Final Project Amelia Noor Oshiro

A Cross-Sectional Pilot Study Proposal: Exploring the Links Between Being Visibly Muslim, Stress Level, and Perceived Racism Among Muslim American Students at Majority-White and non-Majority White Institutions

Final Project Quantitative Research Design for the Social Sciences Dr. Patrick A. Wilson

Amelia Noor-Oshiro Department of Sociomedical Sciences Mailman School of Public Health Columbia University in the City of New York May 23, 2016



CONTENTS



I.

BACKGROUND & AIMS ………………………………….……….…… 2

II.

STUDY DESIGN ………………………………………………….…….. 6

III.

SAMPLING APPROACH …………………………………………..…… 7

IV.

DATA COLLECTION METHOD ……………………………………… 8

V.

MEASURES ………………………………………………………………9

VI.

DATA ANALYSIS PLAN …………………………………………..……13

VII.

STRENGTHS & LIMITATIONS …………………………………………16

VIII.

REFERENCES ………………………………………………………...… 18

1



I.

BACKGROUND & AIMS A. TITLE Exploring the Links Between Being Visibly Muslim, Stress Level, and Perceived Racism Among Muslim American College Students B. BACKGROUND Prior research studies have established a basis for the causal pathway of stress faced by minorities, whether due to race or stigmatized identity (such as sexual orientation) (Calabrese 2014). This stress—sometimes referred to as ‘weathering’, ‘minority stress’, and/or ‘allostatic load’—manifests physiologically, resulting in negative health outcomes (Duru 2012). Presently, minimal extant research exists pertaining to how such stress affects religious minorities, namely Muslim Americans, in the post-9/11 era. Although Muslim Americans constitute a religious minority, and should thus be studied as one, the mainstream media and academia within the United States fails to distinguish Muslims as a religious group due to various reasons. Instead, it should be noted that, since 9/11, Muslim Americans have been referred to as a homogenous racial/ethnic group as opposed to a diverse faith group like Christian Americans (Meer 2014). The implicit conflation of a religious identity to a racial/ethnic identity lends itself to the concept of the “racialization of Muslims” (Meer 2014). Moreover, identifying as Muslim in the post-9/11 period—or even being perceived in public as identifying as such—is becoming an increasingly stigmatized due to a negative social climate fostered against Muslims through media outlets and political officials. Thus, this survey establishes two major points before continuing with its objectives: that identifying as Muslim is (1) a “racialized” concept, and therefore the study design will reflect race-based methodologies and measurements in order to align with this concept; and (2) a highly stigmatized identity, presumably beholding a milieu of tribulations for Muslim Americans.



2



In order to get a more crisp understanding of how stress levels may or may not vary among Muslim Americans, the study seeks to contrast two different populations: those who are visibly Muslim and those who are not. Additionally, it contrasts two different social environments: White Majority Institutions and Non-White Majority Institutions. The survey seeks to assess whether stress levels vary among those who are visibly Muslim and also in order to find out whether their social environment plays a role in the level of stress they experience in order to understand the bigger picture of how differing social environments may predict stress levels for Muslim American college students in the U.S. at large. C. SIGNIFICANCE In the entire field of public health, there is little to no existing research on the topic of the stigmatization and/or racialization of religious minorities in America, much less Muslim Americans. Currently, most of the literature involving religion as a factor in public health studies relates to the role of religion in determining a health outcome (e.g. Catholicism and abortion). There is a dearth of research indicating how religious minorities manage to navigate everyday health issues in an environment that is structurally unsupportive to them (i.e. Muslim Americans in a Christian-dominant societal structure) (Nadal 2012). Therefore, this pilot study acts as a starting point for future research that seeks to investigate potential health outcomes of religious stigmatization and racialization, namely for the Muslim American population. Such studies can be implemented in the future as models for other religious minorities such as Jews, Sikhs, Hindus, and other nonChristian/Catholic faiths, and especially faiths that have been, or currently are, stigmatized and/or racialized (e.g. Jewish people). Research on religious minorities is imperative in order to discover unobserved public health concerns and to develop appropriate interventions, especially with regard to the importance of intersectionality. While many minority identities exist (race, sexual

