Brigham Young UniversityBYU ScholarsArchive

All Faculty Publications

2018

Religious Discrimination Scale: Development andInitial Psychometric EvalutationKawika AllenBrigham Young University, gekawika_allen@byu.edu

Kenneth T. FullerFuller Theological Seminary, ktwang@fuller.edu

See next page for additional authors

Follow this and additional works at: https://scholarsarchive.byu.edu/facpubPart of the Mental and Social Health Commons

Original Publication CitationAllen, G. E. K., Wang, K. T., Richards, P. S., *Ming, M. & *Suh, H. N. (2018). Religiousdiscrimination scale: Development and initial psychometric evaluation. Journal of Religion andHealth, 1-14.

This Peer-Reviewed Article is brought to you for free and open access by BYU ScholarsArchive. It has been accepted for inclusion in All FacultyPublications by an authorized administrator of BYU ScholarsArchive. For more information, please contact scholarsarchive@byu.edu,ellen_amatangelo@byu.edu.

BYU ScholarsArchive CitationAllen, Kawika; Fuller, Kenneth T.; Richards, P. Scott; Ming, Mason; and Suh, Han Na, "Religious Discrimination Scale: Developmentand Initial Psychometric Evalutation" (2018). All Faculty Publications. 3184.https://scholarsarchive.byu.edu/facpub/3184

AuthorsKawika Allen, Kenneth T. Fuller, P. Scott Richards, Mason Ming, and Han Na Suh

This peer-reviewed article is available at BYU ScholarsArchive: https://scholarsarchive.byu.edu/facpub/3184

ORIGINAL PAPER

Religious Discrimination Scale: Development and InitialPsychometric Evaluation

G. E. Kawika Allen1 • Kenneth T. Wang2 • P. Scott Richards3 •

Mason Ming3 • Han Na Suh4

! Springer Science+Business Media, LLC, part of Springer Nature 2018

Abstract This study presents the development and initial psychometric evaluation of theReligious Discrimination Scale (RDS). This 11-item instrument identified three dimen-sions based on perceived discrimination experiences of members of The Church of JesusChrist of Latter-day Saints (LDS): Perceived Prejudice, Closet Symptoms, and NegativeLabels. The psychometric evaluations of the RDS indicated a strong and clear factorstructure as well as good internal consistency reliability. The test of measurement andstructural invariance across gender also suggested that the RDS scale is equally appropriate

& G. E. Kawika Allengekawika_allen@byu.eduhttp://education.byu.edu/directory/view/kawika-allen

Kenneth T. Wangktwang@fuller.eduhttp://fuller.edu/faculty/kwang/; http://kennethwang.com/apsr/

P. Scott Richardsscott_richards@byu.eduhttp://education.byu.edu/consortium; http://ce.byu.edu/cw/division36/;http://education.byu.edu/consortium/bridges/index.html

Mason Mingmshaleming@gmail.com

Han Na Suhhsm7b@mail.missouri.edu

1 Department of Counseling Psychology and Special Education, McKay School of Education, 273MCKB, Brigham Young University, Provo, UT 84602, USA

2 Department of Clinical Psychology, Fuller Theological Seminary, 180 N. Oakland Ave, Pasadena,CA 91101, USA

3 Department of Counseling Psychology and Special Education, McKay School of Education, 340MCKB, Brigham Young University, Provo, UT 84602, USA

4 Department of Educational, School and Counseling Psychology, University of Missouri, Columbia,USA

123

J Relig Healthhttps://doi.org/10.1007/s10943-018-0617-z

to be used with both men and women. Implications for practice and research as well asfuture directions are discussed.

Keywords Scale development ! Religion ! Discrimination ! Latter-daySaints ! Reliability ! Validity

Introduction

Current research on discrimination often defines it as the unjust treatment or action towarda person belonging to a different category, specifically minority groups (Woodford et al.2014). Studies have addressed various forms of discrimination including: overt, interper-sonal, subtle, and witnessing (Woodford et al. 2014; Jones et al. 2016). Overwhelmingly,the research on discrimination provides evidence that individuals who perceive discrimi-nation or are discriminated against suffer negative psychological effects related to thediscrimination (Thijs and Piscoi 2016). Some of these negative effects are: higher levels ofpsychological distress (Ajrouch et al. 2010), increased odds of major depressive disorder ordepressive symptoms in general (Molina and James 2016; Rosenthal et al. 2015), decreasedsocial competence (Myrick and Martorell 2011), raised daily cortisol levels (Huynh et al.2016), increased risk of lifetime alcohol use disorders (Molina et al. 2016), and increasedanxiety symptoms (Rosenthal et al. 2015). These findings have been discovered in a varietyof populations.

Research on discrimination among diverse and marginalized groups seems to beprevalent, especially related to race, LGBT, and gender populations. Numerous studieshave linked racial discrimination to negative mental and physical health outcomesincluding depression, anxiety, stress and lowered self-esteem (Ayalon and Gum 2011;Cano et al. 2016; Hudson et al. 2016; Kim and Noh 2014; McDonald et al. 2014; OtinianoVerissimo et al. 2014; Spence et al. 2016). Similarly, individuals in the LGBT communityhave reported several negative effects related to discrimination including: psychologicaldistress (Lyons, 2016), PTSD symptoms (Reisner et al. 2016), depressive symptoms,suicidal ideation, and self-harm (Almeida et al. 2009). Additionally, gender discriminationhas been positively associated with psychological symptoms, physical health difficulties(Greer 2011; Hahm et al. 2010; Perry et al. 2013), and depression and anxiety symptoms(Foynes et al. 2013). Some have reported that women, but not men, are affected byperceived discrimination because they are the minority group; however, there are noreported gender differences in the prevalence of discrimination (Schmitt et al. 2002;Kessler et al. 1999).

