Service-learning experiences have the potential to improve participants’ attitudes and values toward those whom they serve, but if the experience is poorly designed or poorly implemented, it runs the risk of reinforcing stereotypes and deficit perspectives of the intended beneficiaries of service. This study examines the extent to which Contact Theory predicts the efficacy of service-learning courses in promoting positive attitude change among participants. Contact Theory stipulates the conditions under which attitude change toward an “out-group” becomes possible. Comparing pre-test and post-test scores for 220 students enrolled in service-learning courses in two different institutions, we find that courses that reflect more tenets of Contact Theory are more effective than those less aligned with Contact Theory in reducing students’ overall colorblindness and improving their awareness of blatant racial issues.

Many who practice service-learning have the goal of affecting participants’ attitudes and values, and research suggests that service-learning has the potential to impact learners in ways other forms of teaching may not (Delve, Mintz, & Stewart, 1990; Holsapple, 2012). Service-learning has been found to be associated with a host of positive outcomes, including greater sensitivity and empathy (Bernacki & Jaeger, 2008; Wilson, 2001); increased commitments to social justice (Eppler, Ironsmith, Dingle, & Erickson, 2011; Fenzel & Dean, 2011; Simons, Blank, Fehr, Barnes, Georganas, & Manapuram, 2012); improved cultural competence or multicultural skills (Einfeld & Collins, 2008; Meaney, Bohler, Kopf, Hernandez, & Scott, 2008); and stereotype reduction (Conner, 2010a; Meaney et al., 2008; Wright, Calabrese, & Henry, 2009). However, previous research also cautions that when poorly implemented, service-learning may result in unanticipated outcomes, such as increased prejudice and bias on the part of learners toward the very groups intended to benefit from their service (Erickson & O’Connor, 2000; Erickson & Santmire, 2001; Hollis, 2004; Jones, 2002; Kendall, 1990; Sperling, 2007). Those attempting to implement service-learning in their classrooms may actually do more harm than good if they engage students in service-learning experiences that afford casual contact; that is, contact between groups that is short-term, superficial, and lacking deep mutual engagement (Erickson & O’Connor, 2000; Erickson & Santmier, 2001; Houshmand, Spanierman, Beer, Poteat, & Lawson, 2014). According to Allport (1984), “Such evidence as we have clearly indicates that such contact does not dispel prejudice; it seems more likely to increase it” (p. 263).

Changing attitudes and beliefs involves substantial reflection on one’s values as well as the acquisition of new knowledge and skills (Holsapple, 2012). Lasting attitude changes are very difficult to achieve, and even if the service experience is carefully designed and well implemented, change may not occur for participants. Although numerous studies have examined the effects of service-learning experiences on participants and uncovered some of the design features that seem to promote desired outcomes, such as the amount of time participants are engaged in service and the number of structured opportunities for reflection, the field of service-learning could benefit from more rigorous, large-scale studies that investigate the conditions under which desirable attitude change among participants does transpire. This study responds to that need.

Literature Review

How Participation in Service-Learning Affects College Students

Two recent literature reviews have affirmed the host of positive outcomes associated with service-learning experiences. In their meta-analysis of 62 studies in which service-learning participants were compared to control students, Celio, Durlak, and Dymnicki (2011) found significant gains for service-learning participants in five areas: attitudes toward self; attitudes toward learning; academic performance; civic engagement; and social skills. Holsapple (2012) examined 55 studies that explored students’ reports of the effects of service-learning courses, and found that 32 of the studies documented findings related to students' confrontation of stereotypes they held about the diverse populations with whom they worked. This stereotype reduction was typically facilitated by the development of positive relationships with the served populations. Another key theme – increased knowledge about the served population – was found in 28 of the studies. Students reported greater awareness of cultural backgrounds, marginalization, and differences among individuals of a specific population. Additionally, students who participated in service-learning reported an enhanced appreciation of cultural diversity upon completion of the course, particularly among participants who worked with international and immigrant populations and K-12 school students.

Service-learning courses have demonstrated positive effects on students’ prejudices, understanding of social problems, and commitment to social change. These effects are strengthened when developed in the context of high-quality instructional methods and strong, supportive relationships with site-supervisors (Batchelder & Root, 1994). Indeed, a study conducted by Chang, Anagnostopoulos, and Omae (2010) sought to examine the effects of service-learning site characteristics, such as the quality of supervision students received, the type of organizations with which they worked, and the type of academic instruction they received, along with the degree of rigor of the reflection in which they were expected to engage. The particular course under examination, a multicultural service-learning course, was designed to facilitate interdependent partnerships and collaboration as well as critical thinking about social injustices. The study focused on 18 sections of this course, which engaged 212 students in 22 different service sites. After one semester, Chang et al. found that students who reported receiving more support from their site supervisors also reported a greater awareness of their personal biases and a reduction in previously held stereotypes toward low-income youth of color. Similarly, students who participated in courses with instructors who facilitated intense discussions surrounding the students’ service-learning experiences and issues related to social inequities reported an increased consciousness of their prejudices and previously held stereotypes, a greater comprehension of cultural diversity and systemic inequities, and a dedication to working with disadvantaged populations in the future. In addition, the results indicate that forging positive relationships between the university and service sites as well as between the participants in the service-learning course and the populations with whom they serve are integral components of effective service-learning pedagogy. In addition to these pedagogical factors, Chang et al. found that white students became more aware of their biases over the course of the semester than did students of color. While the results of this study are informative, a major limitation is that the data were only collected from students at one university. Further studies are needed to replicate these findings to determine whether the results may be generalized to other populations of students participating in service-learning courses.

Forming interdependent relationships between individuals participating in service and the individuals with whom they are engaging is widely understood as crucial to effective service-learning (see review by Murphy & Rasch, 2008). Without the formation of interdependent relationships and guided reflection, service-learning experiences may strengthen perceptions of privilege and reinforce stereotypes, just as some programs that attempt to reduce the stigma of a particular group end up reinforcing it (Hinshaw & Stier, 2008). Houshmand et al. (2014) illustrate the potential harmful effects of service-learning experiences in their study of a service-learning course in which students were expected to collaborate with community partners from East St. Louis, a nearby predominantly African-American community, with nearly half of the population living below the poverty line. While students did not receive any direct instruction related to cultural diversity or whiteness, some course readings addressed multiculturalism, and students participated in some activities to increase their awareness of the community's cultural background, such as taking tours of the city to learn more about its history. Students made three site visits, each lasting two days, and engaged in various activities during these site visits, including conducting site analyses, holding individual meetings with community members, attending community meetings, and participating in service activities within the community. Houshmand et al. found, that upon completing the course, students confirmed the negative misconceptions they held regarding East St. Louis and its community members, students' experience as the "minority" in East St. Louis caused them discomfort as they perceived their appearance brought them extra unsolicited attention, and students attributed community problems to the fault of the residents of East St. Louis.