3

orientation, gender), it is possible that they exist within the context of religious identity as well; therefore, it is important to consider faith as a factor in issues affecting public health. Once knowledge is available on the role of religious minority status as it relates to health, public health strategies can utilize intersectionality to improve interventions, thus benefiting minority communities as a whole. D. RESEARCH QUESTION The study seeks to understand stress that occurs as a result of being perceived as Muslim. Theoretically, this perception would be especially stigmatized in environments that amplify an individual’s ‘otherness’, as would likely be the case at a Majority White Institution. After further studies are conducted, we hope to be able to generalize the results to the greater Muslim American college-attending student population at large. For the purpose of this study, being Muslim will be treated as a racialized construct, as it is often the case that non-Muslims conflate the identity of adherents to the Islamic faith with racial/ethnic identities (e.g. Middle Eastern, Arab). Therefore, the research seeks to investigate: (1) the possible association between the variables of being ‘visibly Muslim’ and ‘stress level’, questioning the variable of ‘perceived racism’ as a mediator; and (2) the variable of ‘attendance at a Majority White Institution’ as a moderator between being ‘visibly Muslim’ and ‘stress level’. E. AIMS & HYPOTHESES 1. Aim 1: To test a model linking ‘visibly Muslim’ and ‘stress level’, using ‘perceived racism’ as the mediator. Hypothesis 1: Muslim American college students who are visibly Muslim will report a higher perception of racism, leading to higher stress levels. 2. Aim 2: To test a model linking ‘visibly Muslim’ and ‘stress level’, using ‘attendance at Majority White Institution’ as a moderator. Hypothesis 2: Attending a Majority White Institution will strengthen the relationship



4

between visibly Muslim and stress level among Muslim American college students. 3. Notes: i. The variable of ‘visibly Muslim’ will be dichotomous; it is a score calculated from survey responses based on a Likert scale of 1-7, where a mean score of 5.5+ indicates clearly visibly Muslim, and 5.4 and below as unclearly visibly Muslim. ii. Majority White Institutions (MWI) will be defined as schools with 65% or more White students; Non-Majority White Institutions (NMWI) will be defined as Historically Black Universities (HBU), Hispanic Serving Institutions (HSI), and colleges with 65% or more students of color. F. CAUSAL MODEL









Perceived Racism

Stress Level

Being Visibly Muslim

Attendance at White Institution Causal Model of Relationship Between Being Perceived as Muslim and Stress Level, with Perceived Racism as a Mediator and Attendance at White Institution as a Moderator

II.

STUDY DESIGN A. METHODOLOGICAL APPROACH



5

This will be a correlational and analytical observational study that will utilize an online survey to address the research aims. It will be conducted cross-sectionally as a group comparison to observe possible differences in correlation between the independent variable, “visibly Muslim,” and dependent variable, “stress level,” when populations are situated in the context of a Majority White Institution (MWI) or a Non-Majority White Institution (NMWI).

B. RATIONALE As it is a pilot study, the focus is to first gather information about the Muslim American college population and its levels of stress based on perceived racism and also attendance at a majority or non-majority White institution; however, a longitudinal study to explore other possible causal pathways to the dependent variable of stress should be established to understand the nuances in the relationship between being perceived as Muslim and a Muslim American college student’s stress level.



6



III.

SAMPLING APPROACH A. IDENTIFYING THE POPULATION & SAMPLE FRAME The target population is Muslim American college students who attend undergraduate college institutions in the U.S., full-time and part-time, between the ages of 17 – 24. This constitutes the population construct to which the findings will be generalizable. B. SAMPLING METHOD A non-probability quota sample will be utilized, as this is a stratified convenience sample in that the stratification occurs by distinguishing institutions based on homogeneity of White student majority vs. non-White student majority. The sample population will be extracted from those institutions based on their membership to the Muslim Student Association on campus. C. STEPS TO OBTAIN SAMPLE I will contact large conglomerate MSA organizations to obtain a list of MSA organizations at various institutions in the U.S. I will also conduct an Internet search to obtain any missing MSA organizations to ensure that a large representative sample of schools is selected to meet the quotas for MWIs and NMWIs. Then, I will assemble a database of Muslim American college student members to assemble membership lists. D. RATIONALE Considering that the population of Muslim Americans in the U.S. is around 7 million (Pew 2016),1 and the number of Muslim Americans who attend college is just a proportion of that number, the target population of Muslim American college students is not only small, but also potentially difficult to access outside the realm of MSA. Thus, a convenience sample from within the MSAs would facilitate access to this small population in order to capture the largest sample size possible.





