments against the practice (Popenoe & Whitehead, 1999). More specifically, research has found religiosity to be a correlate of traditional attitudes concerning cohabitation and family (Feltey & Poloma, 1991). Opponents also tend to cite moral arguments underscoring the importance and permanence of marriage, and the inappropriateness of cohabitation, sexual relations, and child-bearing outside of marriage (Thornton & Young-DeMarco, 2001). In addition to the religious and moral arguments, other research has demonstrated a connection between attitudes toward family and other general values and beliefs such as political ideology (Lye & Waldron, 1997; Thornton, 1989). Specifically, individuals with more conservative ideology tend to have more traditional attitudes toward sex and marriage (Lye & Waldron, 1997).
These alternative views are reflected in the different forms that UC may take (Surra et al., 2004). In one form, the UC couple characterizes themselves as members of a dating relationship who choose to live together. Coresidential daters typically report having no plans to marry, a moderate to high likelihood of dissolution, and uncertainty about their compatibility with their partners (Casper & Bianchi, 2002; Manning, 1993). When compared to married partners, coresidential daters tend to experience less happiness and more conflict than their married counterparts (Brown & Booth, 1996). A second form UC may take is as a stage in the courtship continuum beyond dating and mate selection but before formal engagement or marriage. Couples in this group frequently report a low likelihood of relationship dissolution and a strong intention to ultimately marry (Bumpass & Lu, 2000). UC couples who plan to marry their partners experience relatively similar levels of relational happiness and conflict as their married partners (Brown & Booth, 1996). In a third form, the UC couple sees their relationship as the informal equivalent of marriage, with partners living as if they were married in all respects except for official documentation. Couples in this group typically cite pragmatic considerations (e.g. economics) and lack of concern about broad social approval in their decision to remain unmarried (Casper & Bianchi, 2002). A comparison of the research on the relational quality of permanent UC relationships versus married relationships indicates generally higher levels of relational quality for married couples (Brown & Booth, 1996). Research suggests marital relationships are happier and experience lower conflict, less physical aggression, and more stability than UC couples (Brown, 2000; Stets, 1991; Thomson & Colella, 1992). However, these findings are somewhat mitigated if the cohabitators plan to marry their partners (Brown & Booth, 1996).
The UC debate clearly derives from differences in the attitudes and beliefs proponents and opponents hold about relationships, the institution of marriage, and the definition of “family.” However, the contentiousness of the dispute is aggravated by a belief the two sides’ share– that the perspectival gap between themselves and their adversaries is vast and irreconcilable. Perceptions of this alleged polarity are evidenced by the labels partisans use to describe one another. Activist and outspoken proponent Bella DePaulo (2015) describes the opposition as “marriage opportunists” who “seem to think that married people are morally superior people” (para. 9). DePaulo further criticizes “’matrimania’ . . . the over-the-top hyping of marriage and coupling and weddings that are pervasive in American society” arguing that “the marriage opportunists” would like to see more of it (DePaulo, 2015, para. 16). At the other extreme, opponents describe cohabitation as an “embracing of evil” (McManus, 2004, para. 8), and liken the practice of UC with a “trial divorce” (McManus, 2011, p. 168). Such inflammatory remarks notwithstanding, how accurate are partisans in the UC debate at perceiving the attitudes, beliefs, and motives of their adversaries?
The reported research explores the hypothesis that in this debate, as in many other social conflicts, the attitudinal differences may be smaller and the common ground greater than the adversaries recognize. This hypothesis is predicated on past research demonstrating that people discount the subjective nature of their viewpoints, a tendency which ultimately leads them to inflate perceptions of subjectivity in their behavioral attributions and predictions about others (Griffin, Dunning, & Ross, 1990; Robinson, Keltner, Ward, & Ross, 1995; Ross & Sicoly, 1979). Ross and Ward (1996) referred to this phenomenon as “naïve realism” – i.e., one’s conviction that he or she is privy to a knowable, objective reality that others will also grasp if they are rational and reliable themselves. By equating one’s own perspective of the world with objectivity, naïve realism leads the perceiver to treat others who hold a different perspective as victims of lapsed reason, self-interest, ethical compromise, and/or ideological bias. Scholars of intergroup conflict have noted numerous negative consequences of this naïveté, including the tendency to blame adversaries for shared problems, to doubt their sincerity, and to misattribute genuinely benign actions to malicious intent (Chambers, Baron, & Inman, 2006; Keltner & Robinson, 1993, 1996, 1997; Robinson et al., 1995; Ross & Ward, 1996). Perhaps the most troubling consequence is an unwarranted pessimism about the prospect of finding common ground with adversaries whom we presume to be biased or irrational merely because we disagree with them. For example, pro-choice and pro-life partisans often overlook mutually agreeable interests (the need for programs to prevent unwanted pregnancies, on-site day care for working mothers, etc.) by virtue of erroneous assumptions they make about their adversaries’ rationality and motivation (Sherman, Nelson, & Ross, 2003).