As evidenced by Chang et al. (2010) and Houshmand et al. (2014), certain factors may facilitate or hinder the success of service-learning in influencing positive attitude change among participants. To gain a deeper understanding of how students’ attitudinal changes occur as well as their personal views on their experience working at their service-learning sites, researchers have analyzed in-depth student reflective writing, journal entries, interview data, and focus group transcripts (e.g., Baldwin, Buchanan, & Rudisill, 2007; Boyle-Baise & Kilbaine, 2000; Endres & Gould, 2009; d’Arlach, Sánchez, & Feuer, 2009; Greene, 1998; Hollis, 2004; Nenga, 2011; Simons, Fehr, Hogerwerff, Blank, Georganas, & Russell, 2011). Deely (2010) conducted a study to fill the gap in research on the effects of service-learning from the students' perspective. Analyses of the data revealed several themes indicating students' positive and negative development throughout the service-learning experience. Asked to identify salient contextual and learning environment features, students indicated that their critical thinking skills were enhanced through coursework, community service, dialogue with small groups, analysis of readings, and journal writing. Moreover, students reported that having small class sizes made open dialogues easier and more comfortable, which facilitated honest discussions regarding personal experiences.

Limitations of Existing Research

As noted by Holsapple (2012), though qualitative measures provide insight into students’ experiences and beliefs regarding social justice, an overemphasis on reflective measures in service-learning research makes it difficult to evaluate in a quantifiable, statistically rigorous way the effects of participation in service-learning for different students. Specifically, reliance on reflective writing strategies as the sole means of data collection may limit the validity of the findings, as such measures may influence the nature of participants’ responses, leading the responders to be vulnerable to response bias or responses driven by emotion (Hobbs, 2007). Additionally, research that has assessed the degree of reflection in student journals across an array of disciplines in higher education has found minimal, if any, consistency across methodological strategies for evaluating students’ reflective writing (Dyment & O’Connell, 2011).

Similarly, research conducted by instructors of service-learning courses may lead students to respond in a particular way. Students who completed reflective writing prompts as part of a course assignment (e.g., Endres & Gould, 2009) may respond differently than if they had be given an anonymous survey that was not viewed as a course requirement (e.g., d’Arlach et al., 2009). While qualitative measures provide researchers with meaningful insight into students’ experiences in service-learning, sense-making, and views of social injustices before and after their participation in the course, supplemental quantitative measures, administered by researchers not associated with the course, may serve to limit response bias and enhance validity. Indeed, a hybrid or mixed methods approach may offer the most promising research design, as exclusively quantitative studies run the risk of overlooking or over-simplifying the complex ways in which participants make meaning from their experiences (Snow & Reeb, 2013).

In addition to the methodological limitations described above, studies of service-learning have been critiqued for their lack of theoretical frameworks to guide the research (Whitley, 2014). Without a guiding theory, it may be difficult to understand the exact mechanisms that contribute to or hinder the social and academic effects of service-learning. Indeed, a strong theoretical framework has been identified as “the most critical part of the research plan” (Ennis, 1999, p. 132) as it provides boundaries and guidelines to help develop the methods and interpret the results of the study. Without a conceptual focus, it may be difficult for researchers to pinpoint which components of the service-learning experience were most impactful for students (Moely, McFarland, Miron, Mercer, & Ilustre, 2002).

This study responds to the call for service-learning research that demonstrates empirical and methodological rigor and builds from a strong theoretical base in its investigation of the conditions under which service-learning courses facilitate authentic and desirable attitude change among participants. For more than 60 years, Gordon Allport's (1954) Contact Theory (CT) has been used by social psychologists to inform the design of strategies aimed at changing attitudes through human contact as well as evaluate the effectiveness of these strategies; however, few studies of service-learning have drawn on CT (for exceptions, see Conner 2010a; Murphy & Rasch, 2008). Using CT to frame an analysis of the effects of service-learning courses at two institutions of higher education, this investigation examines to what extent service-learning courses promote changes in learners' attitudes and which design features of the service-learning experience are most associated with such changes.

Theoretical Framework

Contact Theory, also called the contact hypothesis, was introduced and developed by social psychologists to understand and evaluate the various conditions under which face-to-face contact would promote greater personal and social understanding between members of different groups (Allport, 1954; Amir, 1969; Cook, 1985). At its core, CT traces the minimum necessary conditions through which favorable experiences with individual members of an out-group may be transmitted or generalized to one’s group-related attitudes (Pettigrew, 1988; Rothbart, 1996). The five main conditions of CT as outlined by Allport (1954) in his classic book, The Nature of Prejudice, include equal status contact; pursuit of common goals; intergroup cooperation; support of authorities, custom, or law; and sustained, long-term contact.

Since Allport first introduced his theory, researchers have extended and elaborated upon this original formulation. This new research has highlighted two additional factors that play an important role in facilitating attitude change for participants. These factors include instructional designs that attend to participants’ affect, particularly students’ anxiety about entering an unfamiliar environment and working with individuals with whom they lack previous experience (Pettigrew & Tropp, 2000; 2005), and designs that ask participants to reflect on their identity and values (Celio, Durlak, & Dymnicki, 2011; Fenzel & Dean, 2011; Gaertner & Dovidio, 2000; Kenworthy, Turner, Hewstone, & Voci, 2005; Simons et al., 2011). Research using CT suggests that these seven features of the learning environment represent the minimum conditions under which attitude change becomes possible. Although a meta-analysis (Pettigrew & Tropp, 2006) across 713 independent samples raises some question about the necessity of these conditions, it does find “those samples that experienced carefully structured contact situations designed to meet Allport’s optimal conditions achieved a markedly higher mean effect size than did other samples” in terms of prejudice reduction (p. 766).