7



IV.

DATA COLLECTION METHOD A. METHOD Considering the age range and context of specifically targeting college students, the most appropriate and cost-efficient method is a self-administered web-based survey. Following the compilation of the list of students in the sample, an email will be sent with information, a link, and instructions to the survey. The respondents will be encouraged to forward the link to any other Muslim students in their network via social network sites and or e-mail. The email invitation to the survey will be sent two more subsequent times at intervals of one week in order to maximize response rate.

B. RATIONALE A web-based survey best fulfills the aims of the research proposal, as it will facilitate a quick turnaround for the cross-sectional “snapshot” the pilot study seeks. As the items on the survey will be self-explanatory and or guided by instructions, a web-based survey is more suitable than an in-person interview or other method requiring inperson or live responses. Being web-based may also mitigate any biases related to interviewer characteristics and or the respondent’s social desirability in answering questions truthfully. The web-based survey may also facilitate greater accessibility to the population sampled, possibly maximizing response rate. The survey will be very costeffective by using Google Forms, a free survey-analyzing software.



8



V.

MEASURES A. FIRST MEASURE. INDEPDENDENT VARIABLE: VISIBLY MUSLIM Defining “visibly Muslim” is challenging, especially when considering multidimensional factors such as skin tone, hair type, facial features, and religious garb as just some of the subjective methods of assessment. However, in order to ensure that the construct of visibly Muslim would withhold, it was imperative that we focus only on accessory-type items to include in our measurement. This meant that we disregarded skin tone, hair type, facial features or any other phenotypic clues and characteristics that would supposedly indicate that someone is a Muslim. Instead, by focusing only on accessories such as religious garb, clothing, and characteristics of the beard (which is arguably an accessory comparable to the hijab for women), it allowed our construct to be more sound and measurable. Therefore, in order to clearly demarcate who would be clearly perceived as a Muslim, we chose to establish a new measure that allows the respondent to answer a series of questions regarding their status of being ‘visibly’ Muslim. Some sample items are included in the following page.



9

Instructions: These set of statements seek to assess the extent to which you are visibly Muslim. Please respond as truthfully and accurately as possible. Keep in mind that all of your responses are anonymous. If you are unsure, please select “unsure.” Some questions may only apply to males or females, so please answer accordingly. Use the following scale for your responses to all questions.

1 1

UNSURE

2 VERY UNTRUE OF ME

3 UNTRUE OF ME

4 SOMEWHAT UNTRUE OF ME

5 SOMEWHAT TRUE OF ME

6 TRUE OF ME

7 VERY TRUE OF ME

MALES Statement 1: Every day when I am out on campus, I always wear Islamic garb indicating that I am Muslim, such as a thawb or kurta (note: not long shirts, loose pants or rolled up pant bottoms—refer to image (L). Statement 2: Every day when I am out on campus, I always wear an item of clothing indicating that I am Muslim, such as a kufi/topi (note: not jewelry, tattoos— refer to image (R). Statement 3: Every day when I am out on campus, I always wear a hat, cap or beanie or some other Western style of head-covering for the purpose of modesty (refer to image (R). [Reverse code this item.] Statement 4: Every day when I am out on campus, I always wear facial hair that is noticeably less than one inch under my chin (note: this includes a clean-shaven face and beard stubble—refer to previous image.) [Reverse code this item.] Statement 5: Every day when I am out on campus, I always wear a beard that is noticeably longer than one inch below my chin, indicating that I am Muslim (refer to image (L). Statement 6: Every day when I am out on campus, I always wear my hair openly, without any covering for the purpose of modesty. [Reverse code this item.] Statement 7: Every day when I am out on campus, I always wear a turban that is similar to the Sunnah style of the Prophet (refer to image (R).

