Race Still Matters: The Relationship Between Racial and Poverty Attitudes Among Social Work Students

The attitudes that social work students hold about race and poverty impact the effectiveness of their practice in the field. This study assessed color-blind racial attitudes and attitudes towards poverty of graduating BSW students (n=41) and MSW students (n=128) from three accredited social work programs. Results indicate a correlation between color-blind racial attitudes and attitudes toward poverty for BSW students, but not MSW students. BSW students had fewer color-blind racial attitudes and more favorable attitudes toward poverty than MSW students. Several predictors of their attitudes were found: their educational status, personal experience of poverty, political ideology, and type of diversity course taken. Implications include the need to approach diversity education from an anti-oppression approach.

Within American society, there are two broad and overarching narratives illuminating the attitudes about poverty and people who are poor. First is the narrative that poverty is the result of individual deficits or failures. This set of attitudes about poverty is akin to color-blind racial attitudes. Since the Civil Rights Movement, color-blind racial attitudes have been the dominant ideology held and taught in the U.S. (Bonilla-Silva, 2006;Neville et al., 2000). This false belief contends that race does not matter and does not play a role in shaping outcomes. It perpetuates racism and White privilege by not acknowledging and addressing the very real disparities that exist between White people and People of Color. A corollary argument could be made about the presence and impact of poverty: individuals are impoverished because of poor choices and personal shortcomings, rather than inequitable societal structures. On a micro level, social workers endorsing this attitude may engage in approaches to service delivery that ignore race and cultural differences, including norms, values, and strengths. On a macro level, social workers may fail to advocate for systemic change to ameliorate racism and White privilege (Abrams & Moio, 2009).
The second narrative is that poverty is the result of structural or systemic failures, such as insufficient employment opportunities and low-paying jobs (Hill et al., 2016;Rank, 2006). From this perspective, social identities such as race and social class are considered overlapping and interconnected and may confer differing levels of privilege and disadvantage (Collins & Bilge, 2016). A growing body of research has documented the intersection of social identities and their effects on social and economic outcomes (e.g., Black & Veenstra, 2011;Caiola et al., 2014).
Although attitudes toward race and poverty have been studied separately, little is known about their relationship or their responsiveness to pedagogical practices within social work education. More understanding is needed about the relationship between attitudes toward race and poverty, as these attitudes affect the strategies used to address persistent and pervasive inequities. Minimal research exists on color-blind racial attitudes among social workers (Loya, 2011) or social work students (Davis, 2019), and to our knowledge, no research exists on the relationship between social work students' endorsement of color-blind racial attitudes and their attitudes towards poverty.
Additionally, understanding how social workers think about and understand the causes of poverty is imperative since these attitudes in turn affect the strategies and policies they suggest to address poverty (Weaver & Yun, 2011). Therefore, the current study examined the association between color-blind racial attitudes and poverty attributions among social work students preparing to graduate. The goal of the inquiry was to inform the inclusion and delivery of content on race and poverty within social work education.