Much of the academic research in the multicultural field has been focused on race/ethnic, LGBT, and gender discrimination, but religious discrimination as a multiculturaltopic has been largely ignored (Sheridan 2006; Jordanova et al. 2015). For example,research on religious discrimination in the workplace has received much less attention thanother workplace discrimination protections (Lund Dean et al. 2015; Ali et al. 2015).Furthermore, Hodge (2007) noticed a paucity in social work literature regarding religiousdiscrimination. Despite the lack of research on religious discrimination, over the lastdecade religious discrimination claims have risen more rapidly than other categoriesprotected under the Civil Rights Act (Ghumman et al. 2013). Sheridan (2006) noted thatreligious affiliation may be a more meaningful predictor of prejudice than race or ethnicity.Therefore, as the demand for more research on religious discrimination has risen, thecurrent body of research on multicultural discrimination has failed to supply the necessary

J Relig Health

123

literature and research tools to meet the demand (van der Straten Waillet and Roskam2012).

Even possibly more concerning and non-existent is the lack of discrimination instru-ments to measure religious discrimination. There are virtually no adequate scales thatreliably and validly assess perceived religious discrimination. The scales and instrumentsthat do exist are few and lacking. Rippy and Newman (2007) developed the PerceivedReligious Discrimination Scale (PRDS), which assesses one’s discriminatory experiencesbecause of one’s religion. However, this measure includes items specifically designed for aMuslim population, thereby limiting the generalizability and making it difficult for otherreligious populations to be studied using the scale. Some example items include: ‘‘Haveyou been called a terrorist, foreigner, Osama Bin Laden?’’ ‘‘Has a wounded Iraqi orAfghani woman or child reminded you of a relative or friend?’’ ‘‘Do you have a strongerassociation with other Muslims/Arabs?’’ Rippy and Newman’s (2007) PRDS examinesthree factors: Racial Prejudice and Discrimination, Bicultural Identification, and RacistEnvironment. The confounding relationship between religion and race among these factorsis also a limitation for the use of this measure with diverse intersections between one’sreligious orientation and race. The PRDS possesses adequate internal consistency relia-bility estimates: PRDS Total (a = .91), Racial Prejudice and Discrimination (a = .91),Bicultural Identification (a = .82), and Racist Environment (a = .85). However, furtherpsychometric properties of the PRDS have not been thoroughly tested beyond these initialalpha values. Since the PRDS has not been tested for more psychometric estimates, it isdifficult to rely on the PRDS as a trustworthy and generalizable religious discriminationmeasure.

Non-religious discrimination-focused scales which include specific items regardingreligious discrimination have also been developed such as the Perceived Ethnic Discrim-ination Questionnaire (PEDQ; Brondolo et al. 2005), and the Spiritual Competence Scale(SCS; Hodge 2005). The PEDQ, like the PRDS above, combine some aspects of ethnicityand religion, which make it difficult to focus on RD alone. The SCS contains items thatspecifically ask about the environment within a social work program. Some sample itemsare: ‘‘To what degree does your social work program (or educational program, department,agency, etc.) foster respect for religious and spiritual cultures?’’ and ‘‘How acceptable is itin your social work program to share religious or spiritual views?’’ Thus, these scales areinadequate because they are not entirely focused on the construct of perceived religiousdiscrimination or are specific to a social work environment and therefore lack importantdetails and generalizability/use.

Given the paucity of religious discrimination instruments, there is a need to develop aninstrument that is able to reliably and adequately measure one’s experiences of religiousdiscrimination, and identify the different dimensions/aspect of this construct. Such aninstrument must solely assess perceived religious discrimination, be adaptable acrossdiverse religious groups, and contain sound psychometric properties. The research isabundant related to the negative psychological effects of religious discrimination amongmultiple populations (Ajrouch et al. 2010; Molina and James 2016; Rosenthal et al. 2015;Thijs and Piscoi 2016), but to what extent have we really assessed religious discriminationin detail regarding specific factors and experiences? The stakes are high when it comes tothe negative effects of perceived discrimination; therefore, a religious discriminationmeasure is needed to better assess religious discrimination among various religious groups.The aim of this study was to develop and initially provide psychometric properties for theReligious Discrimination Scale (RDS).

J Relig Health

123

Methods

Participants

Sample 1 consisted of 270 (96 men, 174 women) participants. The majority of participantswere White (87%) and from suburban settings (71%). The mean age of participants was20.89 (S.D. = 3.94). The household income of these participants was mostly in the$100,000–$200,000 (36%) and $75,000–$100,000 (27%) ranges. Sample 2 consisted of274 (100 men, 174 women) participants. The majority of participants were White (89%)and from suburban settings (66%). Participants’ mean age was 20.68 (S.D. = 2.78). Thehousehold income of the sample 2 participants was mostly in the $100,000–$200,000(30%) and $75,000–$100,000 (24%) ranges. Participants across two samples were almostexclusively Latter-day Saints (99%).