10

FEMALES Statement 1: Every day when I am out on campus, I always wear Islamic garb indicating that I am Muslim, such as an abaya, niqab or long dress (note: not long shirts, loose pants, or indowestern clothing—refer to image (R). Statement 2: Every day when I am out on campus, I always wear a head-covering that is wrapped similarly to a bun, turban or hat/beanie for the purpose of modesty (note: not a hijab—refer to image (L). [Reverse code this item.] Statement 3: Every day when I am out on campus, I always wear a hijab that covers my neck, indicating that I am Muslim (note: not jewelry, tattoos, or hair buns/turbans—refer to image (L). Statement 4: Every day when I am out on campus, I always wear my hair openly, without any head covering for the purpose of modesty. [Reverse code this item.]

B. RATIONALE FOR FIRST MEASURE Thus, for this study, measuring the visibility of Muslimness is actually connected to how we would hope non-Muslims identify and distinguish Muslims from other groups (racial, ethnic or religious). The logic in this train of thought is that the more a person is visibly Muslim through their apparent use of accessories, the more likely he would be perceived as a Muslim—a stigmatized identity—ergo, the more likely he or she will experience the stressors that come with exposing said identity. Some other considerations to measure this construct include the Racism Reaction Scale, which was developed specifically for minority students attending predominantly White universities. Its inventory measures items such as ‘sense of being personally threatened’, ‘differentially treated’, or ‘singled out for differential treatment’ all based on race. While the RRS is particularly well suited for this study, it will not be used as the primary measurement tool for the construct of visibly Muslim; however, should the survey seek to gain complementary information for questions regarding how individuals may be perceived as Muslim by their peers, then the use of the items from this scale would be appropriate. When employing such items, there would be a need to adapt



11

the measure to reflect “Muslim” instead of “Black/African American”; it will also assume the racialization of the term “Muslim” as it is a race-based measure.

C. SECOND MEASURE. DEPENDENT VARIABLE: STRESS LEVEL The definition and measurement of “stress level” encompasses multiple dimensions. First, the survey must seek to establish a distinction between stress that occurs in the life of a typical college student from the social stress that relates to the stigmatization of Muslim identity that we seek. Therefore, the definition of stress involves such factors as hypervigilance of one’s surroundings, trauma or fear of trauma, fear of or stress related to discrimination and or racism (which will be explored through the meditational model of perceived racism) and other racism-related stress. Ultimately, the objective is to assess whether Muslim American college students experience a statistically different level of stress based on their appearance of being visibly Muslim and whether attendance at a MWI or NWMI moderates their stress level. D. RATIONALE FOR SECOND MEASURE The Racism and Life Experience Scale (RaLES), a five-scale measure of racism-related stress, coping, socialization and attitudes, assesses levels of stress. The RaLES uniquely incorporates both context-dependent and stable-over-time scales, which is what distinguishes it from similar overlapping constructs. This measure is specifically fitting to the construct of stress for this study due to its incorporation of frequency, degrees of influence, and an inventory of racism occurrences while also capturing direct, vicarious and collective racism. The RaLES will be slightly adapted to include terminology relating to religion as opposed to exclusively race as some of the questions will need wording to specifically identify the participant as a Muslim and adherent of Islam; for example, “Rate how often you had the experience and how stressful the experience was if/when there was a conflict between you and someone of a different religion.” Other questions relating exclusively to race, however, will remain unchanged as they may not need to be specifically addressing religion alone; for example, “Rate how often you had the experience and how stressful the experience was if/when you heard an offensive joke.”



12

VI.