Color-blind Racial Attitudes and Social Work
Color-blind racial attitudes describe the belief that "race should not and does not matter" (Neville et al., 2000, p. 60). More specifically, if race does not have a salient role in affecting people's experiences and outcomes, then it should be de-emphasized. However, critical race scholars have documented the opposite, i.e., the reality of colorcoded differences in society, as well as the unique experiences of People of Color (Bonilla-Silva, 2006;Delgado & Stefancic, 2017;Neville et al., 2000). Thus, race still matters. Recent research on implicit bias challenges the notion of a color-blind racial ideology; it would be harmful-and impossible-to espouse these attitudes when socialization about race is deeply and often unconsciously held (FitzGerald & Hurst, 2017). There are three main tenets of color-blind racial attitudes. First, there is a denial of White privilege, where people do not acknowledge that White people receive societal benefits by virtue of their skin color. Second, the existence of structural discrimination is denied, which implies that policies and programs are neutral in their impact on People of Color. The final tenet is a lack of awareness of discrimination. Many who hold this belief feel as though the U.S. has arrived at a post-racial society, where discrimination is a matter of the past.
Scholars have theorized that color-blind racial attitudes can negatively impact behavior, thereby impacting the quality of services provided by helping professionals. For example, Burkard and Knox (2004) presented 247 psychologists with vignettes of clients from different racial backgrounds. The participants who endorsed more color-blind racial attitudes had more difficulty empathizing with Black clients and were more likely to attribute responsibility to their clients. Despite the potential impact on practice, only a few empirical studies exist on color-blind racial attitudes in social work. Existing studies begin to document the presence of these attitudes among social workers. For example, using the Oklahoma Racial Attitude Scale, Loya (2012) found that over one-third of a sample of White social workers had negative racial attitudes. Specifically, they were categorized as being dominative and conflictive. Individuals holding the conflictive negative attitudes do not condone racism but oppose policies like affirmative action for minorities based on the assumption that the U.S. already promotes equal opportunity (La Fleur et al., 2002). Those with "dominative" attitudes deny the persistence of racism and uphold stereotypes of racial minorities (La Fleur et al., 2002).
Less is known about the presence of color-blind racial attitudes among social work students. However, a growing body of literature demonstrates cause for concern. In one recent study, social work students who demonstrated more knowledge about the Civil Rights Movement endorsed fewer color-blind racial attitudes (Davis, 2019). It appears difficult for students to deny the effects of racism and White privilege when they possess knowledge about historical oppression and its legacy. Color-blind racial attitudes have also been explored among students in related disciplines. Neville et al. (2006) found that psychology students and mental health workers who held more color-blind racial attitudes had lower self-reported multicultural counseling awareness and knowledge and lower multicultural case conceptualization ability. This link between racial attitudes and effectiveness in practice points to the need for research to examine this connection.
Endorsement of color-blind racial attitudes may relate to a person's socialization, lived experience, and educational attainment. Beginning early in life, White people are socialized "to see everyone as equal and as individuals rather than as members of a particular ethnic group" (Ryan et al., 2007, p. 618). Conversely, People of Color have lived experiences that confirm the reality of racism and influence their beliefs about race and ethnicity. People from other marginalized groups, such as women and members of the LGBTQ community, may also be more cognizant of power differentials and injustice. Further, individuals with multiple intersecting oppressed identities, such as being Black and a woman, are even less likely to endorse color-blind racial attitudes (Neville et al., 2000(Neville et al., , 2014. Education level and collegiate lived experience also influence the development of color-blind racial attitudes. With a sample of practicing social workers (n=179), BSW-level workers demonstrated less awareness of color-blind racism and racial privilege than their MSW counterparts (Loya, 2011). Further, in a longitudinal study of White college students (n=847), Neville and colleagues (2014) found that those who engaged in a greater number of diversity experiences (e.g., diversity-related courses and activities) and those who had a greater number of close Black friends held fewer colorblind racial attitudes after four years in college.

