The Importance of Identity Development, Principled Moral Reasoning, and Empathy as Predictors of Openness to Diversity in Emerging Adults

Programs that attempt to increase tolerance and openness to diverse cultures, religions, and ethnicities through education and exposure are popular and often successful in higher education, but at times backfire, leading instead to an increase in prejudice. The present study considered several intrapersonal and developmental factors that may be important to consider. Identity development (Measures of Psychosocial Development, based on Erikson’s theory of psychosocial development), moral reasoning (Defining Issues Test, based on Kohlberg’s theory of moral development), and empathy (a subscale of the California Personality Inventory) were tested as predictors of openness to diversity (Miville-Guzman Universality-Diversity Scale) with entering first-year university students (n = 282) at a small, private comprehensive university. As hypothesized, identity development, principled moral reasoning, and empathy each significantly and positively predicted openness to diversity, and the combination of predictors accounted for a moderately large portion of the criterion’s variance. Implications are discussed for developing effective programs to facilitate openness to diversity in emerging adulthood.

The effects of increasing exposure to diversity can be complex (Costandius & Rosochacki, 2012) and mixed (Jones, 1994;Killen & Smetana, 2010;Thijs & Verkuyten, 2013), however, and exposure "may sow the seeds of conflict or compassion" (Jones, 1994, p. 39), depending on the conditions in which it occurs. Positive consequences can include increasing empathy, perspective-taking, tolerance, and the celebration of various cultural traditions; but at times, prejudice and discrimination increase instead. A report prepared by the State Higher Education Executive Officers (1987) cautioned that, while lack of cross-cultural contacts can be detrimental, contacts can also be damaging if participants begin the interactions with prejudice or ignorance. As noted by Taylor (1994), "engendering tolerance is not simply a question of demographics-gathering a diverse group of people together-but is dependent on the quality of that experience" (p. 32). Multicultural campus programs may increase ethnic identity (Simmons, Wittig, & Grant, 2010) and, although feeling a strong connection to one's own group does not necessarily lead to a hatred of nonmembers (Brewer, 1999), negative intergroup attitudes have been found at times to increase as a result (Verkuyten, 2008). Although there is considerable evidence that diversity initiatives can foster cultural knowledge and interracial understanding among college students, some attempts at increasing openness to diversity solely through education and discourse can have unfavorable results (McCauley et al., 2000;Wood & Sherman, 2001).
A number of conditions have been posited by contact theory (Allport, 1954) to be important for favorable consequences, including adequate opportunity for the members to get to know each other, similarity in status, a cooperative situation, support by the institution and those in authority, and friendship (Cook, 1978;Pettigrew, 1998). Meaningful, positive relationships with peers from other groups may be relevant in that they both build positive feelings and reduce uncertainty and anxiety about interacting with diverse others. They may also promote perspective-taking and empathy. Even when the conditions noted are met, however, a metaanalysis of studies on children and adolescents in school settings indicated that intergroup contact accounts for only a medium-sized effect on negative ethnic attitudes (Tropp & Prenevost, 2008). In a review of the evidence for positive and negative consequences of interethnic contact in schools, Thijs and Verkuyten (2013) called for a consideration of individual characteristics and social and psychological processes as important predictors of tolerance.
In fact, research on a number of intrapersonal variables is pertinent to understanding the development of tolerance. As argued by Witenberg (2007), tolerance may be best conceptualized, not as the absence or control of prejudice, but as something related to morality, with concerns for justice, equality, and fairness. Examining the reasoning used by children and adolescents for their tolerant or intolerant attitudes in response to dilemmas portraying prejudicial behaviors, Witenberg found moral considerations of fairness as well as empathy for the plight of the target of prejudice to underlie most participants' tolerant attitudes. She also found that an individual's degree of tolerance and underlying reasoning tended to be cross-situational, wherein some individuals were tolerant irrespective of the circumstances described, reflecting an enduring intrapersonal attribute. Furthermore, tolerance, morality, prosociality, and aspects of identity development may be linked in their shared quest for understanding what factors allow people to move away from protecting only their self-interests and toward a perspective that may include or benefit others or society. The extensive literature on prosocial behavior and morality also points to the roles of both cognitive (Davis et al., 2004;Kohlberg, 1976;Rest, Narvaez, Thoma, & Bebeau, 2000) and affective domains (Batson, Chang, Orr, & Rowland, 2002;de Vignemont & Singer, 2006;Eisenberg, Shea, Carlo, & Knight, 1991;Hoffman, 2001). Issues related to identity have received attention as well (Blasi, 2005;Perugini & Leone, 2009), and many researchers have argued that a combination of factors must be considered (Frimer & Walker, 2008;Smetana & Killen, 2008).
The process of teaching intercultural sensitivity to college students is a complex challenge, which some have also noted should include a consideration of developmental issues (Mahoney & Schamber, 2004). For example, some social justice and diversity courses in college curricula require a certain level of developmental maturity, including critical thinking and self-reflective abilities, openness to conflicting perspectives, and perspective-taking (Adams, 2002). King and Baxter Magolda (2005) have argued that intercultural maturity in the college years should be understood as a developmental process and research supports the importance of maturity in promoting openness among college students to exposure to diversity (Kubal, Stone, Meyler, & Mauney, 2003). With emerging adulthood, comes the potential for a number of developmental strides, including new cognitive abilities (Inhelder & Piaget, 1958), corresponding advances in moral reasoning (Gibbs, Moshman, Berkowitz, Basinger, & Grime, 2009;Kohlberg, 1976;Rest et al., 2000), and an integrated sense of identity (Erikson, 1959(Erikson, , 1968. The process of facilitating positive growth in emerging adults is complex (Pike & Kuh, 2006); emerging adulthood, with its new-found potentials for maturity, may be, in fact, a particularly fruitful time to facilitate the acceptance of others.
In considering the potential roles of both cognitive and affective variables relevant to openness to others in emerging adulthood, two developmental theories are particularly pertinent: Erikson's psychosocial theory of identity development and Kohlberg's theory of moral reasoning. Accordingly, the present study focused on both identity development and the development of moral reasoning as understood by these theories. The role of empathy, as possibly "the most prototypical moral emotion and . . . certainly the most widely discussed" (Pizarro, 2000, p. 355), was also addressed. These three factors were also selected to represent both cognitive and affective domains, so that we could test the relevance of a variety of intrapersonal areas.