In addition to prejudice reduction, CT has been profitably applied in research on stigma reduction. Anti-stigma programs involving direct contact between the general public and individuals with mental illnesses have been found to produce lasting changes in both behaviors and attitudes when the conditions of CT are met (Corrigan & O’Shaughnessy, 2007). Interventions that comport with the conditions of CT also have been found to reduce stigma and myths about homelessness and promote positive attitudes and greater empathy toward people experiencing homelessness (Snow & Reeb, 2013). Although CT has been effectively utilized in the study of anti-stigma programs, very limited work considers how CT might be applied to the context of service-learning.

To illustrate how CT might be used to inform the design of a service-learning course, we offer two hypothetical examples of courses that show different degrees of alignment with the tenets of Contact Theory. For the sake of the examples, we presume that both courses are offered at the same institution, and that this institution demonstrates strong support for service-learning, thereby meeting Allport’s condition of support by authorities or institutional support.

“Understanding Poverty in America” is a large, lecture-based course in which students sign up to volunteer for one five-hour shift at a soup kitchen. The instructor does not provide an introduction or orientation to the soup kitchen for the students; instead, she gives them a handout with instructions about where to go, what to wear, and how to help the staff. In addition, students are expected to participate in a class “Day of Service” by volunteering to help with administrative tasks at a local homeless shelter or a welfare agency. The students are asked to draw on these experiences in their final exam; however, class time is devoted to lecture, and there is no opportunity during the semester to discuss or write about these experiences in greater depth. This course does not satisfy any of the following conditions of CT: equal status contact; pursuit of common goals; inter-group cooperation; long-term or sustained contact; or efforts to address participants’ affective states. The course meets CT’s condition of opportunities for personal reflection at a very low level.

“Urban Youth and Their Institutions” is a small, discussion-based course in which students spend two hours each week over the course of a semester at an under-resourced public high school where the college and high school students partner to develop design proposals for the vacant lot across the street from the school. As they are members of the community, the high school students possess local knowledge useful to completing this assignment, and the instructor is careful to draw attention to their expertise. Prior to their first meeting with the high school students, the college students are given an orientation by the course instructor and the high school principal who take them on a tour of the high school and spend time answering their questions. The high school students also meet with the university course instructor and the principal ahead of time to learn about the project and to have their questions answered. The service-learning course requires the college students to respond each week to prompts that engage them in personal reflection that synthesize the course readings and their service experiences, and the instructor sets aside class time for the students to discuss and make meaning about their experiences at the high school. In its design, this course satisfies the following conditions of CT: equal status contact; pursuit of common goals; inter-group cooperation; long-term or sustained contact; opportunities for personal reflection; and efforts to address participants’ affective states. These two hypothetical examples of college courses illustrate how service-learning courses may demonstrate either high or low alignment with basic CT tenets.

Research Questions and Hypotheses

Based on extant research, we developed linked hypotheses for each of our three research questions. First, we asked if students enrolled in service-learning courses demonstrate significant change in their attitudes toward racial minority groups, specifically with regard to colorblindness? Colorblindness is a viewpoint that holds that race and racial categories are no longer salient, and therefore these categories should be ignored and everyone should be seen as an individual (Richeson & Nussbaum, 2004). We expected to see a statistically significant decline in overall colorblindness across the sample of students enrolled in service-learning courses; however, we also expected that the extent of change would vary across courses, with some courses showing strong mean levels of change and some showing no or little mean change. Second, we asked to what extent individual student characteristics might explain their degree of change in colorblindness. Here, we expected that students of color would experience less change than white students. We did not expect to see differences based on any other individual characteristics, such as class year or gender. Third, we asked to what extent a course’s alignment with CT predicts students’ degree of change in colorblindness. We hypothesize that the more a course is aligned with Contact Theory, the greater the change in colorblindness students would demonstrate, becoming less colorblind and more aware. We expected that each tenet of Contact Theory would be associated with positive attitude change, with no one tenet exerting a stronger or weaker influence than any other.

Method

Research Design

This study is designed to investigate the utility of CT in predicting attitude change among service-learning participants. The study uses a pre- and post-test design. Other data sources included an instructor assessment of the course and course artifacts, such as syllabi and handouts, which were used to evaluate each course’s alignment with CT, as explained below.

Participants

Participants included instructors and students in all service-learning courses (N=13) offered during the spring semester at two mid-sized, private institutions of higher education, one located on the east coast (n=7) and one in the midwest (n=6). In one of the institutions, many of the students involved in service-learning courses also lived in a sophomore service-learning community, which required participation in at least one service-learning course a semester and involvement in a weekly reflective discussion group comprised of students across different courses and facilitated by a trained staff member. The 13 service-learning courses involved in this study spanned a range of disciplines from the humanities to social sciences, including Theology, Ethics, Sociology, Education, and Political Science. Class sizes ranged from nine students to 32 students. The service arrangement in each course also varied, with some instructors requiring students to serve together at the same site and other instructors offering students a range of pre-approved sites at which students could complete service hours. These sites included public schools and community centers, where college students mentor and tutor younger students; prisons and adult literacy centers, where college students work with adults to develop their reading and writing skills; and homeless shelters, where students prepared and served meals to temporary residents.

Of the 243 students in all 13 courses, a total of 220 students completed both surveys, for a 91% response rate. Fifty-five percent of the sample attended the east coast university and 45% attended the midwest college. Eighty-one percent of the sample reported that they were engaged in service-learning for their course. Students enrolled in the courses who were exempted by their instructors from fulfilling the service-learning requirements were included in the study, and this gave us a chance to examine whether or not their results differed from those engaged in service. The demographic makeup of the participants was as follows: 34% were males, 66% were females; 74% identified as white, 7% as Black or African-American, 11% as Asian or Pacific Islander, 6% as Latino/a or Hispanic, and 2% as mixed ethnicity or other; and 50% were sophomores, 38% were seniors, 4% were juniors, and 8% were Master’s students; the sample did not include any first-year students.