Attitudes Toward Poverty and Social Work
Social work has a long tradition of addressing poverty and serving the poor. From the earliest days of the profession, charity organization societies coordinated efforts to address urgent urban social problems affecting individuals, families, and communities. Their activities would later develop into well-known approaches to social work practice, including casework, family counseling, and community organizing (Zastrow, 2013). Friendly visitors of the charity organization societies closely monitored the requests for assistance from those in need, and encouraged the poor to be thrifty, pursue work, and become self-sufficient.
The profession continues to address poverty and works to ameliorate its deleterious effects. In fact, the National Association of Social Workers' (2017) Code of Ethics states that the mission of the profession is: "to enhance human well-being and help meet the basic human needs of all people, with particular attention to the needs and empowerment of people who are vulnerable, oppressed, and living in poverty" (para. 1). Given this overarching and clear mission to improve the lives of marginalized and impoverished communities, social work students are socialized into the values of the profession early in their educational training, and as practicing professionals they are expected to develop the ability to advance social and economic justice (CSWE, 2015). The extent to which social work education fulfills this expectation is unclear. Krumer-Nevo et al. (2009) argue that social work educators cover poverty in an "extraordinarily superficial manner" (p. 226), which renders it a "marginal issue in social work practice" (p. 225).
Several studies have explored the effect of social work education on students' commitment to the mission of the profession. Mizrahi and Dodd (2013) followed a cohort of MSW students who were asked to rank the importance of the goals of the profession. At both the beginning and end of their studies, students endorsed the same primary goal of the profession: "to work to improve the conditions and quality of life for vulnerable/oppressed communities and populations" (p. 588). Clark (2007) had similar findings with a 14-year study of MSW students' (n=2,213) perceptions of poverty. Students began their graduate studies with a preference for societal/institutional methods for addressing poverty over individual adaptation, and by graduation, this preference persisted and increased. It appears as though the profession may attract students who already possess a commitment for addressing issues of social injustice and poverty rather than creating such commitments though pedagogical practices.
Professional attributions of poverty. Explanations of the causes of poverty generally fall under three dimensions, as delineated by Feagin (1972): individualistic, structural, and fatalistic. The individualistic explanation suggests that poverty is the result of personal failure and deficits such as laziness, lack of morals, and lack of middle-class values and beliefs. The structural explanation suggests that people are poor because of structural deficits such as lack of jobs, discrimination, and economic policies. The fatalistic explanation suggests that people are poor because of bad luck, or destiny, accordingly there is nothing they can do to escape poverty .
Understanding social workers' attribution of the causes of poverty is essential because perceptions influence behavior in practice, as well as relative support for policies to address poverty. Robinson (2011) found that social workers who were more likely to attribute poverty to a breakdown in family were more likely to support the restrictive features of the Temporary Assistance to Needy Families program, such as time limits on benefits and family caps. In a qualitative study, Carlson (2016) found that how child welfare workers explained the construct of poverty had implications for the services provided and the realms in which they sought to effect change.
The influence of personal factors. Just as racial attitudes are influenced by social identities and lived experiences, these personal factors have been found to affect social work students' perceptions of poverty. Using a sample of social work students (n= 264) from two Western universities, Castillo and Becerra (2012) found that White students and male students were more likely than their counterparts to believe that people were poor because of individual shortcomings. Those students who experienced poverty as a youth were less likely to attribute poverty to a lack of desire to work. In terms of educational status, Castillo and Bacerra (2012) found that MSW students were more like to agree that poverty is the result of social-structural factors than BSW students. This study used a brief, researcher-constructed instrument to measure beliefs, and could be repeated using an instrument with demonstrated reliability and validity. In an earlier study of only undergraduates (n=119), Schwartz and Robinson (1991) found that BSW students developed beliefs about the causes of poverty that are consistent with the profession's values, and as students approached graduation, they had a greater ability to differentiate between Feagin's (1972) three explanations of poverty. These studies lend support to further investigating the relationship between educational status and attitudes about poverty.