Data for this study were collected as part of a larger multiple-measure research projectconducted through Brigham Young University in Provo, Utah, in collaboration withColumbia University in New York. BYU’s SONA online research participation system wasutilized to gather participants and directed them to the study on a Qualtrics survey.Research participants were recruited from among a convenience sample of studentsattending Brigham Young University. Students in various classes (both undergraduate andgraduate) at BYU were offered extra credit as an incentive for participating in the study.Students were sent a link by their instructors which directed them to the SONA website.The survey took approximately 30 min to complete.

RDS Item Development

Other discrimination scales were analyzed to gather possible items for the RDS regardingcontent and meaning of experiences, feelings, and perceptions related to discrimination.The authors examined items from the Everyday Discrimination Scale (Williams et al.1997) and the Daily Life Experience (DLE) subscale of the Racism and Life ExperienceScale (Harrell 1994). After the initial development of the RDS item pool, the initial itemswere evaluated thoroughly and rigorously to check for technical and semantic accuracy,correctness, and sound language structure. Any item that did not fit the criteria of religiousdiscrimination and prejudice accurately was either edited or eliminated. The RDS itemswere revised four times between three authors to capture the essence of religious dis-crimination initially beginning with a pool of 22 items. The final RDS 11 items wereselected among the 22-item pool through factor analyses while maintaining the coresubjective meaning of religious discrimination. There were three subscales in the final RDSscale; five items were designed for Perceived Prejudice, three for Closet Symptoms, andthree for Negative Labels. Example items included: ‘‘I sense hostility from others becauseof my religious affiliation’’ (Perceived Prejudice), ‘‘I felt inclined to keep my religiousaffiliation private’’ (Closet Symptoms), and ‘‘Others hold negative stereotypes of peoplewith my religion’’ (Negative Labels). Each item was rated on a 5-point Likert scale(1 = never, 2 = rarely, 3 = sometimes, 4 = frequently, 5 = always). The instructions toparticipants were: ‘‘Please rate how often during your life you have had the followingexperiences.’’

J Relig Health

123

Results

We used the first sample (N = 270) for exploratory factor analysis (EFA) to select the scaleitems. The second sample (N = 274) was used for confirmatory factor analyses (CFA) tocross-validate the factor structure results from the EFA. The reliability and validityanalyses were conducted with both sample 1 and sample 2, separately.

Preliminary Analyses

The normality of the RDS items were examined to determine whether using robustmaximum likelihood (MLR) as the estimator for CFA analyses was appropriate. Amongthe 22 items in the initial RDS pool used for EFA in sample 1, only three items had anabsolute skew value over 1.0 and two items had absolute kurtosis values over 1.0, but allunder 1.30, which indicated that the items did not significantly deviate from normal dis-tributions. Of the 11 RDS items used for CFA in sample 2, two of the items had an absoluteskew value over 1.0 and one item had absolute kurtosis values over 1.0, but all under 1.30,which also indicated that the item distributions approximated normality. Thus, MLRestimation procedure was used in the CFA analyses of this study.

Sample 1

We first conducted exploratory factor analyses (EFA) for item selection with sample 1(N = 270). The Kaiser–Meyer–Olkin measure of sampling adequacy for the initial EFAwas .94, and Bartlett’s test of sphericity [v2(231) = 4589.42, p\ .001] indicated that thecorrelation matrix was appropriate for factor analysis. To determine the number of factors,we conducted a parallel analysis and scree plot. Parallel analysis was conducted bycomparing initial eigenvalues of this sample with those generated through random data andsuggested a two-factor solution. Scree plot suggested a two- or three-factor solution. Wethus conducted principal axis factor analyses on the 22 items with two- to three-factorsolutions using an oblique (Promax) rotation. The most interpretable solution was a three-factor solution, which had at least three clean items (i.e., one factor loading greater than .40and no cross-loading over .30) on all the factors. Compared to the three-factor solution,there were much more cross-loaded items in the two-factor solution. The three factorsincluded items that generally described discrimination in the forms of: Perceived Prejudice,Closet Symptoms, and Negative Labels. Among the 22 items from the initial pool, 11 wereselected based on the following criteria: (a) factor loadings[ .40 (Netemeyer et al., 2003),(b) cross-loading\ .30, (c) item content within each factor consistent but not redundant,and (d) no more than five items representing each factor given the goal to develop a briefmeasure (Tabachnick and Fidell 2007). Another EFA using principal factor was conductedwith the 11 selected items. The three-factor solution accounted for 65.17% of the totalvariance explained before rotation. After the oblique rotation all factor loadings exceeded.40 on the respective factor, and no item had a cross-loading over .30 on another factor.The five Perceived Prejudice items had loadings ranging from .62 to .83 and accounted for47.3% of the total variance before rotation. The three Closet Symptoms items had loadingsranging from .62 to .89 and accounted for 17.0% of the total variance before rotation. Thethree Negative Labels items had loadings ranging from .64 to .94 and accounted for 9.9%of the total variance before rotation. Each of the items representing the three factors and

J Relig Health

123

their factor loadings, communality estimate, mean, and standard deviation are presented inTable 1.