Procedure

After securing IRB approval for the study at both institutions, a two-pronged recruitment process was used to engage participants. First, the directors of service-learning at each university contacted instructors who used service-learning in their courses and invited them to participate in the study. The invitation letter and consent forms explained that the study would examine students’ attitude change over the course of the semester. It otherwise ensured that the instructors were blind as to the focus of the study. Participation entailed completing a short survey about their course and agreeing to allow a researcher to come into their class to explain the study, distribute consent forms, and administer a baseline and follow-up survey. The baseline survey was administered during the first or second week of each course. Rather than putting their names on the surveys, students were instructed in the process of inventing a unique identifying number. The researcher returned during the last week of each course to administer and collect the post-surveys. Both the researcher and the survey instructions encouraged students to be as open and honest as they could be, assuring them that “there are no right or wrong answers.” Because the course instructors did not see the student survey, they were not influenced by its content in how they taught the course. And because the instructors did not see the students’ responses, and because the students’ names did not appear on the surveys, any pressure students might have felt to respond in socially desirable ways was mitigated.

Measures

In addition to demographic questions, the survey included the Color-Blind Racial Attitudes Scale (CoBRAS), which measures social attitudes toward racial and ethnic minorities (Neville, Lilly, Duran, Lee, & Browne, 2000). CoBRAS has been used extensively in studies of higher education (e.g., Lewis, Neville, & Spanierman, 2012; Worthington, Navarro, Loewy, & Hart, 2008), and validation studies have found that higher scores on each of the three CoBRAS factors and the total score are related to greater racial and gender intolerance, racial prejudice, and belief that the world is just (Neville, Lilly, Duran, Lee, & Browne, 2000). The scale consists of 20 items loading onto three factors: Unawareness of Racial Privilege, Unawareness of Institutional Discrimination, and Unawareness of Blatant Racial Issues. Representative items that load onto the Unawareness of Racial Privilege factor read: “Everyone who works hard, no matter what race they are, has an equal chance to become rich;” and “Race plays a major role in the type of social services (such as type of health care or day care) that people receive in the U.S.” The latter item is reverse-coded. Sample items for the Unawareness of Institutional Discrimination factor are: “It is important that people begin to think of themselves as American and not African American, Mexican American or Italian American” and “White people in the U.S. are discriminated against because of the color of their skin.” Sample items for the Unawareness of Blatant Racial Issues are: “Racism may have been a problem in the past, but it is not an important problem today” and “Talking about racial issues causes unnecessary tension.” Each factor also includes more than one reverse-scored item. Reliability analyses for the current study yielded Cronbach’s alphas on the pre- and post-surveys ranging from .71 to .85. Change in these three factors and in overall colorblindness scores from pre- to post-surveys constitute the dependent variables of this study.

Analytic Approach

To assess the degree to which each course satisfied the seven tenets of Contact Theory, we relied on an instructor survey, course artifacts, and follow-up instructor interviews. Each course instructor completed a survey that contained several questions about the design of the course and its approach to service-learning. These questions mapped onto the CT tenets by asking about the extent to which the instructor experienced institutional support for the service; designed service experiences that were spread out and sustained over the course of the semester; attempted to reduce status disparities between the server and the served; designed the experience around shared goals for the server and the served; encouraged intergroup cooperation between the server and the served; attempted to address participants’ anxiety and concerns about service; and attempted to rouse a sense of personal identity among students in the course. In addition, syllabi and relevant course handouts were coded by two researchers at each institution (the co-principal investigators and their trained graduate students) for the extent to which these artifacts reflected these seven tenets of CT. In coding these documents, we used a protocol (see Appendix A) to determine the extent to which each condition of CT each course met. We then compared instructor survey data with our ratings, and used decision rules and in some cases instructor interviews to reconcile any discrepancies. Following the collection of all survey data, we interviewed six of the course instructors as a means of member-checking.

For each course, pre- and post-surveys were matched according to the unique identifying numbers students had created. Pre- or post-surveys without a match were dropped from the analysis. Other missing data were handled through a process of listwise deletion. Data were analyzed using inferential statistical tests, including paired t-tests for comparing pre- and post-scores; independent t-tests for comparing post-scores by groups, such as gender and service participation; two-way ANOVAs; and hierarchical linear regressions.

Results

To address our first research question, we compared the pre- and post-test scores of all students and found statistically significant declines in overall colorblindness, t(202) = 3.71, p < .000, as well as on two of the three factors for all participants: unawareness of racial privilege, t(202) = 2.70, p < .01, and unawareness of blatant racial issues, t(202) = 2.98, p < .01. In addition, as we expected, we found that mean change scores varied considerably across courses (See Table 1).

Table 1. Trends in Mean Change Scores across Courses
Number of courses with negative change (students become less aware/more blind)Number of courses with no changeNumber of courses with positive change (students become more aware/less blind)
RangeRange
Unawareness of Racial Privilege319
-.14 to -.26.06 to .58
Unawareness of Blatant Racial Issues3010
-.01 to -.17.10 to .35
Unawareness of Institutional Discrimination508
-.02 to -.17.04 to .49
Overall Colorblindness2011
-.02 to -.11.04 to .33

Before turning to our second research question about how differences in degree of change from pre-post were related to students’ individual characteristics, we first examined whether there were any significant baseline differences in pre-test scores based on student characteristics. While there were no significant differences by year in school, white students were significantly more colorblind than students of color in the pre-test, posting significantly higher levels of unawareness on each factor and on overall colorblindness, t(213) = -3.48, p < .001. White students showed greater unawareness of racial privilege, t(135) = -1.96, p < .05, unawareness of institutional discrimination, t(213) = -4.58, p < .000, and unawareness of blatant racial issues, t(213) = -2.24, p < .05 than their non-white counterparts at the baseline. Males were also significantly more colorblind than females at the baseline, on overall colorblindness, t(213) = 2.61, p < .01, on unawareness of institutional discrimination, t(213) = 2.61, p < .01, and on unawareness of blatant racial issues, t(213) = 3.47, p < .001. Finally, those students who did not plan to do service for the course were significantly more colorblind than their counterparts who did intend to do service at the baseline, on overall colorblindness, t(212) = -1.95, p < .05, on unawareness of institutional discrimination, t(212) = -2.15, p < .05, and on unawareness of racial privilege, t(212) = -2.24, p < .05.