Relationship between color-blind racial attitudes and attitudes towards poverty
Presently, there are no empirical studies directly linking the endorsement of color-blind racial attitudes and attitudes to poverty. The ideas, however, are connected conceptually through frameworks such as Critical Race Theory (Delgado & Stefancic, 2017;Milner, 2013). Attitudes towards poverty and color-blind racial attitudes are intrinsically linked when individuals espouse the belief that we are living in a world that is fair, just, and equitable, and that the causes of injustice fall squarely on the shoulders of individuals (Lerner, 1980). Furnham (2003) surmises that adopting beliefs that the world is just and equitable provides "psychological buffers against the harsh realities of the world as well as personal control over one's own destiny" (p. 796). In essence, people long to feel as if they have done nothing to deserve negative outcomes that others experience and can deny the presence and effects of oppression and privilege. Both color-blind racial attitudes and attitudes towards poverty are suggestive of the notion that the more strongly individuals believe that the world is just and equitable, the more strongly they would support the use of color-blind racial attitudes and individualistic causes of poverty as the means by which to explain racial and social class disparities, respectively. Furthermore, individuals who believe that the world is fair and just are more likely to engage in victim-blaming and less likely to attribute the cause of social injustice, such as poverty and racism, to structural issues (Bonilla-Silva, 2006;Feagin, 1972;Neville et al., 2000). In essence, both color-blind racial attitudes and attitudes towards poverty minimize the role of structural factors in shaping individual outcomes and proposing realistic, equitable solutions.

Research questions
Based on the hypothesized conceptual link between color-blind racial attitudes and attitudes towards poverty, this study examined these attitudes among social work students. The study tested three hypotheses: a) there will be a negative correlation between colorblind racial attitudes and attitudes towards poverty; b) MSW students will endorse more structural explanations of poverty than BSW students; c) MSW students will have lower levels of color-blind racial attitudes than BSW students. Additionally, the literature review supports the investigation of student-specific characteristics as possible predictors of students' attitudes, including racial background, educational status (MSW or BSW), personal experience of poverty, political ideology, and type of diversity course taken (Bray & Balkin, 2013;Griffin & Oheneba-Sakyi, 1993;Robinson, 2009;Toporek & Pope-Davis, 2005).

Procedure
Convenience sampling was used to recruit participants at three CSWE-accredited institutions: two private urban colleges in the Northeast and one public university in the Southeast. Researchers administered a survey with standardized measures, as well as researcher-created questions to elicit information about demographic variables and courses taken. When completed together, the survey elucidated information about the background of students relative to their interactions with poverty and education as well as their attitudes about race and poverty.
Prior to data collection, the study was granted approval from the Institutional Review Board. Data were collected using both an electronic survey hosted by Qualtrics and a paperand-pencil version. Electronic data acquisition began with students reading an informed consent statement and then deciding whether to proceed with the online survey. Penciland-paper data acquisition consisted of the distribution of a hard copy of the survey at the beginning of Social Work Practice classes for both BSW seniors and graduating MSW students with the permission of the instructors of those classes. Again, students could read the informed consent and decide whether to proceed with the survey. At one of the sites, data were collected both ways to reach the most students. Participation was voluntary, and no identifying information about the student or institution was collected. The response rate is not known due to the sampling method. Nine online surveys were started but not finished; with less than 30% of the survey completed, these surveys were discarded as unusable. The remainder of the surveys (n= 169) were completed in their entirety with no missing data.
Student-specific variables. Demographic variables in this study included age, gender (male, female, non-binary), race (White or Person of Color), student location (Northeast or Southeast), and educational status (undergraduate BSW or graduate MSW). In addition to basic demographic variables, information about the students' early life and family was obtained, including eligibility for free and reduced lunch (FRL) in childhood (No = 0 and Yes = 1), and whether or not their parents had a college degree (No = 0 and Yes = 1). Lastly, students were asked to report their political leanings (left, center, or right) and their current economic standing (I often go without basic needs, I sometimes go without basic needs, I never have any extra but my needs are met, I have enough money to meet my needs and a bit extra, or I have plenty). The categories of economic standing were used with permission from Gentlewarrior et al. (2008).