Sample 2

Cross-Validation—Confirmatory Factor Analysis

Confirmatory factor analysis (CFA) was conducted with sample 2 (N = 274) using Mplus 7to cross-validate the measurement qualities of the CCLS with MLR as the estimator andGeomin as the rotation method. The CFA model constrained the 11 RDS items to load ontotheir corresponding factors based on the EFA results. The three factors were permitted tocorrelate with one another. The range of standardized factor loadings for the factors was:.59 to .80 for Perceived Prejudice, .57 to .96 for Closet Symptoms, and .63 to .86 forNegative Labels. The fit statistics for this three-factor oblique model [CFI = .948,SRMR = .059, RMSEA = .068 (90% C.I. .049–.086)] were adequate. We also examinedthree competing models: a three-factor orthogonal model, a bifactor model (i.e., each itemloads on a general Religious Discrimination factor and one of the three orthogonal factors),and a one-factor model. The fit indices for all four models are presented in Table 2. Based

Table 1 Summary of pattern matrix for principle axis factoring/Promax rotation of RDS items

Factor loading h2 M S.D.

1 2 3

Perceived prejudice

8. I sense hostility from others because of my religiousaffiliation

.83 - .04 .04 .69 1.94 0.94

7. I was passed over for opportunities due to my religion .81 - .01 - .11 .57 1.76 0.93

6. Felt socially avoided by others due to my religion .79 .13 - .06 .71 1.98 0.94

2. I was ignored because I am a religious person .78 .00 .03 .63 2.00 0.95

1. I felt disrespected because of my religious views .62 - .02 .26 .62 2.42 0.99

Closet symptoms

4. I felt inclined to keep my religious affiliation private. - .11 .89 .01 .70 2.14 0.90

5. I was afraid of others finding out about my religiousbeliefs

.05 .86 .00 .78 1.79 0.82

11. I do not feel free to express who I am religiously .20 .62 - .01 .56 1.85 0.93

Negative labels

10. Others hold negative stereotypes of people with myreligion

- .17 .10 .94 .77 3.42 1.01

9. I have heard people make unfriendly remarks about myreligion

.16 - .05 .76 .72 3.18 0.95

3. People assumed things about me because of myreligion

.04 - .07 .64 .42 3.44 1.05

Final 11 RDS items. Unique factor loadings[ .40 are in bold. N = 270 participants. Factor 1 perceivedprejudice, Factor 2 closet symptoms, Factor 3 negative labels, h2 item communalities at extraction. Eachitem is rated on a 5-point Likert scale (1 = never, 2 = rarely, 3 = sometimes, 4 = frequently, 5 = always).The instructions to participants were: Please rate how often during your life you have had the followingexperiences

J Relig Health

123

on the general guidelines, the CFI, SRMR, and RMSEA all indicated adequate data tomodel fit for both the three-factor oblique and the bifactor models, but not the three-factororthogonal model or one-factor model. To compare between the three-factor oblique andthe bifactor model, we examined the Akaike information criteria (AIC). Although thebifactor model was marginally better, the three-factor model is reasonably close and sig-nificantly easier to use in research and practice.

Table 2 Goodness-of-fit indicators for the competing models of the 11-item RDS

Model MLRv2 df CFI RMSEA [CI] SRMR AIC

Three-factor oblique 92.31 41 .948 .068 [.049–.086] .059 6942.96

Three-factor orthogonal 196.09 44 .846 .112 [.097–.129] .203 7051.66

Bifactor 66.22 33 .966 .061 [.039–.082] .034 6918.27

One-factor 373.09 44 .667 .165 [.150–.181] .105 7264.23

N = 274. RDS religious discrimination scale, CFI comparative fit index, RMSEA root-mean-square error ofapproximation, CI confidence interval for RMSEA, SRMR standardized root-mean-square residual, AICAkaike information criteria

Table 3 Testing for measurement and structural invariance across gender groups

MLRv2 df Mcomp

MLRDv2 Ddf CFI RMSEA SRMR

Male 46.92 41 .986 .038 .055

Female 74.27 41 .944 .068 .066

[M0] Unconstrained 120.60 82 .962 .059 .062

[M1] Loadings 128.69 90 M1-M0

8.08 8 .970 .056 .068

[M2] Loadings, Intercepts 141.41 98 M2-M1

12.87 8 .957 .057 .071

[M3] Loadings, Intercepts, ResidualVariances

157.58 109 M3-M2

16.13 11 .952 .057 .074

[M4] Loadings, Intercepts, FactorVar/Cov

148.64 104 M4-M2

7.26 6 .956 .056 .091

N = 274 (174 women, 100 men)

[M0], The Baseline Model (i.e., all parameters freely estimated); [M1], The Invariant Factor LoadingsModel (i.e., constraining all factor loadings to be equal across the two groups); [M2], The Invariant FactorLoadings and Intercepts Model (i.e., constraining all factor loadings and intercepts to be equal across the twogroups); [M3], The Invariant Factor Loadings, Intercepts and Residual Variances Model (i.e., constrainingall factor loadings, intercepts, and residual variances to be equal across the two groups); [M4] The InvariantFactor Loadings, Intercepts and Factor Variances/Covariances Model (i.e., constraining all factor loadings,intercepts, and factor variances and covariances to be equal across the two groups)

MLR robust maximum likelihood, CFI comparative fit index, RMSEA root-mean-square error of approxi-mation, SRMR standardized root-mean-square residual