We then considered differences in the degree of change pre-post by these same characteristics. We found no statistically significant differences based on students’ year in school or race and ethnicity. Males showed significantly more change than females on unawareness of blatant racial issues, t(201) = 2.5, p < .01 and on overall colorblindness, t(201) = 2.18, p < .05. Two-way repeated measure ANOVA tests, using pre- and post-tests and comparing those engaged in service for the course to those exempt from service, showed a significant main effect for time (pre- to post-test) on colorblindness, F(1, 202) = 7.836, p = .006, ƞ2 = .038 (See Table 2). Both those students who did service for the course and those exempt from service declined in overall colorblindness over the course of the semester; however, the interaction between service participation and time was not significant.

Table 1. Mean Level of Colorblindness by Time and Service Participation
Engaged in Service for CourseMeanS.D.
Colorblindness Pre-testYes2.84.68
No3.09.80
Colorblindness Post-testYes2.74.69
No3.00.87

Turning to our third research question and the regression analyses, we found that when we controlled for demographic factors and pre-test scores, courses with higher alignment with CT conditions did predict lower levels of overall colorblindness on the post-test, consistent with our hypothesis. In addition, we found higher alignment with CT conditions to be associated with lower levels of unawareness of blatant racial issues on post-tests (See Table 3). Not doing service in the course was associated with higher levels of unawareness of institutional discrimination on the post-test, and being part of a service-learning community was associated with lower levels of colorblindness and lower levels of unawareness of blatant racial issues on the post-test.

Table 3. Regression Results for Overall Colorblindness and Each Factor on Post-Test
Colorblindness CoefficientUn RP CoefficientUn ID CoefficientUn BRI Coefficient
Gender (0 = male)-.07-.02-.05-.12
Year-.05.08-.03.07
Service (0 = yes).05-.08.18*.02
Service-Learning Community (0 = yes).20*.14.11.27**
Race and Ethnicity (0 = non-white).07.02.13.02
Pre-test.70****.77****.38****.48****
Course Contact Theory-.17*-.10-.09-.25**
F16.23****21.55****3.87***6.03****
Adjusted R2.51.59.17.26
Note. * p < .1; ** p < .05; *** p < .01; ****, p < .001. Un RP indicates Unawareness of Racial Privilege; Un ID indicates Unawareness of Institutional Discrimination; Un BRI indicates Unawareness of Blatant Racial Issues.

Finally, analyzing each tenet of CT separately and controlling for gender, students' racial minority status, and whether or not he/she participated in service, we found that certain tenets of CT appear to play more important roles in colorblindness reduction than others. Specifically, cooperation between students and those with whom they served and institutional support for the service each generates significant associations with the degree of change in students’ CoBRAS results from pre to post-test. Whether or not the course encouraged cooperation between students and those with whom they served is significantly associated with change in unawareness of institutional discrimination, B = .32 t(201)=2.72, p < .01 and change in overall colorblindness, B = .236 t(201)=1.99, p < .05, even when all other CT tenets are included in the model (See Table 4). Furthermore, when added independently through a stepwise regression, cooperation explained a significant portion of the variance in change in institutional discrimination scores, R2 = .05, F(10, 191) = 1.97, p < .05. A second CT tenet, the level of institutional support for the service activity, is significantly associated with change in unawareness of blatant racial issues, B = .25 t(201)=1.71, p < .01 and when added independently through a step-wise regression, institutional support explained a significant portion of the variance in change in blatant racial issues scores, R2 = .05, F(10, 191) = 2.15, p < .05. For the three outcome variables mentioned above (unawareness of institutional discrimination, unawareness of blatant racial issues, and overall colorblindness), gender was also a significant predictor; being female was consistently associated with less pre-post change than being male.

Table 3. Regression Results for Change in Colorblindness and Each Factor between Pre-Test and Post-Test
Colorblindness Change Score CoefficientUn RP Change Score CoefficientUn ID Change Score CoefficientUn BRI Change Score Coefficient
Gender (0 = male)-.176**-.014-.15**-.19***
Race and Ethnicity (0 = non-white).089.043.029.112
Service (0 = yes).005.049-.051.003
Institutional Support.095-.17.172.250*
Spread Out.163.064.286-.059
Reduce Status Disparity-.084.106-.305.026
Common Goals-.175-.157-.034-.131
Cooperative Work.236**.116.319***-.015
Concerns Addressed-.286-.069-.418-.062
Reflection on Identity.221.03.213.206
F1.55.811.97**2.15**
Adjusted R2.027.009.046.054
Note. * p < .1; ** p < .05; *** p < .01; ****, p < .001. Un RP indicates Unawareness of Racial Privilege; Un ID indicates Unawareness of Institutional Discrimination; Un BRI indicates Unawareness of Blatant Racial Issues.

Discussion

Our findings affirmed some of our study’s hypotheses, while challenging others. First, as expected, we found that on average, students who were enrolled in service-learning courses, regardless of whether or not they were actually engaged in the service activity, demonstrated improved attitudes, showing declines in colorblindness over the course of the semester. Nonetheless, as we predicted, some courses were more effective at promoting positive attitude change than others. Though limited to a small number of cases, certain courses indeed appeared to do more harm than good by impeding positive attitude change and reinforcing or increasing students’ average colorblindness over the course of the semester.

With respect to our second set of hypotheses regarding individual differences, we were not surprised to find that students of color demonstrated significantly less colorblindness than white students at the outset of the study; however, we were surprised to find that white students did not experience more change over the course of the semester on our dependent variable than students of color. Also contrary to our expectations, we found significant results by gender. Male students came into the courses with higher levels of unawareness than their female classmates. They then experienced significantly greater change in their levels of unawareness over the course of the semester, such that by the end of the course, their scores were statistically indistinguishable from those of the females; in their end-of-term levels of awareness, males had caught up to the females. This result suggests that service-learning can have a particularly profound effect on male students’ attitudes, helping to make men more aware of institutional discrimination and blatant racial issues.