Studies of naïve realism have demonstrated significant inaccuracies in partisans’ perceptions of intergroup attitudes. For example, Robinson et al. (1995) asked pro-choice and pro-life college students to evaluate a series of vignettes describing cases of abortion (e.g., a businesswoman terminates a pregnancy that poses a threat to her career aspirations). For each scenario, participants reported the level of sympathy they personally felt and also estimated the degree of sympathy felt by typical members of the pro-choice and pro-life camps. Their results indicated widespread perceptions of disagreement among both partisan groups. However, these perceptions proved to be greatly exaggerated when compared to partisans’ self-report ratings. For example, pro-choice participants assumed they felt much more sympathy for women seeking abortion than the typical pro-life partisan, when in fact the differences between the groups in self-reported sympathy were quite modest. This pattern of “false polarization” – i.e., a bipartisan exaggeration of differences in intergroup attitudes and beliefs – was also observed in partisans’ perceptions of their adversaries’ ethical, moral, and/or religious convictions, as well as the influence of political ideology on their abortion stance (see also Chambers et al., 2006).
“False polarization” (Keltner & Robinson, 1997; Robinson et al., 1995) occurs when partisans perceive the attitudinal differences between themselves and their adversaries to be larger than is indicated by a comparison of the attitudes members of each group attribute to themselves. In this case, it was predicted that UC proponents and opponents would overestimate the extent to which the groups differ in a) their sympathy toward scenario characters who choose to live in a UC relationship and b) the role of ideology and political orientation as a basis for their stance on the practice of UC.
Relatedly, naive realism proffers predictions regarding partisan self-serving interpretations of both their supporters and opponent’s attitudinal stances—referred to as egocentric bias. Egocentric bias (Ross, Green, & House, 1977) occurs when individuals interpret attitudinal differences between themselves and others in self-serving ways. In particular, people tend to interpret such differences as indicating that others are less rational, ethical, and/or more prone to ideological bias than they are. In this case, it was predicted that partisans would perceive their adversaries as more influenced by ideology and political orientation, than either allies or themselves.
The reported research extends the naïve realism paradigm to the UC debate as a means for identifying adversarial misperceptions that arise in this dispute. As the subject of significant social and moral scrutiny, UC seems fertile ground for naïve realism phenomena. Although UC proponents might view their endorsement as the product of a thoughtful, pragmatic evaluation of relationships, opponents are inclined to see it as a consequence of loose liberalism or moral slouching. In contrast, opponents of cohabitation frame their opposition as reflecting a moral stance, but proponents see it as a product of reactionary conservatism. Thus proponents and opponents both feel that their own views are based on careful deliberation, but their adversaries’ views are tainted by self-interest and bias. Attributing their adversaries’ views to these sources leads naïve realists to see these views as inflexible and extreme, in contrast to the more complex or ambivalent views they attribute to themselves (Ross & Ward, 1996; Sherman et al., 2003).
We employed a survey format based on materials and procedures developed by Robinson et al. (1995) to explore naïve realism in the UC debate. College students who identified themselves as UC proponents or opponents were presented with a series of vignettes describing cases in which men and women choose to live together as an unmarried heterosexual couple. Participants were asked to describe not only their own impressions of the couple in each scenario, but also to speculate about the impressions of the typical UC proponent and opponent. The comparison of respondents’ self-reported attitudes and those they attributed to typical allies and adversaries provided the basis for a systematic measure of naïve realism’s operation in partisans’ perceptions. Two hypotheses regarding different aspects of naïve realism were tested.
H1: The attitudes partisans attribute to themselves, their allies, and their adversaries exhibit a pattern of false polarization.
“False polarization” (Keltner & Robinson, 1997; Robinson et al., 1995) occurs when partisans perceive the attitudinal differences between themselves and their adversaries to be larger than is indicated by a comparison of the attitudes members of each group attribute to themselves. In this case, it was predicted that UC proponents and opponents would overestimate the extent to which the groups differ in a) their sympathy toward scenario characters who chose to live in a UC relationship and b) the role of ideology and political orientation as a basis for their stance on the practice of UC.
H2: The attitudes partisans’ attribute to themselves, their allies, and their adversaries exhibit a pattern of egocentric bias.
Egocentric bias (Ross, Green, & House, 1977) occurs when individuals interpret attitudinal differences between themselves and other in self-serving ways. In particular, people tend to interpret such differences as indicating that others are less rational, ethical and/or more prone to ideological bias than they are. In this case, it was predicted that partisans would perceive their adversaries as more influence by ideology, political orientation, and other non-pragmatic considerations (e.g. philosophical beliefs) than either themselves or their allies.
Method
Participants
One hundred ninety-one undergraduate students enrolled in introductory communication courses at a large university in the southwestern U.S. received course extra credit for their participation in this study. Participants ranged in age from 18-41 years and the mean age of the sample was 20.26 (SD = 2.56) years. Over half of participants (58%) identified themselves as European-American and the remainder of the sample was comprised of various ethnicities including Hispanic (17.8%), African-American (4.7%), Asian (12.6%), and ‘Other’ (6.8%). Of the participants, 94 identified themselves as pro-UC, 57 identified themselves as anti-UC, and 38 ranked themselves as neutral (2 recorded no response). Approximately 20% of the participants had either previously engaged in UC (n = 24) or were currently cohabiting (n = 11). In addition, four of the participants were currently married and two reported that they had been married previously.