DATA ANALYSIS PLAN A. AIM 1, VARIABLES, HYPOTHESIS, DATA ANALYSIS, RATIONALE Aim 1: To test a model linking ‘visibly Muslim’ and ‘stress level’, using ‘perceived racism’ as the mediator. Hypothesis 1: Muslim American college students who are visibly Muslim will report a higher perception of racism, leading to higher stress levels. Variables: The independent variable is perceived as Muslim. The dependent variable is stress level in relation to the independent variable of perceived as Muslim and also in relation to perceived racism. Mediating Variable. The mediator is perceived racism. Data Analysis: In regards to recoding, the measure that calculates a score on the scale of “visibly Muslim” will be utilized to create a dichotomous variable through the following method: if the score is lower than 5.50, the individual will be categorized as a “0” to indicate not visibly Muslim; if the score is 5.50 or greater, then the individual will be categorized as “1” to indicate visibly Muslim. The rationale for this cutoff is that we are looking for only extreme or nearly extreme responses to the items related to “visibly Muslim” otherwise there would be too much ambiguity and clouding of who or what constitutes “visibly” Muslim. To ensure that a person would be—without a doubt—visibly Muslim, only the highest values (6 or 7) for means would ideally be isolated; however, to add nuance to these extreme values, a half-point cutoff of 5.5 was instead used. In other words, while some Muslims may definitely be visibly Muslim all of the time (answers correspond to “true of me and very true of me,” there may be some questions that lower their average score (e.g. they may not wear a conservative outfit every single day, but most days) and, thus, that should not disclude them from being—overall—a very visible Muslim (refer to example statements in measurement section). In regards to tests, there are two pathways that need to be examined—see diagram on the next page.



13



Perceived Racism

two-sample t-test linear regression

continuous

Correlation multiple regression

Being Visibly Muslim

Stress Level continuous

dichotomous

two-sample t-test linear regression

a. The first pathway is ‘visibly Muslim’ causing ‘stress level’; for this pathway, the tests are (1) two-sample t-test and (2) linear regression. i. Rationale: (1) First, a two-sample t-test will compare the means of the visibly Muslim and not visibly Muslim variables to stress level in order to determine whether the the mean scores between the two groups are statistically significant; (2) a linear regression test will explain how much of the stress level, the continuous outcome variable, is explained by being visibly Muslim compared to those not being visibly Muslim. b. The second pathway involves controlling for ‘perceived racism’ as the mediator between ‘visibly Muslim’ and ‘stress level’; for this pathway, there are several steps for tests. First, the relationship of being visibly Muslim and perceived racism will be evaluated with (1) two-sample t-test and (2) linear regression. Then, the relationship of being visibly Muslim and perceived racism will be evaluated with (3) correlation and (4) multiple regression. i. Rationale: (1) Conducting a two-sample t-test to compare mean scores between visibly and not-visibly Muslim will indicate whether their perception of racism is statistically significantly different; (2) conducting a linear regression will explain how much of the perceived racism, the continuous outcome variable, is explained by being visibly Muslim compared to not being visibly Muslim. Next, a (3) correlation test will establish whether there is a relationship between perceived racism and stress level; and (4) multiple regression will involve the input variables of ‘visibly Muslim’ and ‘perceived racism’ to see which is a better predictor of stress level.

14



B. AIM 2, VARIABLES, HYPOTHESIS, DATA ANALYSIS, RATIONALE Aim 2: To test a model linking ‘visibly Muslim’ and ‘stress level’, using ‘attendance at Majority White Institution’ as a moderator. Hypothesis 2: Attending a Majority White Institution will strengthen the relationship between visibly Muslim and stress level among Muslim American college students. Variables: The independent variable is perceived as Muslim. The dependent variable is stress level in relation to the independent variable of perceived as Muslim and also in relation to perceived racism. Mediating Variable. The mediator is perceived racism. Data Analysis: In regards to recoding, those attending a NMWI will be categorized as “0” and those at a MWI will be categorized as “1” as this is a dichotomous variable. In regards to tests, there is only one causal pathway that needs evaluation for this moderational model: determining whether attendance at a MWI strengthens the relationship between being visibly Muslim and stress level (presumably there is a higher stress level for those who are visibly Muslim at a MWI).

Attendance at MWI dichotomous Being Visibly Muslim

Stress Level continuous

dichotomous

logistic regression chi-squared

a. The tests involved are: (1) logistic regression and (2) chi-squared. ii. Rationale: (1) Conducting a logistic regression will tell us if the odds of stress level increase for visibly Muslim students attending MWI compared to visibly Muslim students attending NMWI; it will also tell us the odds of not being visibly Muslim and stress level at a MWI and NMWI. (2) By conducting a chi-squared test, we determine whether there is any significant difference among the populations being tested, i.e. visibly and not-visibly Muslim at MWI and NMI institutions.