In terms of what might explain different courses’ effectiveness facilitating attitude change, the results from this study show that courses that align with more CT tenets are more successful than courses that do not reflect as many of the guiding CT principles in promoting attitudinal change among students. This result was consistent with our hypothesis that greater alignment would be associated with greater positive change. However, where we expected all CT tenets to matter equally, the results of our study draw attention to two in particular as most strongly associated with attitude change: promoting cooperation between students and those with whom they serve and receiving or communicating institutional support for the service. These findings generate important implications for the design of service-learning experiences, and suggest concrete ways in which instructors can be aided in fostering attitude change for their students. Inter-group cooperation has long been considered a hallmark of effective service-learning. It is widely understood as an important component of reciprocal relationships, which are considered the gold-standard in service-learning (Anderson & Hill, 2001; Bailis, 2000; Conner, 2010b; Donahue, Bowyer & Rosenberg, 2003; Jacoby, 2003). Comparatively less attention has been paid to the importance of institutional support in the service-learning literature. Nonetheless, institutional support has emerged as an important condition in effective anti-stigma programs (Hinshaw & Stier, 2008). Our findings underscore the salience of institutional support. Higher levels of institutional support may signal to students that the work they are doing for the course is important and valued by others beyond the professor and the beneficiaries of service. Further research is needed to understand how these messages of institutional support might influence students and the nature of their interactions with the service recipients; it does, however, seem possible that higher levels of institutional support could establish a foundation of institutional relationships upon which stronger interpersonal relationships can then be built.

Our finding about the importance of institutional support also aligns with our finding of the modest but significant association between being a member of a “service-learning community” and post-test colorblindness and unawareness of blatant racial issues scores. Even after controlling for pre-test scores, we found that those students who live with other students in a service-learning community are more likely to show lower levels of colorblindness and greater awareness of blatant racial issues on the post-test than their peers who do not live in such a community. A service-learning community may be one way an institution communicates its support for student engagement in service-learning. In addition, students in a service-learning community may have more informal opportunities to discuss, reflect on, and make sense of their service experiences and more occasions to hear from speakers who challenge them to think about issues of race and class than those who do not live in such a community. Of course, self-selection may also play a role in explaining the effects found among members of such a community.

Of the three factors that make up colorblindness, students’ unawareness of their racial privilege presented a bit of a puzzle in our results. We found that students’ levels of unawareness dropped significantly between pre- and post-surveys, meaning that they became more aware of their privilege over the course of the semester. This result is consistent with results found in other studies investigating the connection between service-learning and changes in students’ awareness of white privilege (Fenzel & Dean, 2011; Simons et al., 2012). However, where these previous studies could link this outcome to particular course features, such as personal identity reflection activities, we found no significant associations between this factor of CoBRAS and CT that could help explain declines in unawareness of racial privilege. Other contextual conditions and learning design features than those specified by CT may be responsible for these changes among our study participants, and future research is needed to resolve this inconsistency in the literature. Future research could also productively explore classroom-based orientation and content differences, which may have shaped our results, but which we did not consider in this study.

As with any study, the results should be viewed in light of the study’s limitations. First, although we were careful to check the instructor’s assessment of the course’s alignment with CT against our own coding of course artifacts before ascribing a value for each tenet in our database, our ascriptions could have been strengthened had we either used raters who were blind to the purpose of the study and who achieved a strong, consistent inter-rater reliability score or had we engaged in further triangulation, including interviews with students and service-site coordinators and observational data that could corroborate (or help us to refine) our ascriptions of value. Furthermore, for simplicity of analysis, we used dummy variables for each CT tenet; however, future research could use more graduated assessments of alignment. In addition, given the structure of our design, with students nested in courses, and courses nested in separate institutions of higher education, structural equation modeling might have been a useful analytic tool. Finally, it is worth noting that although we had a fairly robust sample in terms of its size, the sample was not very diverse, and our technique of grouping together students of color might have masked some finer-grained differences that would have emerged had we been able to compare students of different racial and ethnic backgrounds.

Despite these limitations, this study marks an advance in the field of service-learning, which has come under criticism for its lack of empirical and methodological rigor (Hironimus-Wendt & Lovell-Troy, 1999; Sperling, 2007). We build on a small but growing body of research that applies quantitative methods to the study of service-learning outcomes for students (Astin & Sax, 1998; Bernacki & Jaeger, 2008; Borden, 2007; Bowman, Brandenberger, Mick, & Smeddley, 2010; Eppler et al., 2011; Eyler & Giles, 1999; Fenzel & Dean, 2011; Hamilton & Fenzel, 1988; Simons et al., 2012). In addition to using previously validated measures, such as CoBRAS (Neville et al., 2000) to address threats to internal validity, we designed the study to include student participants in multiple courses at two different institutions, thereby responding to concerns about external validity leveled at researchers who study students in only one course or at one institution of higher education (Sperling, 2007). Finally, by using CT as the theoretical basis for this study, we respond to the critique that much service-learning research has been conducted without strong theoretical frameworks.

Contact Theory has been studied for over half a century. It cannot be considered new; however, this research suggests that the application of the theory to the service-learning field may offer new insights into the question of when service-learning works to facilitate positive attitude change among student participants. It is, after all, not only service-learning research, but also service-learning practice that has lacked a robust theoretical undergirding. Though Allport might have predicted the findings of this study sixty years ago, his words serve as a helpful reminder to those who design and implement service-learning with the goal of effecting attitude change among their students:

Prejudice may be reduced by equal status contact between majority and minority groups in the pursuit of common goals. The effect is greatly enhanced if this contact is sanctioned by institutional supports, and provided it is of a sort that leads to the perception of common interests and common humanity between members of the two groups. (1954, p. 231)