Materials and Procedure
A survey was created to assess participants’ attitudes regarding UC. The first portion of the survey focused on general demographic information (age, sex, year in school, and ethnicity), marital and UC history, and degree of support or opposition towards UC. Participants responded to four questions regarding marital and UC history, “Are you currently married?”, “Have you ever been married?”, “Are you currently living with a romantic partner?” and “Have you ever lived with a romantic partner you were not married to?” by circling “yes”, “no”, or “no answer” on their questionnaire. To gauge degree of support or opposition toward UC, participants responded to the statement, “Rate the degree to which you support or oppose the practice of unmarried cohabitation” on a 1-7 scale, where 1 equals “strongly support” and 7 equals “strongly oppose”. Participant responses to this question were then recoded into three groups: pro-UC (responses 1, 2 or 3), anti-UC (responses 5, 6, or 7) or neutral (response 4) (Sherman, Nelson, & Ross, 2003). Participant responses were compared as a member of the pro-UC group or as a member of the anti-UC group when analyzing the remainder of items on the survey. The responses of those who were classified as neutral were excluded from data analysis. The re-coding of participants as proponents, opponents or neutrals based on their degree of support is a process commonly utilized in research exploring naïve realism (e.g. Sherman et al., 2003).
The remaining portions of the survey featured four vignettes (see Appendix) created specifically for this study and designed to elicit strong or weak feelings of sympathy toward a UC couple based on the controllability of their circumstances. Each vignette dealt with a couple’s decision to cohabitate without being married. Two were designed to elicit strong feelings of sympathy. One involved a couple who have been dating and learn of an unexpected pregnancy; due to this pregnancy the couple decides to cohabitate once the child is born. A second involved a couple under financial strain due to an unexpected job loss. The other two vignettes were designed to elicit weak feelings of sympathy. One involved a couple who had simply “not gotten around to getting married” yet (Jamieson et al., 2002). The second described a couple who chose to cohabitate in order to afford a larger, more luxurious home. The first two vignettes were selected as likely to elicit strong feelings of sympathy towards the UC couple because they place the circumstances for the cohabitation outside of the couple’s immediate, present control. The second two vignettes were selected as likely to elicit weak feelings of sympathy because they place the grounds for the cohabitation within the couple’s immediate control. Attributions regarding controllability can serve as triggers for sympathy; specifically, research has shown that attributions of uncontrollability lead to feelings of sympathy by observers (Weiner, 1986). Participants viewed the vignettes in varying order and assessed how much sympathy they felt for the couple deciding to cohabitate in each case by responding to the question, “How sympathetic do you feel toward the couple in this scenario” on a 7-point Likert-type scale with “1” indicating “minimal sympathy” and “7” indicating “very sympathetic”.
After participants had assessed all four vignettes, they were asked about the bases for their attitudes and beliefs regarding UC. Specifically, they were asked to indicate on a 1-7 scale (where 1 equals “no influence” and 7 equals “extremely influential”) the extent to which they had been influenced by: 1) actual cases of cohabitation they were aware of, 2) religious, moral, or ethical considerations 3) philosophical beliefs, and 4) their political orientation. We chose to assess respondents’ perceptions of the influence of these factors because they are central to opposing sides views regarding UC (e.g. Feltey & Poloma, 1991; Lottes & Kuriloff, 1992; Lye & Waldron, 1997; Thornton & Young-DeMarco, 2001) as well as to other socially contentious issues (see Robinson, Keltner, Ward, & Ross, 1995).
Following the design structure of previous studies exploring naïve realism (e.g. Sherman, Nelson, & Ross, 2003), when participants had completed all of these ratings pertaining to their own construals and beliefs, they were again presented with the same set of vignettes, but this time asked about the responses that would be made to each item by “typical proponents of unmarried cohabitation” and “typical opponents of unmarried cohabitation,” using the same scales as their self-focused assessments.
We chose to assess participants’ perceptions of “typical proponents of unmarried cohabitation” and “typical opponents of unmarried cohabitation” because studies of naïve realism have noted that partisans involved in conflicts (such as the debate over UC) tend to see the opposing side as extreme, unreasonable, and unreachable, and also to see their own side as similarly (albeit somewhat less) extreme and biased (Robinson et al., 1995). These often faulty and exaggerated viewpoints may preclude efforts to reach conciliation among involved parties. Participants were given as much time as they needed to complete the questionnaire, but most took approximately 30 minutes. Upon completion, participants were thanked for their participation and debriefed regarding the purposes of the study.