Social work education variables.
Respondents were asked to provide information about their social work education experience by reporting whether they had taken a class related to diversity (No = 0 and Yes = 1) and if they reported taking such a course, they were asked about its focus (multiculturalism or anti-oppression). Multicultural education was described as increasing awareness about cultural norms and beliefs, and learning about diverse populations; anti-oppression education was described as confronting unequal power dynamics in society that reinforce oppression and privilege (Morelli & Spencer, 2000).
Color-blind racial attitudes. The Color-Blind Racial Attitudes Scale (CoBRAS; Neville et al., 2000) is a 20-item instrument that aims to measure three constructs: unawareness of racial privilege (e.g., "Everyone who works hard, no matter what race they are, has an equal chance to become rich"), unawareness of institutional discrimination (e.g., "Social policies, such as affirmative action, discriminate unfairly against White people"), and blatant racial issues (e.g., "Racial problems in the U.S. are rare, isolated situations"). Participants indicated their level of agreement with each statement using a 6-point Likerttype scale from 1 (strongly disagree) to 6 (strongly agree). Several items were reverse-scored. Higher scores indicated a greater unawareness or denial of racism. The CoBRAS was normed with a social desirability scale, making it less likely that participants' responses reflect this bias (Neville et al, 2000). Previous studies have established that CoBRAS is both reliable and valid with Cronbach alpha scores ranging from α = .68 twoweek test-retest reliability to α= .80 internal consistency for the Unawareness of Racial Privilege subscale (Neville et al., 2006;Neville et al., 2000).
Attitudes towards poverty. The short form of the Attitudes Toward Poverty scale (ATP; Yun & Weaver, 2010) is a 21-item instrument that measures attitudes related to poverty and poor people. The ATP aims to measure three attitudes: personal deficiency (e.g., "Children raised on welfare will never amount to anything"), stigma (e.g., "An ablebodied person collecting welfare is ripping off the system"), and structural perspectives (e.g., "Poor people are discriminated against"). Participants indicated their level of agreement with each statement using a 5-point Likert-type scale from 1 (strongly agree) to 5 (strongly disagree). Several items were reverse-scored. Higher scores indicate greater belief in the structural explanation of poverty, while lower scores indicated greater belief in individual explanations of poverty. This measure has been found to be both reliable and valid with Cronbach's alpha ranging from α = .93 upon assessment of internal consistency and a split-half reliability of α = .87 (Yun & Weaver, 2010).

Data Analysis
Descriptive statistics were analyzed to categorize students' demographics and characteristics, and assess their scores on the overall CoBRAS, three CoBRAS subscales, overall ATP score, and three ATP subscales. Additionally, several tests of inferential statistics were conducted in order to gain further insight into relationships among variables including the relationship between color-blind racial attitudes and attitudes towards poverty, which enabled insight into the first aim of the study. T-tests were conducted to compare the scores between BSW and MSW students on both measures, and correlations were conducted to assess the relationship between CoBRAS and ATP for BSW and MSW students, respectively, illuminating insight into the second aim of the study. Lastly, a stepwise linear regression was conducted to assess the demographic and educational determinants of color-blind racial attitudes and attitudes toward poverty to provide insight into the third aim of the study. It was important to assess not only the overall differences between MSW and BSW students, but other variables that may contribute to lower levels of color-blind racial attitudes in students. IBM SPSS (Version 21.0) was used to perform the statistical analyses.