15



VII.

STRENGTHS & LIMITATIONS A. STUDY DESIGN The main threats to validity for the study design are (a) restriction of range in which the spectrum of ‘visibly Muslim’ is not properly represented due to the fact that not enough variability is present in the population sampled and (b) selection bias in that respondents must be members of the Muslim Student Association and thus disclude those who are Muslim but not members. However, because it is a pilot study, it is easy, efficient and inexpensive. Since the study design aims to generalize results to the greater population of Muslim American college students, and therefore focus on external validity, the tradeoff, then, is a compromise to internal validity. Although there are threats affecting the validity of the study design, it is still an acceptable and worthy study to conduct because it can be reproduced for other marginalized populations using similar variables. The most important justification of this study design is to have a starting point in which to identify and infer the possible allostatic load of the religious minority of Muslim American college students. B. SAMPLING APPROACH It is anticipated that the obtained sample will be quite representative of the sample frame and more broadly the population to which it will be generalized, however, there could be institutions that have Muslim students but don’t have MSAs, and/or MSAs that exist on campus but have no internet presence (therefore creating gaps in the database), and also that mostly practicing Muslims are part of the MSA, therefore those who aren’t practicing but still identify as Muslim may be excluded. These all contribute to selection bias. I anticipate, however, that these will be minority populations and will not have a significant effect on generalizability. C. DATA COLLECTION METHOD Although many college students have access to computers, there may be a sizeable number of individuals who lack such access; other individuals may find the Google



16

software as unfamiliar or challenging. This may lead to a higher refusal rate than expected. Also, the lists from which the email addresses of the population are gathered may be flawed, dated, or incorrect. Lastly, there is no control over the environment or circumstances under which the respondent takes the survey. The tradeoffs in choosing the web-based data collection strategy is justified, then, in that this is a pilot study. Finally, as the nature of the items assessed is somewhat sensitive, it is imperative to consider that an online format may put respondents at ease. D. MEASURES The primary limitation for the construct invented to measure “visibly Muslim” is that there is no established construct validity or content validity. The scope of the items and therefore item representativeness may also be inadequate to fully capture the range of the variable that the study seeks. This poses a threat to the strengths of inferences between the dependent and independent variables. The primary limitation for the construct measuring stress level is that it is long and may be discouragingly timeconsuming for survey respondents. E. DATA ANALYSIS Should the results of the tests be validating to the causal model, the test strategies enumerated would create a strong case to establish further studies that will bring more evidence for a significant, correlational and predictive relationship between the independent and dependent variables. One weakness, however, includes the loss of variability in data due to recoding for a dichotomous independent variable. Further stratifications of the population (not visibly Muslim but apparently Muslim in an otherwise apparent fashion) could yield more nuanced results; this population is excluded from the population and could act as a confounder or even as a second moderator to the relationship between being visibly Muslim and stress level and the context of attending a MWI. The strength, however, is that there is at least a starting point with which to begin data analysis.



17



VIII.

REFERENCES Calabrese, S.K. et. al. (2014). Exploring discrimination and mental health disparities faced by Black sexual minority women using a minority stress framework. Psychology of Women Quarterly. doi#10.1177/0361684314560730 Duru, O.K. et. al. (2012). Allostatic load burden and racial disparities in mortality, Journal of the National Medical Association, 104(1-2), pp. 89-95. Meer, N. and Modood, T. (2014). For “Jewish” read “Muslim”? Islamophobia as a form of racialization of ethno-religious groups in Britain today. Islamophobia Studies Journal, 1(1), pp. 34-53. Mohamed, B. (2016) A new estimate of the U.S. Muslim population. Retrieved March 4, 2016, from http://www.pewresearch.org/fact-tank/2016/01/06/a-new-estimate-of-the-u-smuslim-population/ Nadal, K., et. al. (2012). Subtle and overt forms of Islamophobia: Microaggressions toward Muslim Americans. Journal of Muslim Mental Health. 6(2). Doi: http://dx.doi.org/10.3998/jmmh.10381607.0006.203



18