Results
To explore the manifestation of naïve realism in the UC debate, proponents’ and opponents’ responses to three types of questions were compared. Self-focused questions assessed respondent’s own attitudes (e.g. How sympathetic do you feel toward the couple in this scenario?); proponent-focused questions gauged respondent’s perceptions of the attitudes held by those who support UC (e.g. How sympathetic would a typical person who supports cohabitation feel toward the couple in this scenario?); and opponent-focused questions probed respondent perceptions of the attitudes held by those that oppose UC (e.g. How sympathetic would a typical person who opposes cohabitation feel toward the couple in this scenario?). The ratings respondents provided to these questions were analyzed in a series of 2 X 3 mixed-model ANOVAs with UC attitudinal stance (proponent or opponent) as a between- subjects factor and question focus (i.e. self, proponent, or opponent) as a within-subjects factor. One-tailed planned comparisons based on univariate analyses for each dependent measure were conducted to test the predicted differences between cell means (Keppel, Saufley, & Tokunaga, 1992).
The principal statistical evidence for naïve realism in the ratings data is an interaction between UC stance and question focus. Two specific interaction patterns provide evidence for naïve realism. One is a pattern of “false polarization” (Robinson et al., 1995), in which both proponents’ and opponents’ actual attitudes (as measured by the self-focused questions) is smaller than the gap they perceive between “typical” partisans (as measured by the proponent-focused and opponent-focused questions). For example, if the actual difference between proponents’ and opponents’ self-report ratings of sympathy for a UC couple in a vignette is small, but the difference in sympathy ratings they assign to typical proponents and opponents is large, this pattern indicates that partisans have exaggerated the “sympathy gap” between themselves and their adversaries.
The second pattern we explored reflects “egocentric bias” (Ross et al., 1977), in which both proponents and opponents perceive the basis for their attitudes regarding UC as more sensible or rational than those of their allies and adversaries. For example, if participants rate the basis for their own views regarding UC as rooted in practical considerations (e.g. actual cases of cohabitation they are aware of), but rate the basis for their adversaries view’s (as well as their allies albeit to a lesser extent) to be rooted in less rational, ideological considerations (e.g. political orientation), this pattern indicates that partisans perceive an egocentric bias in their perceptions of the bases for the views of themselves, their allies, and their adversaries towards UC. To assess perceptions of the impact of partisans’ ideological considerations, we asked respondents to identify the bases of their own views toward UC, as well as those of a typical proponent of UC, and those of a typical opponent of UC, by indicate on a 1-7 scale (where 1 equals “no influence” and 7 equals “extremely influential”) the extent to which they had been religious, moral, or ethical considerations.
False Polarization
We predicted that a pattern of “false polarization” would emerge in partisans’ perceptions concerning attitudes towards the couples in four UC vignettes. Each vignette dealt with a couple’s decision to cohabitate without being married. Two were designed to elicit strong feelings of sympathy. Scenario one was designed to draw out strong feelings of sympathy towards the couple choosing to engage in UC; specifically, it described an unmarried couple who engage in UC due to an unexpected pregnancy. Analyses of the responses for this scenario indicate a significant main effect, F(2, 298) = 165.98, p < .001, η2 = .51 and position X focus interaction, F(2, 298) = 8.41, p < .001, η2 = .03. As illustrated in Figure 1, the mean difference in proponents’ and opponents’ mean ratings of their sympathy toward the unmarried cohabiting couple in scenario one was approximately one point (M = 5.29, SD = 1.41 and M = 4.26, SD = 1.79 respectively), with proponents rating themselves as slightly more sympathetic than opponents rated themselves. Despite the relatively harmonious picture painted by proponents’ and opponents’ mean rating self-reports, analyses of partisan construal of these differences presents a much more dichotomized illustration. Specifically, proponents perceived the “sympathy gap” between allies and typical opponents to be much larger (M = 5.97, SD = 1.28 and M = 2.99, SD = 1.62 respectively) than the actual differences indicated by self-focused ratings, F(1, 298) = 5.67, p = .02, η2 = .02. Opponents displayed a similar misperception, viewing typical proponents as more sympathetic than their typical allies when compared to their self-focused ratings (M = 6.09, SD = 1.01 and M = 3.14, SD = 1.69 respectively), F(1, 298) = 5.54, p = .02, η2 = .02. Thus, in response to an unmarried couple’s decision to cohabitate as a result of an unexpected pregnancy, proponents’ and opponents’ actual degree of sympathy varied slightly, but both sides perceived an exaggerated version of this difference.
Figure 1
Actual and perceived differences in sympathy ratings among opponents and proponents of UC, Vignette 1.