J Relig Health

123

Measurement Invariance Across Genders

We conducted a multiple-group CFA to examine measurement invariance between menand women using a forward (sequential constraint imposition) approach based on Dim-itrov’s (2010) guidelines (see Table 3). The first step involved establishing configuralinvariance by examining models for each gender group separately. Results indicatedadequate fit for both groups [men: MLRv2(41, n = 100) = 46.92, p = .24, CFI = .986,RMSEA = .038, SRMR = .055; women: MLRv2(41, n = 174) = 74.27, p = .001, CFI =.944, RMSEA = .068, SRMR = .066]. Measurement invariance was examined next,which involved establishing a baseline model (Model 0: Unconstrained Model), and thentesting for equal factor loading across groups (Model 1: Invariant Factor Loadings Model),equal item intercepts across groups (Model 2: Invariant Factor Loadings and InterceptsModel), and equal item error variances/covariances across groups (Model 3: InvariantFactor Loadings, Intercepts, and Residual Variances Model). Nested models were com-pared using MLR scaled Chi-square difference tests. In Model 0 (M0), no parameters wereconstrained to be equal across the two gender groups. Factor loadings were constrained tobe equal across groups in Model 1 (M1). A nonsignificant MLRDv2 difference between M1and M0 [MLRDv2(8) = 8.08, p = .43] indicated metric invariance (i.e., invariant factorloadings). Both factor loadings and item intercepts were constrained to be equal across twogender groups in Model 2 (M2). The MLRDv2 between M2 and M1 was nonsignificant[MLRDv2(8) = 12.87, p = .12], indicating that the intercepts were also invariant across thetwo gender groups. Model 3 (M3) added constraints for residual item variances/covari-ances to be equal across genders. The nonsignificant MLRDv2 difference between M3 andM2 [MLRDv2(11) = 16.13, p = .14] indicated that item error variances/covariances werealso invariant across genders. Testing structural invariance was the last step where con-straints were added to factor variances and covariances across genders in Model 4 (M4).The MLRDv2 difference between M4 and M2 was nonsignificant [MLRDv2(6) = 7.26,p = .30], supporting structural invariance between genders. In sum, multiple-group CFAresults indicate that the RDS demonstrated the strictest level of measurement and structuralinvariance between men and women in this sample (Figs 1, 2).

Reliability

The reliability for the RDS total and subscale scores were overall adequate, based on theGeneralizability Theory Study (G-study) results. G-study has been highly utilized for itsstrength to distinguish multiple sources of error, which provides reliability of the mea-surement collected in a given study that could be generalized to other situations (Raykovand Marcoulides 2009). The Perceived Prejudice subscale score had generalizabilitycoefficient of .89 in sample 1 and .85 in sample 2. The Closet Symptoms subscale scorehad a generalizability coefficient of .85 in sample 1 and .78 in sample 2. The NegativeLabels subscale score had generalizability coefficient of .82 in sample 1 and .79 in sample2. The coefficient estimates were near .80, which is often used as a rule-of-thumb foracceptable score in GT studies (Mushquash and O’Connor 2006).

J Relig Health

123

Discussion

The purpose of the present study was to develop a reliable and valid measure of perceivedreligious discrimination across two samples. The majority of research on multiculturaltopics related to discrimination has focused heavily on race, LGBT, and gender issues tothe exclusion of religious issues. Accordingly, there are virtually no psychometricallysound measures of perceived religious discrimination that are generalizable across avariety of religious groups. However, religion and religious discrimination are importantfactors to consider in the growing realm of multiculturalism given highly religious indi-viduals who may experience discrimination based on their faith and religious beliefs. Thus,

Fig. 1 Diagram for RDS three-factor oblique model

J Relig Health

123

the RDS scale and subscales were constructed in order to address this gap in the currentpsychological and multicultural measurement regarding religious discrimination.

This study identified three dimensions based on the experiences of LDS individuals. An11-item Religious Discrimination Scale with three different subscales—Perceived Preju-dice, Closet Symptoms, and Negative Labels—was developed. The Perceived Prejudicedimension included discrimination through both overt hostility and subtle avoidance. Thisis similar to many prejudice scales that have split prejudice into both overt and implicittypes (Huntsinger et al. 2016; Cascio and Plant 2015). The Closet Symptoms dimensionfocused on believers’ hesitancies or fears around identifying publicly and openly abouttheir religious faith. This subscale is similar to scales and ideas surrounding coming outwith mental illness (COMIS; Corrigan, et al. 2010) and LGBT ‘‘coming out of the closet’’;however, the RDS items are specifically religious in nature. The Negative Labelsdimension addressed the negative stereotypes and assumptions that others have toward thebeliever’s religious orientation. Other scales have measured Negative Labels, but in thecontext of mental health symptom stigmas (Rockett et al. 2007; Yang et al. 2015). The

Fig. 2 Diagram for RDS bifactor model

J Relig Health

123

correlations between these subscale scores were of medium to strong strength. The cor-relations along with the bifactor model fit results indicate that these three factors haveshared commonalities but represent distinct experiences of religious discrimination.However, due to the inadequate model fit indices of the single-factor model, we do notsuggest using a composite RDS total score.