References

  • Allport, G. (1954). The nature of prejudice. Cambridge, MA: Addison-Wesley.
  • Amir, Y. (1969). Contact hypothesis in ethnic relations. Psychological Bulletin, 71(5), 319-342.
  • Anderson, J. & Hill, D. (2001) Principles of good practice for service-learning in preservice teacher education. In J.B. Anderson, K. J. Swick, & Y. Joost (Eds.), Service-learning in teacher education: Enhancing the growth of new teachers, their students, and communities (pp. 69-84). Washington DC: American Association of Colleges of Teacher Education.
  • Astin, A. W., & Sax, L. J. (1998). How undergraduates are affected by service participation. Journal of College Student Development, 39(3), 251-263.
  • Bailis, L. N. (2000). Taking service-learning to the next level: Emerging lessons form the national community development program. Springfield, VA: National Society for Experiential Education.
  • Baldwin, S. C., Buchanan, A. M., & Rudisill, M. E. (2007). What teacher candidates learned about diversity, social justice, and themselves from service-learning experiences. Journal of Teacher Education, 58(4), 315-327. doi: 10.1177/0022487107305259
  • Batchelder, T. H., & Root, S. (1994). Effects of an undergraduate program to integrate academic learning and service: Cognitive, prosocial cognitive, and identity outcomes. Journal of Adolescence, 17(4), 341-355. doi: 10.1006/jado.1994.1031
  • Bernacki, M.L., & Jaeger, E. (2008). Exploring the impact of service-learning on moral development and moral orientation. Michigan Journal of Community Service Learning, 14(2), 5-15.
  • Borden, A. W. (2007). The impact of service-learning on ethnocentrism in an intercultural communication course. Journal of Experiential Education, 30(2), 171-183. doi: 10.1177/105382590703000206
  • Bowman, N. A., Brandenberger, J. W., Mick, C. S., & Smedley, C. T. (2010). Sustained immersion courses and student orientations to equality, justice, and social responsibility: The role of short-term service-learning. Michigan Journal of Community Service Learning, 17(1), 20-31
  • Boyle-Baise, M., & Kilbaine, J. (2000). What really happens? A look inside service-learning for multicultural teacher education. Michigan Journal of Community Service Learning, 7(1), 54-64.
  • Celio, C. I., Durlak, J., & Dymnicki, A. (2011). A meta-analysis of the impact of service learning on students. Journal of Experiential Education, 34(2), 164-181.
  • Chang, S., Anagnostopoulos, D., & Omae, H. (2011). The multidimensionality of multicultural service learning: The variable effects of social identity, context and pedagogy on pre-service teachers’ learning. Teaching and Teacher Education, 27(7), 1078-1089. doi: 10.1016/j.tate.2011.05.004
  • Conner, J. O. (2010a). Learning to unlearn: How a service-learning project helped teacher candidates to reframe urban students. Teaching and Teacher Education, 26(5), 1170-1177. DOI: 10.1016/j.tate.2010.02.001
  • Conner, J. O. (2010b). Building the reciprocal relationship: How students in an affluent private university and a low-income public high school became partners in service and learning. Information for Action: A Journal for Research on Service-Learning for Children and Youth, 3(2), 1-23.
  • Cook, S. (1985). Experimenting on social issues: The case of social desegregation, American Psychologist, 40(4), 452-460.
  • Corrigan, P. W. & O’Shaughnessy, J. R. (2007). Changing mental illness stigma as it exists in the real world. Australian psychologist, 42(2), 90-97.
  • d'Arlach, L., Sanchez, B., & Feuer, R. (2009). Voices from the community: A case for reciprocity in service-learning. Michigan Journal of Community Service Learning, 16(1), 5-16.
  • Deely, S. (2010). Service-learning: Thinking outside the box. Active learning in higher education, 11(1), 43-53. doi: 10.1177/1469787409355870
  • Delve, C., Mintz, S., & Stewart, G. (1990). Community service as values education. San Francisco: Jossey-Bass.
  • Donahue, D., Bowyer, J., & Rosenberg, D. (2003). Learning with and learning from: Reciprocity in service learning in teacher education. Equity and Excellence in Education, 36(1), 15-27.
  • Dyment, J. E., & O’Connell, T. S. (2011). Assessing the quality of reflection in student journals: A review of the research. Teaching in Higher Education, 16(1), 81-97. doi: 10.1080/13562517.2010.507308
  • Einfeld, A., & Collins, D. (2008). The relationships between service-learning, social justice, multicultural competence, and civic engagement. Journal of College Student Development, 49(2), 95-109.
  • Endres, D., & Gould, M. (2009). “I am also in the position to use my whiteness to help them out”: The communication of whiteness in service learning. Western Journal of Communication, 73(4), 418-436. doi: 10.1080/10570310903279083
  • Ennis, C. D. (1999). A theoretical framework: The central piece of a research plan. Journal of Teaching in Physical Education, 18(2), 129-140.
  • Eppler, M.A., Ironsmith, M., Dingle, S.H., & Errickson, M.A. (2011). Benefits of service-learning for freshmen college students and elementary school children. Journal of the Scholarship of Teaching and Learning, 11(4), 102-115.
  • Erickson, J., & O’Connor, S. (2000). Service-learning’s effect on prejudice: Does it reduce or promote it? In C. O’Grady (Ed.), Transforming education, transforming the world: The integration of service-learning and multicultural education into higher education. Mahwah, NJ: Lawrence Erlbaum Associates.
  • Erickson, J., & Santmire, T. (2001). Psychological bases of effective service-learning. In J. Anderson, K. Swick, & J. Yff (Eds.), Strengthening service and learning in teacher education. Washington, DC: American Association of Colleges for Teacher Education and the ERIC Clearinghouse for Teaching and Teacher Education.
  • Eyler, J., Giles, D., (1999). Where’s the learning in service-learning? San Francisco: Jossey-Bass.
  • Fenzel, M.L & Dean, R. J. (2011). Changes in students’ social justice and racial attitudes in an undergraduate child psychology service-learning course. Journal of Research on Service-learning in Teacher Education, 1(2), 20-30.
  • Gaertner, S. L., & Dovidio, J. F. (2000). Reducing intergroup bias: The Common Ingroup Identity Model. Philadelphia: The Psychology Press.
  •  Greene, D. (1998). Student perceptions of aging and disability as influenced by service learning. Physical & Occupational Therapy in Geriatrics, 15(3), 39-55.
  • Hamilton, S.F., & Fenzel, L.M. (1988). The impact of volunteer experience on adolescent social development: Evidence of program effects. Journal of Adolescent Research, 3(1), 65-80.
  • Hironimus-Wendt, R. J., & Lovell-Troy, L. (1999). Grounding service-learning in social theory. Teaching Sociology, 27(4), 360-372.