Like scenario one, scenario two also was designed to elicit strong feelings of sympathy towards the couple choosing to engage in UC. Scenario two described an unmarried couple who decides to cohabitate after an unexpected job loss. Analyses of the responses to the relevant questions for this scenario indicate a reliable main effect F(2, 296) = 197.28, p < .001, η2 = .55 as well as an issue position X focus interaction, F(2, 296) = 12.38, p < .001, η2 = .03. As depicted in Figure 2, analysis of the actual differences in proponents’ and opponents’ mean ratings of their sympathy illustrates a modest difference (M = 5.27, SD = 1.47 and M = 4.00, SD = 1.56 respectively), with proponents viewing themselves as somewhat more sympathetic as opponents. Despite their relatively small differences in sympathy, both proponents and opponents were equally guilty of exaggerating the differences between themselves and their adversaries. Proponents perceived the “sympathy gap” between allies and typical opponents to be much larger than self –reports indicate (M = 5.72, SD = 1.36 and 2.81, SD = 1.39 respectively), F(1, 296) = 8.67, p = .003, η2 = .02, as did opponents, when comparing the scores of typical proponents and their allies (M = 5.73, SD = 1.23 and M = 2.82, SD = 1.39 respectively), F(1, 296) = 8.64, p = .003, η2 = .02. Therefore, after being presented with a vignette describing an unmarried couple’s decision to cohabitate after an unexpected job loss, proponents’ and opponents’ actual degree of sympathy varied modestly, but both sides perceived the partisan gap to be larger than indicated by self-reports.
Figure 2
Actual and perceived differences in sympathy ratings among opponents and proponents of UC, Vignette 2.
In contrast to scenarios one and two, scenario three was designed to elicit weak feelings of sympathy for the couple choosing to engage in UC. In order to induce weak feelings of sympathy, scenario three described an unmarried couple who lived together but had simply “not gotten around to getting married yet” (Jamieson et al., 2002). Responses to the relevant questions for this scenario indicate a reliable main effect, F(2, 296) = 197.70, p < .001, η2 = .55 and issue position X focus interaction, F(2, 296) = 15.46, p < .001, η2 = .04. As indicated in Figure 3, the actual differences in proponents’ and opponents’ mean ratings toward the unmarried cohabiting couple were moderate (M = 4.02, SD = 1.68 and M = 2.32, SD = 1.49 respectively), with proponents rating themselves as more sympathetic than opponents. Despite their actual relatively modest differences in feelings of sympathy toward the couple in scenario three, proponents and opponents perceived much more severe degrees of difference between one another. Proponents perceived the sympathy gap between their typical allies and typical opponents to be larger (M = 4.98, SD = 1.56 and M = 1.84, SD = 1.32 respectively) than self-report scores reveal, F(1, 296) = 11.01, p = .001, η2 = .03; with opponents close behind, also overestimating the sympathy gap between typical proponents and their typical allies when compared to self-report ratings (M = 4.98, SD =1.54 and M = 1.60, SD = 1.47 respectively), F(1, 296) = 10.88, p = .001, η2 = .03. In response to an unmarried couple’s decision to cohabitate because they had “not gotten around to getting married yet”, proponents and opponents actual degree of sympathy varied moderately. However, both proponents and opponents largely over-estimated this difference (although the latter to a lesser extent).
Figure 3
Actual and perceived differences in sympathy ratings among opponents and proponents of UC, Vignette 3.
The fourth and final scenario was also designed to present relatively less sympathetic circumstances for the couple choosing to engage in UC. This scenario described an unmarried couple who chose to cohabitate in order to afford a larger, more luxurious home. Responses to the relevant questions for this scenario indicate a reliable main effect, F(2, 296) = 180.01, p < .001, η2 = .52 and issue position X focus interaction F(2, 296) = 18.23, p <.001, η2 = .05. As Figure 4 demonstrates, the actual differences in proponents’ and opponents’ mean ratings for the unmarried cohabiting couple was less than two points (M = 4.36, SD = 1.66 and M = 2.63, SD = 1.41 respectively), with proponents rating themselves as more sympathetic than opponents. However, both proponents and opponents exaggerated the degree to difference between themselves and their adversaries. Opponents saw the greatest degree of difference, perceiving a sympathy gap between typical proponents and their typical allies to be larger (M = 4.98, SD = 1.27 and M = 2.02, SD = 1.20 respectively) than was indicated in their self-reports, F(1, 296) = 11.74, p < .001, η2 = .03. Proponents closely mirrored this misperception, also perceiving the sympathy gap between their typical allies and typical proponents to be larger (M = 5.07, SD = 1.46 and M = 2.29, SD = 1.34 respectively) than self-reported scores indicated, F(1, 296) = 13.07, p < .001, η2 = .04.
Figure 4
Actual and perceived differences in sympathy ratings among opponents and proponents of UC, Vignette 4.