Identity Development
Identity has been considered from a number of perspectives (Craig-Bray, Adams, & Dobson, 1988;Hill, Allemand, & Burrow, 2010;Loevinger, 1998), but Erikson's (1959Erikson's ( , 1968 psychosocial theory of identity development may be especially applicable to understanding a new potential for openness to diversity during emerging adulthood. According to Erikson's theory, identity development is an epigenetic process wherein specific issues tend to be addressed and resolved in a predictable order across the life span. A progression through childhood stages, of trust versus mistrust, autonomy versus shame and doubt, initiative versus guilt, and industry versus inferiority, culminate, in adolescence, in a crisis involving identity versus ego confusion. Each of Erikson's stages of identity development can be seen as relevant to the evolution of tolerance and openness to diversity, but the successful resolution of identity in adolescence would seem to be especially pertinent. Erikson (1959) noted that, before a sense of identity is adequately established, adolescents become remarkably clannish, intolerant, and cruel in their exclusion of others who are "different," in skin color or cultural background, in tastes and gifts, and often in entirely petty aspects of dress and gesture arbitrarily selected as the signs of an in-grouper or out-grouper. (p. 97) According to Erikson, It is important to understand . . . such intolerance as the necessary defense against a sense of identity confusion . . . It is difficult to be tolerant if deep down you are not quite sure . . . that you really know who you are. (pp. 97-98) and without a positive resolution of this stage, one is left with " . . . distantiation: the readiness to repudiate, to isolate, and, if necessary, to destroy those forces and people whose essence seems dangerous to one's own . . . " (p. 101). Erikson continued, "This . . . repudiation is an outgrowth of the blinder prejudices which during the struggle for an identity differentiate sharply and cruelly between the familiar and the foreign" (pp. 101-102). Not all who tolerate or even embrace diversity in others do so because of a mature identity. For example, some adolescents may gravitate toward "all that is different" as they explore various aspects of themselves or as a means of rebelling against the status quo. It can also be argued that being indiscriminantly accepting of others' values may not be desirable (Shady & Larson, 2010). Being open to a respectful dialogue and serious consideration of diverse perspectives may be facilitated by feeling secure in oneself, however. We reasoned that successful resolution of the adolescent identity crisis may facilitate the young adult's ability to tolerate others' differences and foster the security helpful for understanding and appreciating others' perspectives, feelings, or needs.