Findings Sample
The sample comprised 41 BSW seniors and 128 MSW students who intended to graduate in May 2017. Participants were asked to self-identify their social identities. Table  1 displays demographic and student-specific descriptive statistics. BSW students primarily identified their race as White (63%), their gender as female (88%), and their political tendency as liberal (64%). Their ages ranged from 21-35, with a mean of 21.5 years old.
The largest proportion of students (42%) described their economic situation as "always having enough money to meet my needs and a little bit extra." Growing up, just under half (47%) were eligible for free or reduced-fee lunch, and just over half (53%) had at least one parent who earned a college degree. Most of the MSW students identified their race as White (66%), their gender as female (89%), and their political ideology as liberal (89%). Their ages ranged from 21-49, with a mean of 28.4. As with the BSW students, most of the MSW students (54%) described their economic situation as "always having enough money to meet my needs and a little bit extra." Growing up, approximately half (51%) were eligible for free or reduced-fee lunch, and just over half (56%) had at least one parent who earned a college degree.
Most BSW students (87.7%) and most MSW students (92%) had taken at least one diversity-related course during their social work education. The students who had taken such a course tended to describe it as having an anti-oppression focus, rather than a multicultural approach.
Aim 1: There will be a negative correlation between color-blind racial attitudes and attitudes towards poverty: the more color-blind racial attitudes held, the fewer structural explanations of poverty are endorsed.
No correlation emerged between ATP and CoBRA for MSW students. However, for BSW students, there was a strong negative correlation between the total scores (r= -0.66; p < 0.001) and most of the subscales. All three subscales of the CoBRA scale were correlated with the total ATP scale as well as the stigma subscale and the structural perspective subscale, but not the personal deficiency subscale. At the undergraduate level, there was a negative correlation between color-blind racial attitudes and attitudes towards poverty (see Table 2). Aim 2: MSW students will endorse more structural explanations of poverty than BSW students.
Significant differences existed between BSW and MSW students in their attitudes towards poverty. BSW students had significantly higher scores (M= 92.2 SD=10.46) than MSW students (M=87.1 SD= 11.26; t(167)= 2.45, p < .05). This means that BSW students were more likely to espouse the structural explanations of poverty. Additionally, there was a statistical difference between BSW students (M=26.6, SD=3.51) and MSW students (M= 24.9 SD= 3.37; t(167)= -2.73, p < .01) on the structural perspective sub-scale, indicating that BSW students were more likely than their MSW counterparts to attribute poverty to structural forces. There were no statistical differences for the personal deficiencies and stigma sub-scales (see Table 3). Thus, the second hypothesis was not supported as the more structural attitudes were seen for BSW students, not MSW students.

Aim 3: MSW students will have lower levels of color-blind racial attitudes than BSW students.
In terms of color-blind racial attitudes, there was a statistically significant difference between BSW and MSW students. Specifically, BSW students had significantly lower CoBRA scores (M=34.5 SD=12.79) while MSW students were nearly five points higher (M= 39.4 SD=12.93;t(167)= -2.15, p < .05). This meant that MSW students had greater unawareness or denial of racism. Upon closer inspection of sub-categories, BSW students were significantly more aware of racial privilege (M= 13.5 SD=5.81) than were the MSW students (M= 15.9 SD=5.96; t(167)= 2.26, p < .05). Again, this was not the hypothesized finding, as the lower levels were seen for BSW students.

Linear Regression Analyses
A stepwise linear regression was conducted to assess the extent to which demographic and educational achievement were determinants of color-blind racial attitudes. The educational status of students (BSW or MSW) explained variations in racial attitudes, with Master's level students having higher CoBRA scores than undergraduates (see Table 3).The type of diversity training received also helped to explain variation in CoBRA scores. Students who perceived their diversity class to focus on anti-oppression had lower CoBRA scores. Poverty status as a youth, as measured by being eligible for free or reduced lunch, also explained the variance in CoBRA scores. Individuals who were eligible to receive free or reduced lunch had lower CoBRA scores (see Table 4). Racial background, political ideology, and current economic standing were additional independent variables that were assessed but lacked sufficient statistical significance to be included in the model.
Stepwise linear regression was also used to assess the demographic and educational correlates of poverty attitudes to gain further insight into the unexpected and interesting findings observed with regard to two of the three study aims. As is shown in Table 5, the educational status of students was a significant determinant of attitudes. The data reveal that Masters level students had lower scores indicating greater belief in the individual explanation of poverty. The type of diversity class that students took was also associated with attitudes toward poverty. Students who perceived their diversity course to focus on anti-oppression, rather than multicultural education, were more likely to subscribe to the role of inequitable structures in causing poverty. Political affiliation also affected poverty attitudes. Compared to liberals, conservatives tended to see poverty as a result of individual failure. Lastly, attitudes towards poverty are determined by previous experience in poverty. Students who were eligible to receive free or reduced lunch as a youth were less likely to blame the individual for poverty. Racial background, parents' level of education, and current economic standing were additional independent variables that were assessed but lacked statistical significance, and thus, were not associated with attitudes toward poverty.