Egocentric Bias
We predicted that a pattern of “egocentric bias” would emerge in partisan perceptions of the basis for attitudes toward the circumstances in which each couple chose to engage in unmarried cohabitation. To assess the impact of participant’s ideological considerations, we asked respondents to identify the bases of their own views toward UC, as well as those of a typical proponent of UC, and those of a typical opponent of UC. Analyses of the relevant responses indicated a reliable main effect, F(2, 374) = 106.30, p < .001, η2 = .33 and issue position X question focus interaction, F(4, 374) = 15.38, p <.001, η2 = .10. Analysis of the difference in mean ratings by proponents’ and opponents’ of the influence of ideological considerations towards the couples in four UC vignettes indicates a moderate difference of approximately two points (M = 3.53, SD = 1.77 and M = 5.72, SD = 1.67). Proponents perceived themselves as less influenced by these considerations than opponents. Interestingly, proponents were relatively accurate in their perception of the ethical gap between typical allies and typical opponents when compared to their self-report point difference (M = 3.79, SD = 1.83 and M = 6.06, SD = 1.55 respectively). Despite their relative accuracy in perceiving the ethical gap between typical allies and typical opponents, proponents’ had distorted perceptions of the influence ethical considerations had on their typical allies—rating them slightly higher than the actual responses provided (M = 3.79, SD = 1.83 and M = 3.57, SD = 1.77 respectively). Proponents were also inaccurate in their ratings of opponents, rating them as slightly more influenced than actual self-reports of opponents indicated (M = 6.06, SD = 1.55 and M = 5.72, SD = 1.67 respectively).
In contrast to the relative accuracy of proponents in assessing partisan differences regarding ethical influence, opponents were less accurate perceiving a much larger ethical influence gap between the partisan groups than the difference in self-reports indicated. Specifically, opponents perceived a larger gap between proponents and opponents (M = 3.09, SD = 1.70 and M = 6.47, SD = 1.03 respectively) than the actual gap taken from self reports, F(1, 374) = 11.6, p < .001, η2 = .07, such that proponents were perceived as being much less influenced by ideological beliefs than opponents. When rating the influence of ethics on UC stance, opponents’ ratings reveal an intragroup bias such that opponents assumed their typical allies were more influenced by ethics than was indicated by self-reports from opponents (M = 6.47, SD = 1.03 and M = 5.72, SD = 1.67 respectively). Along similar lines, opponents also perceived their allies in a faulty manner by underestimating the influence of ethics on proponents’ sentiments. Specifically, opponents rated their adversaries to be less influenced than the actual responses of proponents indicated (M = 3.09, SD = 1.70 and M = 3.57, SD = 1.77 respectively).
Comparison of the partisans’ actual responses with their perceptions of adversaries revealed that the perceived gaps were exaggerated, displaying false polarization for partisans on each side, although particularly strongly for opponents. Proponents rated themselves as the least influenced by ideological considerations (M = 3.57, SD = 1.77), followed by their allies (M = 3.79, SD = 1.83), and rated their adversaries as operating under the most influence of ideological considerations (M = 6.06, SD = 1.55), F(1,374) = 12.51, p < .001, η2 = .06. Ostensibly, opponents’ self-ratings reveal a similar trend, opponents rated themselves (M = 5.72, SD = 1.67) as less influenced than their allies (M= 6.47, SD = 1.33). However, opponents rated proponents as least influenced by ideological considerations (M = 3.09, SD = 1.70), F(1, 374) = 12.08, p < .001, η2 = .05. For opponents, the direction of the polarization was such that their own group was portrayed as being even more influenced by moral considerations (than themselves or their adversaries). This influence is likely one that participants valued, as a high degree of ethical influence would seem a positive group characteristic.
We predicted that a pattern of “egocentric bias” would emerge in partisans’ perceptions of the basis for attitudes toward the circumstances in which each couple chose to engage in UC. To assess the impact of participant’s general political orientation, we asked respondents to estimate the bases of their own views toward UC, as well as those of a typical opponent of UC, and those of a typical proponent of UC. Analyses of the responses to the relevant questions indicated a reliable main effect F(2, 376) = 82.99, p < .001, η2 = .30 and issue position X question focus interaction F(4, 376) = 3.55, p <.05, η2 = .03. Additionally, analysis of the difference in mean ratings of the influence of general political orientation has toward the couple’s in the four UC scenarios was minimal (M = 2.61, SD = 1.64 and M = 2.46, SD = 1.70 respectively). Self-reports indicate that proponents rate themselves as slightly more influenced than opponents.
Proponents’ perceptions of the influence of political orientation on UC perceptions were inaccurate in a number of ways. They perceived the political influence gap between allies and typical opponents to be much larger (M = 3.33, SD = 1.73 and M = 4.91, SD = 1.96) than what actual self-reports indicated. These perceptions thus reflect an intragroup bias, as they portrayed their allies as less influenced by these beliefs than the opponent group. However, proponents’ assumptions of the influence general political orientation had on their typical allies was exaggerated when compared to the actual responses provided (M = 3.33, SD = 1.73 and M = 2.61, SD = 1.64 respectively), F(1, 376) = 3.31, p = .07, η2 = .01. When rating their opponent, proponents felt that opponents were much more influenced than self –reports indicated (M = 4.91, SD = 1.73 and M = 2.46, SD = 1.70 respectively), F(1, 376) = 3.68, p = .055, η2 = .01.