Principled Moral Reasoning
Much research has been rooted in Kohlberg's developmental theory of moral reasoning (Haggbloom et al., 2002). Extending Piaget's work on morality, Kohlberg's theory (Kohlberg & Kramer, 1969) posits that increasingly sophisticated moral structures may emerge as developing cognitive capabilities interact with one's social and cultural environment. Although some have questioned the continued validity of this theory (Lapsley & Hill, 2008) and conceptualizations of prosociality and morality involve heated debate (Frimer & Walker, 2008), recognition of the utility of the Kohlbergian approach (Boom, Wouters, & Keller, 2007) and the importance of considering developmental issues in research on moral functioning (Gibbs et al., 2009) continues.
The neo-Kohlbergian approach of Rest and colleagues (2000) is well suited to studying a level of moral reasoning that, developmentally, first becomes possible shortly before emerging adulthood. This approach describes moral schemas, with development seen as involving shifting distributions of less and more mature types of moral reasoning. The three main schemas-Personal Interest, Maintaining Norms, and Postconventional-are seen as increasingly mature ways of answering "the "macro" question (of) . . . how to get along with people who are not friends, kin or personal acquaintances, i.e. how to organize society-wide cooperation" (Rest et al., 2000, p. 386). Adolescence brings a major social-cognitive advance that allows for reasoning beyond a Personal Interest approach and toward a sociocentric perspective. The Maintaining Norms schema involves appeals to a moral consensus that are based on established practice and existing authority, whereas the Postconventional schema is principled, viewing moral obligations as based on shared ideals, as reciprocal, and as open to scrutiny, debate, and tests of logical consistency. Principled moral reasoning may best reflect a perspective that involves moving away from one's own self-interest and toward one that reflects concerns with fairness, social justice, equality, and openness to diversity.

Empathy
Eisenberg and her colleagues have extensively studied the relationship between empathy and prosocial behaviors and intentions, noting that "other-oriented cognitions and emotions may foster prosocial moral judgment and vice versa" (Eisenberg et al., 2002(Eisenberg et al., , p. 1004. As asserted by Eisenberg and Eggum (2008), "altruists see strangers as fellow human beings" and "the cultivation of sympathy-especially for individuals who . . . may differ in ethnicity, race, nationality, or other ways-is critical for enhancing cooperation and altruistic behavior" (p. 69). Batson, Early, and Salvarani (1997) also noted that once empathic feelings are aroused, they have some "inertia," allowing for positive feelings to generalize to a stigmatized group as a whole and numerous other studies have found various aspects of empathy to predict prosocial behavior (Batson et al., 2003;Eisenberg, Fabes, & Spinrad, 2006;Malti, Gummerum, Keller, & Buchmann, 2009;Van Lange, 2008).

The Present Study
College students are exposed to experiences with ethnicities, religions, political views, values, and lifestyles that differ from their own. Consistent with Erikson's psychosocial theory of identity development, we predicted that resolution of the adolescent crisis involving identity and ego confusion would significantly predict openness to diversity (Hypothesis 1). In accordance with research on prosociality and moral reasoning, we predicted that principled moral reasoning (Hypothesis 2), and empathy (Hypothesis 3) and would also be significant predictors.