The mean scores on the three RDS dimensions in this sample were relatively low,especially for Perceived Prejudice (M = 1.97) and Closet Symptoms (M = 1.94). Anendorsed response of 2 on the Likert scale was ‘‘rarely.’’ These low averages could be dueto the fact that this LDS sample in Utah Valley, UT, lived in a Mormon-friendly envi-ronment and community. In other words, they did not feel the need to hide their Mormonfaith (Closet Symptoms) because a majority of others were also LDS. Moreover, they maynot feel like a minority living in Utah and thus would perceive less prejudice from othersaround them. However, these are all speculation because the sample may not be completelybased on this heavily dense LDS community given that the population of students is drawnwidely from all fifty US states and international countries. Their experiences with religiousdiscrimination may have emerged from their home communities which may not be soLDS-friendly. Interestingly, for this sample, the Negative Labels subscale had a mean(M = 3.36) indicating the frequency of experiencing this type of discrimination between‘‘sometimes’’ and ‘‘frequently.’’ It could be that these participants were aware of thenegative labels associated with their religious faith by outsiders or conveyed through themedia. In other words, although the participants might not feel as strong of a perceivedprejudice in their everyday life, or the need to hide their faith, they may be aware ofnegative perceptions and stereotypes of the LDS community through various forms ofmedia such as the Internet, television, movies.

Overall, the psychometric evaluations of the RDS indicated a strong and clear factorstructure as well as good internal consistency reliability. The test of measurement andstructural invariance across gender also suggested that this newly developed RDS scale isequally appropriate to be used with both men and women. With the gender invarianceestablished for the RDS, future studies can better use this scale to examine and comparereligious discrimination experiences between genders. This could be particularly usefulwhen studying religious populations that have more specific gender roles, which may leadto different experiences based on one’s gender. In sum, the RDS has potential to be a usefulmeasure for research related to religion in the field of psychology.

Limitations and Future Directions

Although the results of this study indicate that the RDS could be a promising measure inthe study of religious discrimination, there are some limitations of this study as well asfuture directions for research that should be noted. First, this initial psychometric evalu-ation was conducted with a sample of LDS participants. It would be important and helpfulto validate the psychometric properties of the RDS in other religious populations in futureresearch. For example, it would be important to see if the psychometric properties hold forbelievers of other religions in diverse international contexts, such as Muslims in the USA,Christians in China, or Buddhists in India. Although this sample may have resided in anLDS-friendly environment, LDS individuals may have very different experiencesregarding religious discrimination depending on whether they are living in an LDSmajority community such as Utah or a community where they are a religious minority.

J Relig Health

123

Future studies may use this scale to measure and compare perceived religious discrimi-nation of believers across environments with varying degrees of friendliness/hostilitytoward their particular religious belief.

Implications

The Religious Discrimination Scale (RDS) makes it possible to assess aspects of religiousdiscrimination, which has not been reliably and validly available across religious groupsfor interested researchers, social scientists, and other instrument developers previously.Previous RD scales have focused almost exclusively on Islamic religious discrimination,but the RDS can certainly be applicable and generalizable to other religious affiliations.Additionally, some preceding scales have contained only limited items or focus on reli-gious discrimination, but the RDS is a sole measurement of perceived religious discrim-ination which can add important information and knowledge about this type ofdiscrimination. Our hope is that the RDS can be utilized in a variety of future researchstudies that will further assess elements of religious discrimination across the globe.Furthermore, the RDS could be used to raise awareness about religious discrimination inthe social, political, and mental health communities, and in the general population.

Compliance with Ethical Standards

Conflict of interest All authors declare that they have no conflict of interest

Human and Animal Rights Research with human subjects has been approved by Brigham YoungUniversity’s Institutional Board of Review. All ethical guidelines were followed for this study.

Informed Consent A legal and appropriate consent form was used for this study.

References

Ajrouch, K. J., Reisine, S., Lim, S., Sohn, W., & Ismail, A. (2010). Perceived everyday discrimination andpsychological distress: Does social support matter? Ethnicity & Health, 15(4), 417–434. https://doi.org/10.1080/13557858.2010.484050.

Ali, S. R., Yamada, T., & Mahmood, A. (2015). Relationships of the practice of hijab, workplace dis-crimination, social class, job stress, and job satisfaction among Muslim American women. Journal ofEmployment Counseling, 52(4), 146–157.

Almeida, J., Johnson, R. M., Corliss, H. L., Molnar, B. E., & Azrael, D. (2009). Emotional distress amongLGBT youth: The influence of perceived discrimination based on sexual orientation. Journal of Youthand Adolescence, 38(7), 1001–1014. https://doi.org/10.1007/s10964-009-9397-9.

Ayalon, L., & Gum, A. M. (2011). The relationships between major lifetime discrimination, everydaydiscrimination, and mental health in three racial and ethnic groups of older adults. Aging & MentalHealth, 15(5), 587–594. https://doi.org/10.1080/13607863.2010.543664.

Brondolo, E., Kelly, K. P., Coakley, V., Gordon, T., Thompson, S., Levy, E., et al. (2005). The PerceivedEthnic Discrimination Questionnaire: development and preliminary validation of a community version.Journal of Applied Social Psychology, 35(2), 335–365.