Discussion
This study explored color-blind racial attitudes and attitudes about poverty among social work students in three CSWE-accredited schools of social work. Results from this study were both surprising and affirming. The hypotheses that undergirded this study were that MSW students, having benefitted from greater instruction and exposure to diversity content, would hold fewer color-blind racial attitudes and possess a greater appreciation for the structural forces that create a system of poverty. Counter to the hypothesized results, findings from this study indicate that BSW students were more likely to have lower scores on the CoBRA and higher scores on the ATP than MSW students.
These findings indicate that MSW students are more likely to deny racism and White privilege and ascribe to the belief that poverty is more situated in individual factors rather than the structural forces that drive oppression. Interestingly, the overall and subscale correlations that were observed between the CoBRA and the ATP among the BSW students vanished when examined in the context of MSW students. BSW students may learn about the concepts of oppression and privilege during a formative time in their development. Often the undergraduate years are a time when students are exposed to new ideas and people, which can powerfully shape their beliefs about diversity and difference. Another possible explanation for this finding is that among graduate students, there is a separate phenomenon that was being measured rather than a clear relationship between lower colorblind racial attitudes and attitudes about poverty. It is much easier to teach about these concepts in a classroom where students can gain appreciation for the theoretical mechanisms that produce and support poverty than it is to cope with the harsh realities and disparities that students observe once they begin working in the field. Therefore, it is possible that after having spent more time working directly with clients, MSW-level social workers are more likely to blame individual clients as it allows for catharsis from the frustrations of dealing with systemic and systematic poverty on a daily basis.
Some results of this study, however, were consistent with what is known about attitudes towards race and poverty and the importance of social work education. Results were consistent with previous studies on students' attitudes, indicating that anti-oppressive education leads to more awareness of racism and White privilege (Abrams & Gibson, 2007;Abrams & Moio, 2009;Davis et al., 2014;Heron, 2004;Singh, 2019;Williams & Parrott, 2014. This finding supports Abrams and Moio's (2009) critique that multicultural education focuses on beliefs and attitudes but does not draw attention to social injustices or structural forces, and thus, fails to mobilize students into social action. Additionally, the lived experiences of students prior to entering the program were observed as predictive of their attitudes towards race and poverty. Not surprisingly, students who reported being eligible for free or reduced lunch as children were more aware of White privilege compared to those students who did not meet this criterion for lowincome status. Based on the Just World Belief (Lerner, 1980), those students with firsthand experience of poverty would be more aware of the role of structural forces, such as White privilege, in shaping people's lives.
In terms of attitudes towards poverty, an expected result is that students who indicated that their diversity course took an anti-oppression approach were more likely to view poverty as a result of structural deficits. This was expected as anti-oppression education emphasizes the role of macro forces in social and economic outcomes. Another expected finding was that students who experienced poverty growing up were less likely to perceive poverty as an individual problem. This finding echoes those in Castillo and Becerra's (2012) study, as they also found that students who were eligible for free or reduced-fee lunch attributed poverty to individual factors. Students who identified with a conservative political ideology were also more likely than their liberal counterparts to blame individuals for their poverty. While social workers who hold conservative views and those who favor liberal views may want to help people who are poor, their beliefs about the cause of poverty may differ greatly, from individualistic to structural to fatalistic (Feagin, 1972). These beliefs, in turn, influence the policies and practices that they endorse to ameliorate poverty (Rosenwald et al., 2012).