Instruments
Sample demographics. The Background Questionnaire assessed participants' age, gender, and ethnicity.
Openness to diversity. We chose to assess tolerance as defined positively, to reflect an interest in diverse others. As college students are typically exposed to others who differ in many ways, including ethnicity, abilities, and sexual orientation, we also selected an instrument that tapped diversity as broadly defined.
Openness to diversity was assessed by the total score on the Miville-Guzman Universality-Diversity Scale (M-GUDS). The M-GUDS was developed by Miville et al. (1999) to measure a universal-diverse orientation (UDO). Questions address attitudes regarding a variety of potential human differences, including age, gender, race, abilities, and sexual orientation, and assess the degree to which there is a recognition and acceptance of similarities and differences, both past and intended interpersonal contact with diverse others, and an emotional bond felt toward others.
The instrument consists of 45 items, rated on 6-point scales (from 1 = strongly disagree to 6 = strongly agree). Positive and negative phrasings are present to avoid a response bias; negative items are reverse-scored, and a higher score indicates higher levels of UDO.
Five judges were involved in item development and assessment and a number of studies have assessed the instrument's validity and reliability. In a series of four studies, Miville et al. (1999) found internal consistency to range from .89 to .94 and test-retest reliability for the measure at 1 to 2 weeks to be high (r = .94). Means ranged from 169.89 to 203.03 and standard deviations ranged from 23.62 to 31.27. Means and standard deviations for the present study fell within these ranges (M = 187.70, SD = 25.08).
When tested with predominantly Caucasian samples (comparable with the present sample), construct validity for the M-GUDS was supported by significant and positive correlations with other measures of openness to others, including the Perspective Taking and Empathic Concern subscales of Davis's (1980) Interpersonal Reactivity Index, the Autonomy scale of Helms's (1990) White Racial Identity Attitude Scale, Fassinger's (1994) Attitudes Toward Feminism and the Women's Movement Scale, as well as with Androgyny and Femininity scores on the Bem Sex-Role Inventory (Bem, 1974). Also as predicted by its developers, the M-GUDS correlated significantly and positively with several measures of psychological health developed by Luhtanen and Crocker (1992) and Slyter (1989) and negatively with both Troldahl and Powell's (1965) Dogmatism Scale and Hansen's (1982) Homophobia Scale. Discriminant validity was supported by low, non-significant correlations with SAT Verbal scores, the Fantasy and Personal Distress scales of the Interpersonal Reactivity Index, and Crowne and Marlowe's (1964) Social Desirability Scale.
Identity development. Identity development was assessed by the Measures of Psychosocial Development (MPD). Hawley (1988) developed this self-report instrument to assess the degree to which each of Erikson's eight stages of psychosocial development has been resolved. It consists of 112 Likerttype 5-point scales (from 1 = not at all like you to 5 = very much like you) about feelings, experiences, and selfconcepts. Each stage is represented by seven positive (e.g., willing to give and take in my relationships) and seven negative (e.g., avoid commitment to others) statements, allowing for the computation of positive, negative, and resolution (positive minus negative) scores for each stage. Hawley (1988) found test-retest reliability coefficients to range from .67 to .91 for intervals from 2 to 13 weeks and Cronbach's alpha coefficients to range from .65 to .84. The MPD correlated positively with corresponding traits on two other measures of Erikson's theory: Constantinople's (1969) Inventory of Psychosocial Development (correlations = .46-.78) and Boyd's (1966) Self-Description Questionnaire (correlations = .28-.65), supporting its construct validity.
Identity was assessed in the present study by subtracting the subscore for adolescent ego confusion (i.e., experiencing difficulty in integrating a central identity) from the subscore for identity (i.e., knowing who one is, where one is going, and what one's goals and values are). Hawley (1988) reported a mean of 7.11 (SD = 8.41) for a normative sample of 18-to 24-year-old men and a mean of 8.44 (SD = 8.60) for a normative sample of women. Means and standard deviations were comparable in the present study (M = 7.48, SD = 8.53).
Moral reasoning. The Defining Issues Test (DIT) was developed by Rest and colleagues (Rest, 1993;Rest & Narvaez, 1998) to assess moral reasoning. Based on Kohlberg's theory of moral development, the DIT is conceptualized as a device for activating moral schemas (Rest et al., 2000). Participants read six moral dilemmas and respond to a series of issue statements which provide fragments of lines of reasoning. Items that both make sense to the participant and activate a preferred schema receive a high rating and rank. Scores represent percentages of the relative importance that a participant gives to each of three modes of reasoning: Personal Interest, Maintaining Norms, and Postconventional (principled) schemas (Rest et al., 2000). Data that do not pass five reliability checks are purged.
Over 400 published articles support the test's construct validity and reliability (Rest, Narvaez, Thoma, & Bebeau, 1999). According to Rest et al. (2000), scores significantly correlate with moral education interventions, many prosocial behaviors, desired aspects of professional decision making, controversial public policy and political attitudes, education post junior high school, age, and numerous cognitive measures. Rest et al. reported that the DIT is equally valid for males and females, and that DIT scores predict seven validity criteria, above and beyond variance accounted for by scores of verbal ability/general intelligence or political attitude. Cronbach's alpha and test-retest reliability scores range from the upper .70s to the lower .80s.
The P score is most often used to reflect a principled orientation, generally averaging in the 30s for senior high students and 40s for college students (Rest, 1993). The mean score fell within this range in the present study (M = 33.34, SD = 16.53) and was used for assessing principled reasoning.
Empathy. The California Personality Inventory (CPI) is a 462-item, self-report inventory, which consists of 20 "folk scales." These scales were developed "to assess the kind of everyday variables that ordinary people use in their daily lives to understand, classify, and predict their own behavior and that of others" (Gough, 1987, p. 1).
In the present study, empathy was assessed by the inventory's Empathy Scale. The scale is composed of 38 true/false statements, measuring the extent to which an individual understands the feelings of others. According to its developer, Gough (1987), a high-scorer would be "comfortable with self and well-accepted by others" (p. 6) and would understand others' feelings, whereas a low-scorer would be "ill at ease in many situations; unempathic" (p. 6). A study with college students indicated internal consistency for the scale to be .58, and test-retest measures on high school students ranged from .56 to .58 (Gough, 1987). Raw scores may be converted to standard scores. A score of 21 on the Empathy Scale is at approximately the 50th percentile for both men and women. The mean for the present study was very close to 21 (M = 21.79, SD = 4.58).