  • Hinshaw, S. & Stier, A. (2008). Stigma as related to mental disorders. Annual Review of Clinical Psychology, 4, 367-393. doi: 10.1146/annurev.clinpsy.4.022007.141245
  • Hobbs, V. (2007). Faking it or hating it: Can reflective practice be forced? Reflective Practice, 8(3), 405-417. doi: 10.1080/14623940701425063
  • Hollis, S. (2004). Blaming me, blaming you: Assessing service learning and participants’ tendency to blame the victim. Sociological Spectrum, 24(5), 575-600.
  • Holsapple, M. A. (2012). Service-learning and student diversity outcomes: Existing evidence and directions for future research. Michigan Journal of Community Service Learning, 18(2), 5-18.
  • Houshmand, S., Spanierman, L. B., Beer, A. M., Poteat, P. V., & Lawson, L. J. (2014). The impact of a service-learning design course on white students' racial attitudes. Journal of Higher Education Outreach and Engagement, 18(2), 19-48.
  • Jacoby, B. (2003). Building partnerships for service-learning. San Francisco: Jossey-Bass.
  • Jones, S. (2002). The underside of service learning. About Campus, 7(4), 10-15.
  • Kendall, J. & Associates (1990). Combining service and learning: A resource book for community and public service. Raleigh, NC: National Society for Internships and Experiential Education.
  • Kenworthy, J., Turner, R., Hewstone, M., & Voci, A. (2005). Intergroup contact: When does it work, and why? In J. Dovidio, P. Glick, & L. Rudman (Eds.). On the nature of prejudice: Fifty years after Allport (pp. 278-292). Malden, MA: Blackwell.
  • Lewis, J., Neville, H., & Spanierman, L. (2012). Examining the influence of campus diversity experiences and color-blind racial ideology on students' social justice attitudes. Journal of Student Affairs, 49(2), 119-136.
  • Meaney, K. S., Bohler, H. R., Kopf, K., Hernandez, L., & Scott, L. S. (2008). Service learning and pre-service educators' cultural competence for teaching: An exploratory study. Journal of Experiential Education, 31(2), 189-208.
  • Moely, B. E., McFarland, M., Miron, D., Mercer, S., & Ilustre, V. (2002). Changes in college students’ attitudes and intentions for civic involvement as a function of service-learning experiences. Michigan Journal of Community Service Learning, 9(1), 18-26.
  • Murphy, J. W., & Rasch, D. (2008). Service-learning, contact theory, and building black communities. Negro Educational Review, 59(1/2), 63-78.
  • Nenga, S. K. (2011). Volunteering to give up privilege? How affluent youth volunteers respond to class privilege. Journal of Contemporary Ethnography, 40(3), 263-289. doi: 10.1177/0891241611400062
  • Neville, H. A., Lilly, R. L., Duran, G., Lee, R. M., & Browne, L. (2000). Construction and initial validation of the Color-Blind Racial Attitudes Scale (CoBRAS). Journal of Counseling Psychology, 47(1), 59-70.
  • Pettigrew, T. (1988). The intergroup contact hypothesis reconsidered. In M. Hewstone & R. Brown (Eds.), Contact and conflict in intergroup encounters (pp. 169-195). Oxford, UK: Basil Blackwell.
  • Pettigrew, T. & Tropp, L. (2000). Does intergroup contact reduce prejudice? Recent meta-analytic findings. In S. Oskamp (Ed.), Reducing prejudice and discrimination (pp. 93-114) Mahwah, NJ: Lawrence Erlbaum Associates.
  • Pettigrew, T. & Tropp, L. (2005). Allport’s intergroup contact hypothesis: Its history and influence. In J. Dovidio, P. Glick, & L. Rudman (Eds.), On the nature of prejudice: Fifty years after Allport (pp. 262-277). Malden, MA: Blackwell.
  • Pettigrew, T. & Tropp, L. (2006). A meta-analytic test of intergroup contact theory. Journal of Personality and Social Psychology, 90(5), 751-783.
  • Richeson, J. & Nussbaum, R. (2004) The impact of multiculturalism versus color-blindness on racial bias. Journal of Experimental Social Psychology, 40(3), 417-423.
  • Rothbart, M. (1996). Category-exemplar dynamics and stereotype change. International Journal of Intercultural Relations. 20(3/4), 305-321.
  • Simons, L., Blank, N., Fehr, L., Barnes, K., Georganas, D., & Manapuram, G. (2012). Another look at the dissemination of the racial identity interaction model in a cultural-based service-learning course. In J. A. Hatcher & R. G. Bringle, (Eds.), Understanding service-learning and community engagement: Crossing boundaries through research (pp. 47-71) Charlotte, NC: Information Age.
  • Simons, L., Fehr, L., Hogerwerff, F., Blank, N., Georganas, D., & Russell, B. (2011). The application of racial identity development in academic-based service learning. International Journal of Teaching and Learning in Higher Education, 23(1), 72-83.
  • Snow, N. & Reeb, R. (2013). Social stigma and homelessness. Journal of Psychological Practice, 18, 104-139.
  • Sperling, R. (2007). Service-learning as a method of teaching multiculturalism to white college students. Journal of Latinos and Education, 6(4), 309-322.
  • Whitley, M. (2014). A draft conceptual framework of relevant theories to inform future rigorous research on student service-learning outcomes. Michigan Journal of Community Service Learning, 20(2), 19-40.
  • Wiersma, W. (2000). Research methods in education (7th ed.) Boston: Allyn & Bacon.
  • Wilson, J.C. (2011). Service-learning and the development of empathy in U.S. college students. Education & Training, 53(2/3), 207-217.
  • Worthington, R., Navarro, R., Loewy, M., & Hart, J. (2008). Color-blind racial attitudes, social dominance orientation, racial-ethnic group membership and college students' perceptions of campus climate. Journal of Diversity in Higher Education, 1(1), 8-19.
  • Wright, A., Calabrese, N., & Henry, J. (2009). How service and learning came together to promote "cura personalis". International Journal of Teaching and Learning in Higher Education, 20(2), 274-283

Appendix A

Authors

JERUSHA CONNER (jerusha.conner@villanova.edu) is an associate professor of Education at Villanova University. Her research focuses on student engagement, student voice, and youth activism and organizing for educational improvement. Recent publications include Student Voice in American Education Policy (2015, National Society for the Study of Education Yearbook) and Contemporary Youth Activism: Advancing Social Justice in the United States (2016, Praeger).

JOSEPH ERICKSON (Erickson@augsburg.edu) is a professor of Education at Augsburg College. His research addresses service-learning in teacher education and technology integration in schools. He recently published Transforming Teacher Education through Service-Learning (2013, Information Age Publishing) and presented “Educating in the digital age: Why service-learning is more important now than ever” at the 2015 Asia Pacific Regional Conference on Service-Learning.