Opponents perceived a smaller (though similar) attitudinal gap between partisan groups due to general political orientation than what the actual self-reports from partisans indicated. Opponents perceived the political influence gap between allies and typical proponents to be significantly larger (M = 4.26, SD = 2.00 and M = 3.86, SD = 1.83) than what the actual self-reports indicated. Opponents assumed their typical allies were significantly more influenced by political orientation than they were themselves (M = 4.26, SD = 2.00 and M = 2.46, SD = 1.70 respectively), F(1, 376) = 4.09, p = .04, η2 = .01, thus placing themselves further inside of the perceptual continuum than their own group. In a similar manner to proponents, opponents felt that their adversaries were much more influenced than self-reports indicated (M = 3.86, SD = 1.83 and M = 2.61, SD = 1.64 respectively), F(1, 376) = 3.87, p = .049, η2 = .01.
Taken together, results from the analysis regarding political ideology support the manifestation of the two primary response patterns associated with naïve realism, false polarization and egocentric bias, in partisan views regarding UC. More specifically, the patterns of proponents’ and opponents’ responses support a false polarization effect, with proponents’ responses revealing the largest gap. Proponents’ displayed an egocentric bias in that they portrayed their allies as much less influenced by political beliefs than the opponent group when compared to self-reports. Further, opponents displayed an egocentric bias in that they assumed their typical allies were more influenced by political orientation than they were themselves.
Discussion
The present research suggest that partisans in the UC controversy tend to exaggerate the degree of difference between opposing sides regarding the degree of sympathy each side experiences toward unmarried cohabiting couples. The incongruency between actual differences and perceived differences regarding sympathy manifests itself in a “false polarization” effect, a phenomenon that arises from the faulty inference that partisans’ views reflect top-down bias, rather than bottom-up reasoning from the “true facts” (Pronin, Gilovich, & Ross, 2004). This finding constitutes evidence that naïve realism operates in partisan perception in the UC debate, an issue characterized by more relational components than previous social issues that the naïve realism phenomena has been applied to.
Our findings also indicate that partisans significantly distort the bases for opposing sides’ viewpoints. Moreover, partisans’ exaggeration of allies and adversaries’ reliance on ideological considerations and political beliefs results in an “egocentric bias” effect, providing further evidence of naïve realism in the UC debate.
In a socially contentious context like the UC controversy, partisans display an extreme, united front for several reasons. One is to strengthen the perception of their vehemence of opinion or to mitigate the appearance of lack of conviction that ambiguity may indicate to opposing parties (Ross & Ward, 1995). Second, partisans may believe that they are “publicly displaying their private ambivalence” (Sherman et al., 2003, p. 287), yet because in reality these feelings are not seen, the individuals appear both more consistent and more extreme than they actually are (Gilovich & Savitsky, 1999; Gilovich, Savitsky, & Medvec, 1998). Miller and colleagues (Miller, Monin, & Prentice, 2000; Prentice & Miller, 1996) note that this incongruency between internal feelings and external expression can lead to a state of pluralistic ignorance – perhaps accounting for the exaggerating of differences and minimizing of commonalities exhibited between UC proponents and opponents we observed.
A core component of naïve realism is the false sense that one perceives the world through a clear lens, unfiltered by one’s expectations or motives. Based on this logic, individuals who do not share one’s perceptions are either uninformed or biased by subjective, ideological considerations (Pronin, Gilovich, & Ross, 2004). In the UC context, proponents’ responses reflect this faulty logic. Specifically, proponents rated themselves as the least influenced by ideological considerations and political orientation, followed by their allies, and finally rating their adversaries as operating under the most influence of ideological and political considerations. This pattern of egocentric bias provides evidence for naïve realism.
For opponents, the picture is slightly more complicated. The direction of the polarization was such that they perceived themselves and their allies as being more influenced by moral considerations than their adversaries. This influence is likely one that participants valued, as a high degree of ethical influence would seem a positive group characteristic. Moreover, this estimation may be backlash to the perception that those who oppose UC are inappropriately judgmental (Wieler, 2002). Thus, for a characteristic that may be viewed as positive for the group, opponents perceived a false polarization in a self-serving manner. However, when political influence (potentially a more negatively valenced variable) was examined as a basis for UC views, results reflected a pattern of false polarization and egocentric bias consistent with the tenets of naïve realism. Specifically, opponents exaggerated the influence of political orientation on both their allies and adversaries and rated themselves as being less influenced by political orientation than their adversaries.
Using the lens of naïve realism to examine the conflict over UC in the United States illuminates several aspects relevant to the resolution of this controversy. First, opposing party members who overestimate differences in their assumptions, priorities, and sympathies are likely to be overly pessimistic about the likelihood of finding common ground in their views (Clark & Marshall, 1981). In the context of the UC debate, we believe that adversaries who share their views might find more common ground than they normally perceive (Keltner & Robinson, 1993). For example, they might find they agree on a need for programs that address financial issues of couples prior to marriage, or programs designed to help couples deal with crisis situations (e.g. unexpected pregnancy). Additionally, they may learn they share an interest in creating and preserving healthy relationships and marriages. Despite the breadth of issues in which those that support and oppose UC may agree on, the discussion necessary to generate awareness of partisans’ common interests is thwarted by faulty assumptions about the strength of opposing sides’ views, as well as the bases for those views.