Cano, M. A., Castro, Y., de Dios, M. A., Schwartz, S. J., Lorenzo-Blanco, E. I., Roncancio, A. M., et al.(2016). Associations of ethnic discrimination with symptoms of anxiety and depression among His-panic emerging adults: A moderated mediation model. Anxiety, Stress & Coping: An InternationalJournal, 29(6), 699–707. https://doi.org/10.1080/10615806.2016.1157170.

Cascio, J., & Plant, E. A. (2015). Overt Prejudice Scale [Database record]. Retrieved from PsycTESTS. doi:http://dx.doi.org/10.1037/t38487-000.

Corrigan, P. W., Morris, S., Larson, J., Rafacz, J., Wassel, A., Michaels, P., et al. (2010). Self-stigma andcoming out about one’s mental illness. Journal of Community Psychology, 38(3), 259–275. https://doi.org/10.1002/jcop.20363.

J Relig Health

123

Dimitrov, D. M. (2010). Testing for factorial invariance in the context of construct validation. ThousandOaks, CA: SAGE Publications.

Foynes, M. M., Shipherd, J. C., & Harrington, E. F. (2013). Race and gender discrimination in the Marines.Cultural Diversity and Ethnic Minority Psychology, 19(1), 111–119. https://doi.org/10.1037/a0030567.

Ghumman, S., Ryan, A. M., Barclay, L. A., & Markel, K. S. (2013). Religious discrimination in theworkplace: A review and examination of current and future trends. Journal of Business and Psy-chology, 28(4), 439–454.

Greer, T. M. (2011). Coping strategies as moderators of the relationship between race-and gender-baseddiscrimination and psychological symptoms for African American women. Journal of Black Psy-chology, 37(1), 42–54. https://doi.org/10.1177/0095798410380202.

Hahm, H. C., Ozonoff, A., Gaumond, J., & Sue, S. (2010). Perceived discrimination and health outcomes: Agender comparison among Asian-Americans nationwide. Women’s Health Issues, 20(5), 350–358.https://doi.org/10.1016/j.whi.2010.05.002.

Harrell, S. P. (1994). The racism and life experience scales. Unpublished Manuscript. Harrell, S. from CE202 at Texas A&M.

Hodge, D. R. (2005). The Spiritual Competence Scale: Validating a population-specific measure of culturalcompetence with a faith-based sample. Advances in Social Work, 6(2), 290–304.

Hodge, D. R. (2007). Religious discrimination and ethical compliance: Exploring perceptions among aprofessionally affiliated sample of graduate students. Journal of Religion & Spirituality in Social Work:Social Thought, 26(2), 91–113.

Hudson, D. L., Neighbors, H. W., Geronimus, A. T., & Jackson, J. S. (2016). Racial discrimination, JohnHenryism, and depression among African Americans. Journal of Black Psychology, 42(3), 221–243.https://doi.org/10.1177/0095798414567757.

Huntsinger, J. R., Sinclair, S., Kenrick, A. C., & Ray, C. (2016). Implicit Prejudice IAT [Database record].Retrieved from PsycTESTS. doi: http://dx.doi.org/10.1037/t49152-000.

Huynh, V. W., Guan, S. A., Almeida, D. M., McCreath, H., & Fuligni, A. J. (2016). Everyday discriminationand diurnal cortisol during adolescence. Hormones and Behavior. https://doi.org/10.1016/j.yhbeh.2016.01.009.

Jones, K. P., Peddie, C. I., Gilrane, V. L., King, E. B., & Gray, A. L. (2016). Not so subtle: A meta-analyticinvestigation of the correlates of subtle and overt discrimination. Journal of Management, 42(6),1588–1613. https://doi.org/10.1177/0149206313506466.

Jordanova, V., Crawford, M. J., McManus, S., Bebbington, P., & Brugha, T. (2015). Religious discrimi-nation and common mental disorders in England: A nationally representative population-based study.Social Psychiatry and Psychiatric Epidemiology, 50(11), 1723–1729.

Kessler, R. C., Mickelson, K. D., & Williams, D. R. (1999). The prevalence, distribution, and mental healthcorrelates of perceived discrimination in the United States. Journal of Health and Social Behavior,40(3), 208–230.

Kim, I., & Noh, S. (2014). Ethnic and gender differences in the association between discrimination anddepressive symptoms among five immigrant groups. Journal of Immigrant and Minority Health, 16(6),1167–1175. https://doi.org/10.1007/s10903-013-9969-3.

Lund Dean, K., Safranski, S. R., & Lee, E. S. (2015). Bad apples, behaving badly? Workplace reli-gious discrimination disputes and resolution behaviors. Journal of Management, Spirituality & Reli-gion, 12(3), 227–256.

Lyons, A. (2016). Mindfulness attenuates the impact of discrimination on the mental health of middle-agedand older gay men. Psychology of Sexual Orientation and Gender Diversity, 3(2), 227–235. https://doi.org/10.1037/sgd0000164.

McDonald, J. A., Terry, M. B., & Tehranifar, P. (2014). Racial and gender discrimination, early life factors,and chronic physical health conditions in midlife. Women’s Health Issues, 24(1), e53–e59. https://doi.org/10.1016/j.whi.2013.09.006.

Molina, K. M., Jackson, B., & Rivera-Olmedo, N. (2016). Discrimination, racial/ethnic identity, and sub-stance use among Latina/os: Are they gendered? Annals of Behavioral Medicine, 50(1), 119–129.https://doi.org/10.1007/s12160-015-9738-8.