Participants
Participants initially consisted of 282 entering first-year students (84% of the class) at a small, private university in southern California. No potential participants were purposefully excluded from the sample. All participants were at least 18 years of age and ranged from 18.02 to 24.17 (M = 18.57, SD = .53). Males (48.6% of the sample) and females (51.4%) were approximately equally represented. The sample was predominantly Caucasian (74.5%), followed by Mixed Race (12.8%), Hispanic/Latino (6.4%), Asian (2.8%), Native American (0.7%), African American (0.4%), or Other (2.5%) ethnicities. Two outliers (scoring over 3 SDs from the M) were omitted from the data: One male scored very low on openness to diversity and one male scored very low on identity. Data were analyzed for the resulting 280 cases.

Procedure
The study was approved by the university's Institutional Review Board (IRB), and the American Psychological Association's (2002) ethical guidelines for research with human participants were followed throughout the process. Before beginning the study, all participants signed an informed consent form detailing the procedure and notifying them of their rights as participants. The consent form explained that their participation would assist in a research project on the thinking processes and attitudes of the freshman class. They were informed that questionnaire packets were coded by identification numbers linked to their names for a voluntary follow-up study to be offered in the future. Participants were assured that all personal information would be kept confidential and that participation was entirely voluntary. All identifiers, as well as informed consent forms, were separated from the questionnaires and secured in locked files accessible only to the researchers. Participants were given contact phone numbers and email addresses of the researchers to obtain a summary of the findings and to address any questions or concerns.
University students completed the packet of questionnaires during their first-year orientation, 2 days prior to the start of the fall semester. Packets were completed simultaneously in quiet classrooms, in groups of approximately 20. The order of questionnaires in packets was counterbalanced and administration took approximately 1½ hr.

Results
An alpha level of .05 was adopted for all statistical tests. Primary statistical analyses were conducted using the Statistical Package for the Social Sciences 20.0 (SPSS 20.0, professional package). One-tailed Pearson correlations were used only for hypothesis tests; otherwise, two-tailed tests were used. Size of effect was assessed by R 2 for multiple regressions and cutoffs recommended by Cohen (1992) were applied for interpreting squared values of r (.01 for small, .09 for medium, and .25 for large). All scores were approximately normally distributed.

Hypothesis Tests
As predicted, openness to diversity correlated positively and significantly with identity, r(280) = .20, p < .001, principled reasoning, r(261) = .16, p = .005, and empathy, r(280) = .33, p < .001. In addition, it did not correlate positively with stages of non-principled moral reasoning and a two-tailed test revealed that it correlated negatively and significantly with stages of moral reasoning that reflect a personal interest orientation, r(261) = −.17, p = .007.