Previous research has shown that partisans in contentious debates tend to assume opposing parties’ interests in a negotiation are completely opposed, when in reality they may be compatible (Thompson, 1995b). Our research highlights the value of open discussions amongst opposing sides in the UC debate. These discussions may be instrumental towards conciliation between opposing sides to the extent that they offer both sides the opportunity to discuss the complexities of their viewpoints rather than simply defending their positions. Although such discussions may not necessarily lead to a change in policy regarding UC (e.g. the repudiation of extant anti-UC laws or the creation of programs designed to help unmarried couples in financial crisis) they could at least allow each side to more accurately perceive the opposing side’s viewpoint as well as the basis for their viewpoint. Moreover, such discussion may also reduce partisans’ false impressions about their adversaries and their allies (e.g. the extremity of allies’ viewpoints), this more accurate perception of adversaries and allies viewpoints may help those with more moderate views on each side feel more comfortable expressing their more temperate viewpoints. The freedom to express milder viewpoints is important as moderates within groups can play a valuable role in facilitating dialogue between adversaries and in encouraging partisans inclined towards extremism to move beyond rhetoric and position statements to underlying assumptions and concerns (Jacobson, 1981; Robinson, Keltner, Ward, & Ross, 1995).
While this research focuses on the UC controversy within the United States, our findings may also contribute to the resolution of UC controversy in other regions, particularly those experiencing relatively recent growth in UC rates juxtaposed with a generational divide in social acceptance for UC. More specifically, data suggests that UC rates in both China and Russia have increased sharply among recent birth cohorts, while social acceptance of the practice has been largely linked to age (National Public Opinion Center, 1989 as cited in Vishnevsky, 1996; Yu & Xie, 2015). The growing prevalence of UC in these countries coupled with lower social acceptance ratings among older members of the population provides fertile ground for the phenomenon of naïve realism generally—and thus these regions may benefit as well from using a naïve realism lens to illuminate patterns of false polarization and egocentric bias-- and thus opportunities for resolution of the UC debate.
Limitations and Future Research Considerations
As is true for any study, there are limitations that should be kept in mind during the evaluation of this research. Several of the limitations concern the design of the study. Although participants encountered the UC vignettes in varying order, we did not truly randomize the order in which the vignettes were presented to participants, the order of the estimates for UC partisans, or the order of the bases for the attitudes and beliefs regarding UC. Thus it is feasible that the order in which these items were presented may have affected participant responses. Additionally, data collection took place during a single session in which participants reported their attitude toward UC and then responded to four UC vignettes. It is possible that both responding to a question regarding UC status and responding to UC scenarios in the same phase may have biased participant responses toward the UC vignettes. Finally, we asked participants to report the basis for their attitudes and beliefs regarding UC, as well as speculate about those of a typical proponent and opponent of UC. In doing so, participants reported on the extent to which “general philosophical beliefs” were a factor in perceptions of UC, however this term was not defined for participants and it is possible that participants may have not fully understood this term.
Additionally, this research asked college-age students about their perceptions of UC. Some research has found that cohabitation compensates for decreasing marriage rates the least for college students (Bumpass, Sweet, & Cherlin, 1991). Thus it would be fruitful for future research to examine the cohabitative attitudes and (mis)perceptions among populations in which cohabitation has a more significant impact upon marriage rates. Additionally, college-age students tend be more progressive socially and thus research incorporating a more socially balanced sample would be useful when generalizing the study’s findings beyond college aged students.
Finally, future research may benefit from using alternative statistical analyses in the measurement of attitudes and perceptions. More specifically, it may be useful to employ Rausch-scaling in the measurement of attitudes, as well as differential-item functioning to identify the gaps or biases related to different groups or when assigning perceptions to others (e.g. Lange, Houran, & Li, 2015a, 2015b).
Research on naïve realism has shown that opposing partisans tend to overestimate the dissimilarity of their views as well as misidentify the bases for these views. These findings have been replicated in a variety of contentious social contexts, and the shortcomings in perception and judgment they document exacerbate the perceived degree of difference between opposing parties and skew each parties’ perception of the objectivity of their opposing side. The reported research supports the hypothesis that in this debate, as in many other social conflicts, the attitudinal differences may be smaller and the common ground greater than the adversaries recognize.
The accurate perception of partisan viewpoints holds important implications for resolving social conflicts. For example, Sherman et al. (2003) found that when considering negotiation with an opponent who held the views that are actually typical of the opposing side’s participants were relatively enthusiastic and optimistic about the negotiation procedure. Additionally, they were less likely to see their interests as incompatible, thought the negotiation would be more successful, and felt more open and less angry about the negotiation procedure than they did when considering the prospect of negotiating with an opponent who purportedly held the views they assumed to be characteristic of the opposing side. Given the societal implications of successful negotiation between opposing parties on UC, future research that considers partisan estimates of resolution when considering actual UC partisan views versus assumed views is encouraged.