Limitations
This study had a number of limitations. First, the BSW sample size was small and limited power to detect results free from type-1 and type-2 errors. A larger sample size could have yielded more significant findings or perhaps the true nature of the relationships; further research is needed to more fully understand these findings and their meaning. Second, the small sample size of overall participants and of BSW students in particular limited the comparison between BSW and MSW across institutions or regions. It is possible that regional differences may exist between undergraduate and graduate students; however, we were unable to thoroughly assess those differences. Third, this cross-sectional study does not capture the students' attitudes at the beginning of their program and at the end. A pre-/post-test design could provide insights into the efficacy of the pedagogical approaches used in the program and could document changes in students' thinking over time. Further, by examining the syllabi of the diversity courses taken by students, it would be possible to determine whether the course was taught from a multicultural or antioppression approach, and to compare this focus to students' perception of the course. Lastly, this quantitative study does not provide a deep understanding of how students' attitudes affect their practice, especially in their field placements. Despite these limitations, however, the current study is still instructive of students' attitudes towards race and poverty and points to implications for social work education.

Implications for Social Work Education
The findings from this study suggest the need to approach diversity education from an anti-oppression approach. An anti-oppression (or specifically, anti-racist) framework highlights the structural forces in society that affect outcomes for marginalized groups (Abrams & Gibson, 2007;Abrams & Moio, 2009). While it is important to develop multicultural knowledge and awareness, Ortiz and Jani (2010) purport that it does not stimulate students to challenge the Just World Belief or engage in social action. Critical Race Theory is an example of an anti-oppression approach that can be helpful in raising students' consciousness about the systems of oppression and privilege (Kolivoski, Weaver & Constance-Huggins, 2014). By zeroing in on the racial power struggles in the U.S., Critical Race Theory suggests that racism and White privilege are tightly and invisibly woven into the fabric of society (Delgado & Stefancic, 2017).
Secondly, educators can help foster more positive and realistic attitudes about race and poverty by infusing such content across the curriculum and inviting students to reflect on the meaning of their own lived experiences. Infusing anti-oppression content across the curriculum, instead of teaching it solely in selected diversity courses, demonstrates a social work program's commitment to social, racial, and economic justice. Social work educators need to help students grasp poverty as a social justice issue that can be addressed on all levels of policy and practice (Hill et al., 2016). Future studies could explore whether and how anti-oppression approaches are used throughout curricula in social work programs that use an infusion model.
If these implications are taken in sum and integrated into BSW and MSW curricula, perhaps meaningful and permanent gains can be achieved with regard to educating our students about the important issues of race, power, and privilege. Such changes are desperately needed in the evolving landscape of social work practice as the U.S. becomes more diverse. Further, the current racial and political climate, in which hate crimes are on the rise and Black people are the most frequent victims (Eligon, 2018), necessitates that social workers are race-conscious and actively work against the systems that reinforce oppression and privilege. Social workers engage in clinical practice and policy practice within the very institutions that have long created and reinforced pervasive inequities, and thus, have an opportunity and a responsibility to advocate for justice on both micro and macro levels. Further, within agencies and communities, social workers often hold leadership positions in which they can help clients and colleagues challenge their beliefs about why racism and poverty exist, and mobilize together to effect social change.

Conclusion
This study makes an important contribution to the diversity discourse by shedding light on social work students' attitudes towards race and poverty. Given that attitudes shape behavior (Carlson, 2016), it is imperative to gauge the attitudes of social work students as they prepare to engage in practice with oppressed populations. People of color and people in poverty can be targeted by systemic oppression, and thus, our professional values should be reflected in the attitudes that practitioners hold. The current study suggests that there are differences in color-blind racial attitudes and attitudes towards poverty based on educational status, personal experience with poverty, and type of diversity course taken, as well as differences in attitudes towards poverty based on political affiliation. These findings hold important implications for social work educators as we prepare students to practice in an increasingly diverse country with a deep legacy of inequities. Together, our work can stimulate positive and realistic attitudes about race and poverty among social work students at all levels, thereby upholding our professional commitment to advance racial, social, and economic justice.