Our Findings and Implications for Future Research
As hypothesized, identity development, principled moral reasoning, and empathy each predicted openness to diversity. Consistent with previous research, empathy was a particularly strong predictor. These findings confirm that intrapersonal variables predict openness to diversity. They also confirm the importance of considering cognitive, affective, and developmental variables in predicting openness to diversity for emerging adults.
We had predicted that resolution of the adolescent identity crisis would be particularly important for facilitating openness to diversity. Exploratory analyses found that resolution of the next stage, intimacy versus isolation, was an even stronger predictor, especially in combination with the earlier stage of industry versus inferiority. Both stages of intimacy versus isolation and industry versus inferiority are interesting, as the first involves issues related to connecting with another person and the second, confidence in one's own competence relative to peers. Erikson (1959) posited resolving identity versus ego confusion to be important for the strength not to engage in "distantiation" and intolerance of diverse others. Interestingly, a recent study (Tadmor et al., 2012) demonstrated that a cognitive need for closure predicts stereotype endorsement, racism, and discrimination. Although their study was based on a very different theoretical framework from that of the present study, individuals in the throes of ego confusion might be expected to have a need to protect themselves from the "threat" of diverse others. Our finding for the relative importance of the next stage of development may in fact reflect how we operationalized tolerance. Rather than assessing the absence of prejudice or intolerance, our instrument (M-GUDS) measured a positive attitude of interest in exploring diversity-a characteristic which may involve issues related more to intimacy and in being open to linking with others. Promoting a positive interest in others may support an appreciation for diversity. Future research should explore whether the mere absence of intolerance involves different intrapersonal mechanisms than does embracing diversity.
Regarding the stage of industry versus inferiority, Erikson (1959) wrote, "this is socially a most decisive stage: since industry involves doing things beside and with others, a first sense of division of labor and of equality of opportunity develops at this time" (p. 93). It is also the first stage in which the child's significant relations move beyond the basic family, to include neighborhood and school (Erikson, 1959), presumably with new-found opportunities for exposure to diversity. It makes sense that successfully resolving issues related to working side-by-side with others outside the family would be conducive to openness to diversity. Further research could explore the role of childhood experiences, as well as ways to promote a sense of competence and developmental readiness to connect intimately with others.
Principled moral reasoning was found to be a significant predictor of openness to diversity and non-principled moral reasoning was not. This is also an interesting finding given how we operationalized tolerance. Other studies (Witenberg, 2007) have found moral reasoning to be important in reactions to descriptions of another's prejudicial behaviors-situations which may more clearly be tied to issues of justice, equality, and fairness. We found moral reasoning to be a significant factor in embracing differences as well, although the effect size for the variable was small. It may be worthwhile to study whether its strength as a predictor depends on how tolerance is operationalized. As noted by Witenberg (2007), understanding its role, as well as how moral reasoning progresses developmentally, may have important implications for curriculum design and social policy.
Finally, empathy was found in our study to be the strongest predictor of openness to diversity. This finding is consistent with previous research linking empathy with both the reduction of prejudice and the presence of prosocial behavior (Batson et al., 2003;Eisenberg et al., 2006;Malti et al., 2009;Van Lange, 2008), and suggests that empathy may be related to an overall interest in being with and understanding other people. Empathy may foster an identification with another person in an emotionally significant way. Further research could explore how far-reaching the effects of empathy may be.
Not only may multiple factors be involved in openness to diversity, there may be individual differences regarding which factors predominate as well. In a person-level analysis of moral exemplars, for example, Walker, Frimer, and Dunlop (2010) found that exemplary moral functioning can arise from different sources across individuals. Specifically, they found three "clusters" of moral exemplars: some heroes were strong in relational and generative motives, some were strong in sophisticated moral reasoning and self-development, and others were quite "ordinary," differing little from non-heroes. Accordingly, it may be productive to explore whether behaviors are motivated by heterogeneous factors, as opposed to universally resulting from a particular characteristic or combination of variables. Consistent with findings of Walker et al. (2010) for multiple possible motives for heroism, predictors may vary for individuals, subgroups, and situations. The sample size of the present study prohibited reliable cluster analyses to identify possible distinct pathways, but an informal review of the data revealed that a relatively high score on no single predictor was a necessary prerequisite for a relatively high score on openness to diversity. Future research may consider exploring large samples using a cluster analytic approach. Better questions than which variable is most important may be for whom are particular variables important and/or most efficiently strengthened, as well as what additional variables should be considered. Finding a single "key" to increasing openness to diversity may not be as effective as designing programs that address multiple factors and/or are tailored to individual needs.
Finally, it may be worthwhile to consider how these intrapersonal factors can be developed. As noted by Butrus and Witenberg (2013), "There is a consensus that pro-social attitudes and behaviours of empathy, sympathy, compassion, and caring, which form the basis of tolerance, can be taught and learned" (p. 297). There is also some evidence for bidirectionality among some of these variables and exposure to multicultural experiences. Tadmor et al. (2012) found that exposure to multicultural experiences may be helpful in that they may lead to a reduction in cognitive rigidity, thereby reducing intolerance. They conceptualized this growth in terms of responses to the cognitive disequilibrium and repeated experiences with dissonance which led individuals to become "epistemically unfrozen" (p. 753). Experience may be helpful by what it teaches a person about others, by resulting cognitive changes, and/or by changes in feelings toward others. The interplay among multicultural experiences, identity development, moral development, empathy, and other factors may be bi-directional. Developmental readiness may be important to consider, however, to minimize the chances of destructive outcomes.