Molina, K. M., & James, D. (2016). Discrimination, internalized racism, and depression: A comparativestudy of African American and Afro-Caribbean adults in the US. Group Processes & IntergroupRelations, 19(4), 439–461. https://doi.org/10.1177/1368430216641304.

Mushquash, C., & O’Connor, B. P. (2006). SPSS and SAS programs for generalizability theory analyses.Behavior Research Methods, 38(3), 542-547. Retrieved from https://search.proquest.com/docview/204305459?accountid=4488.

J Relig Health

123

Myrick, S. E., & Martorell, G. A. (2011). Sticks and stones may break my bones: Protective factors for theeffects of perceived discrimination on social competence in adolescence. Personal Relationships,18(3), 487–501. https://doi.org/10.1111/j.1475-6811.2010.01320.x.

Netemeyer, R., Beardon, W., & Sharma, S. (2003). Scaling procedures: Issues and applications. ThousandOaks, CA: Sage.

Otiniano Verissimo, A. D., Gee, G. C., Ford, C. L., & Iguchi, M. Y. (2014). Racial discrimination, genderdiscrimination, and substance abuse among Latina/os nationwide. Cultural Diversity and EthnicMinority Psychology, 20(1), 43–51. https://doi.org/10.1037/a0034674.

Perry, B. L., Harp, K. H., & Oser, C. B. (2013). Racial and gender discrimination in the stress process:Implications for African American women’s health and well-being. Sociological Perspectives, 56(1),25–48.

Raykov, T., & Marcoulides, G. A. (2009). Estimation of generalizability coefficients via a structuralequation modeling approach to scale reliability evaluation. International Journal of Testing, 6(1),81–95. https://doi.org/10.1207/s15327574ijt0601_5.

Reisner, S. L., White Hughto, J. M., Gamarel, K. E., Keuroghlian, A. S., Mizock, L., & Pachankis, J. E.(2016). Discriminatory experiences associated with posttraumatic stress disorder symptoms amongtransgender adults. Journal of Counseling Psychology, 63(5), 509–519. https://doi.org/10.1037/cou0000143.

Rippy, A. E., & Newman, E. (2007). Perceived religious discrimination and its relationship to anxiety andparanoia among Muslim Americans. Journal of Muslim Mental Health, 1(1), 5–20. https://doi.org/10.1080/15564900600654351.

Rockett, J. L., Murrie, D. C., & Boccaccini, M. T. (2007). Diagnostic labeling in juvenile justice settings: Dopsychopathy and conduct disorder findings influence clinicians? Psychological Services, 4(2),107–122. https://doi.org/10.1037/1541-1559.4.2.107.

Rosenthal, L., Earnshaw, V. A., Lewis, T. T., Reid, A. E., Lewis, J. B., Stasko, E. C., et al. (2015). Changesin experiences with discrimination across pregnancy and postpartum: Age differences and conse-quences for mental health. American Journal of Public Health, 105(4), 686–693. https://doi.org/10.2105/ajph.2014.301906.

Schmitt, M. T., Branscombe, N. R., Kobrynowicz, D., & Owen, S. (2002). Perceiving discrimination againstone’s gender group has different implications for well-being in women and men. Personality andSocial Psychology Bulletin, 28(2), 197–210.

Sheridan, L. P. (2006). Islamophobia Pre- and Post-September 11th, 2001. Journal of Interpersonal Vio-lence, 21(3), 317–336.

Spence, N. D., Wells, S., Graham, K., & George, J. (2016). Racial discrimination, cultural resilience, andstress. The Canadian Journal Of Psychiatry/La Revue Canadienne De Psychiatrie, 61(5), 298–307.https://doi.org/10.1177/0706743716638653.

Tabachnick, B. G., & Fidell, L. S. (2007). Using multivariate statistics (5th ed.). Boston, MA: Allyn &Bacon/Pearson Education.

Thijs, J., & Piscoi, D. (2016). Perceiving discrimination in ‘real life’: Distinguishing negative events fromdiscrimination attributions. Basic and Applied Social Psychology, 38(3), 166–172. https://doi.org/10.1080/01973533.2016.1186027.

van der Straten Waillet, N., & Roskam, I. (2012). Religious discrimination in childhood and adolescence.Archiv fur Religionspsychologie/Archive for the Psychology of Religions, 34(2), 215–242.

Williams, D. R., Yu, Y., Jackson, J., & Anderson, N. B. (1997). Racial differences in physical and mentalhealth: Socio-economic status, stress, and discrimination. Journal of Health Psychology, 2, 335–351.https://doi.org/10.1177/135910539700200305.

Woodford, M. R., Han, Y., Craig, S., Lim, C., & Matney, M. M. (2014). Discrimination and mental healthamong sexual minority college students: The type and form of discrimination does matter. Journal ofGay & Lesbian Mental Health, 18(2), 142–163. https://doi.org/10.1080/19359705.2013.833882.

Yang, L. H., Link, B. G., Ben-David, S., Gill, K. E., Girgis, R. R., Brucato, G., Wonpat-Borja, A. J., &Corcoran, C. M. (2015). Stigma of symptoms measure [Database record]. Retrieved from PsycTESTS.doi: http://dx.doi.org/10.1037/t50896-000.

J Relig Health

123