Limitations of the Study
A number of limitations are important to address. We selected only three factors to explore as potential predictors of openness to diversity. Our intent was not to address all relevant factors, but rather to demonstrate that developmental factors may be important to address, as well as both cognitive and affective domains. Other factors are also undoubtedly influential, including such intrapersonal or interpersonal variables as past and current experiences with diverse others, family attitudes (Castelli, Zogmaister, & Tomelleri, 2009;Stringer et al., 2010;White & Gleitzman, 2006), or even issues related to temperament (Davis, Luce, & Kraus, 1994). Eisenberg and colleagues (Eisenberg & Eggum, 2008), for example, have found that level of personal distress may impair one's ability to be empathic toward others.
Each of our factors could also be assessed by other measures. Our instruments relied on self-report, which can give valuable information about feelings, attitudes, and experiences, but are limited to information about which a participant is both aware and willing to share. We believe that our measure of openness to diversity is particularly rich in that it taps attitudes regarding multiple aspects of diversity, feelings toward diverse others, and both past and intended behavioral contact. Sound alternatives exist for measuring our constructs, however, and findings should eventually be tested with direct behavioral measures. Distinctions between lack of intolerance and positive intergroup attitudes should also be explored (Pittinsky, 2009).
It should also be noted that our sample was limited to undergraduates entering a small, private university in southern California. We were able to include most of our entering first-year students; however, our students may have differed from other emerging adults in the general population or even at other universities. Further research is needed to confirm our findings with other samples. Furthermore, a larger sample would allow for cluster analyses to explore the possibility of multiple pathways to openness to diversity.

Conclusion
We found evidence for the relevance of a developmental approach, as well as for a number of specific intrapersonal variables which included both cognitive and affective factors. Identity development, principled moral reasoning, and empathy each significantly predicted openness to diversity in our sample of emerging adults. Continuing to define specific situational features that support and elicit desired behaviors remains important future work-both exposure and education have been found to be effective in building acceptance of diverse cultures, religions, and ethnicities. Including a consideration of developmental and intrapersonal variables may be important as well, however, both for increasing the effectiveness of such programs and for reducing the likelihood of adverse effects. In fact, it may be fruitful to explore ways in which intrapersonal variables and situational factors may benefit each other in a bi-directional way or even interact, to maximize the effectiveness of future programs.
Definitions of tolerance vary, from "putting up with" others, to being genuinely interested in understanding and knowing about them, and the strongest predictors may depend somewhat on the definition used. Our measure of openness to diversity assessed the degree to which participants accepted similarities and differences, had both past and intended interpersonal contact, and felt an emotional bond with others who differed from them in a variety of ways, including age, gender, race, abilities, and sexual orientation. Although resolution of identity versus ego confusion significantly predicted our measure of openness to diversity, resolution of the next stage of intimacy versus isolation was an even stronger predictor, and principled moral reasoning was a significant but relatively weak predictor. A mere absence of intolerance, however, may be more closely linked to resolution of the adolescent identity struggle as well as to concerns about morality. Research should explore further which predictors are strongest, based on an intervention's goals.
We also found suggestive evidence for multiple possible pathways. Might some individuals be open to diversity because of feelings of empathy, others because of what they have reasoned to be fair, and others because they have enough security in their own identity that they are not threatened by differences? Considering how to assess an individual's developmental readiness and intrapersonal strengths or needs may be helpful for creating successful programs.
Tolerance and mutual respect are crucial for peaceful coexistence and success in today's multicultural and global society. Institutions of higher learning are attempting to address the challenge with programs to build tolerance. Emerging adulthood, with its new-found potentials for maturity, is a particularly fruitful time, developmentally, to address the acceptance of others, and understanding the complex array of factors involved in promoting openness to diversity may be helpful for developing the most effective programs possible.

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 and/or authorship of this article.