Instruments

Racial Microaggressions Scale (RMAS; Torres-Harding, Andrade Jr, &

Romero Diaz, 2012; Appendix D). The RMAS is a self-reporting scale that assesses the frequency and the distress evoked by experiences of racial microaggressions in respondent’s everyday lives. This scale consists of two sets of questions. The first set assesses the frequency of racial microaggressions incidents, and the second set examines the distress elicited by each category of microaggression experience. The original scale consists of six subscales with 32 items. However, in this study, the Environmental Invalidations subscale was deleted, and the contents of some items were revised in order to adapt to international students in Taiwan. The revision of RMAS includes 24 items and five subscales, which are:

(1) Foreigner/Not Belonging, how often individuals reported being treated as a foreigner or as if they didn’t belong in a given setting because of one’s racial background

(2) Criminality, how frequently individuals reported others treating them as if they were aggressive or likely to engage in criminal behavior;

(3) Sexualization, how frequently individuals reported that others treated them in an overly sexual manner or sexually stereotyped them because of their race;

(4) Low-Achieving/Undesirable Culture, how frequently individuals were treated as if their racial culture was dysfunctional, low achieving, or undesirable, and how frequently others viewed academic success as due to preferential treatment; and

(5) Invisibility, how frequently individuals reported that others overlooked, invalidated, or dismissed their views or contributions because of their racial heritage, and the extent to which they felt marginalized by others.

For each item on the subscales, participants were asked “How often does this happen to you?” with no specific time frame given. All of the items are rated on a 4-point Likert scale with 0 = never, 1 = a little/rarely, 2 = sometimes or a moderate amount, and 3 = often/frequently, based on how often participants have encountered a particular racial microaggression. A sample item is: “I receive poorer treatment in restaurants and stores because of my race.” A total score (ranging from 0 to 72) is the sum of all five subscales, with higher scores representing higher frequency of experiencing racial microaggressions.

In this study, the reliability of the RMAS and distress subscales was examined using the Cronbach’s alpha to ensure the internal consistency of the subscales within this study.

It is presented that the RMAS subscales had good reliability: Foreigner/Not Belonging (α

= .67); Criminality (α = .79); Sexualization (α = .89); Low Achieving/Undesirable (α

= .83); and Invisibility (α = .82).

In order to establish the content validity of RMAS, the researcher invited three professors to examine the wording, suitability, and importance of each item from all the measurements, and to give advice of whether to add new item, to revise, or delete a certain item. All of these professors possess a doctoral degree in counseling psychology in Taiwan or in the United States, have cross-cultural experience, and/or is an expert of international student study.

Convergent validity of the original RMAS was examined by Torres-Harding, Andrade Jr, and Romero Diaz (2012) by comparing mean scores of each subscale with the subscales of the Schedule of Racist Events (SRE; Landrine & Klonoff, 1996). The result indicated that all of the RMAS subscales positively correlated with the three SRE

subscales. The SRE is an 18-item self-report instrument specific, negative life events and race-related stressors that occur to African Americans, and was later modified to be applicable more generally to people of color. As the revised RMAS only has subscales contained in the original, this should apply to the revised RMAS as well.

Microaggressions distress subscales. The second set of questions in the RMAS was derived examining the distress elicited by each category of microaggression experience.

Each item includes an additional query assessing how stressful, bothersome, or upsetting respondents found a particular incident is (i.e. if this does happen to you, how stressful, upsetting, or bothersome is this for you?). The distress subscales are Foreigner/Not Belonging Distress, Criminality Distress, Sexualization Distress, Low Achieving/Undesirable Distress, and Invisibility Distress.

Individuals first answer RMAS items. If individuals reported experiencing a particular racial incident at all (i.e. at least a 1 for the frequency of the experience), they were then asked to indicate their distress level to the particular item; if they reported not experiencing a microaggression, they were coded as a ‘0’ for the distress item. The score for each distress subscale was computed by the means of the items composing each subscale that were answered. However, if a frequency item associated with the distress item was endorsed as Never, then the distress response was omitted from the calculation of the distress score.

Reliability of the revised RMAS distress subscales were computed using the Cronbach’s alpha to ensure the internal consistency of the subscales. All of the five distress subscales were found to have good reliability: Foreigner/Not Belonging Distress (α = .74); Criminality Distress (α = .88); Sexualization Distress (α = .90); Low Achieving/Undesirable Distress (α = .86); and Invisibility Distress (α = .86). The coefficient alpha was > .80 for each subscale item except Foreigner/Not Belonging (α

= .74).

Additionally, in this study, all five frequency subscales except Invisibility were found to be significantly and positively correlated with their corresponding distress subscales:

Foreigner/Not Belonging and Foreigner/Not Belonging Distress (r = .22, p < .01);

Criminality and Criminality Distress (r = .49, p < .01); Sexualization and Sexualization Distress (r = .54, p < .01); Low Achieving/Undesirable and Low Achieving/Undesirable Distress (r = .57, p < .01). Previous studies also revealed a similar result, with all six frequent subscales correlated to their corresponding distress subscales (Andrade Jr, 2013;

Torres-Harding, Turner, 2014), indicating that as the frequency of racial microaggressions increases, so does the distress associated with these experiences.

Acculturative Stress Scale for International Students (ASSIS; Sandhu &

Asrabadi, 1994). ASSIS is a self-reporting measure which identifies acculturative stress, known as psychological adjustment problems, of international students. It was originally developed for international students in the U.S., including students from Asia, Latin America, Middle East, Europe, and Africa. It was also used for immigrants and sojourners living in Asian countries, such as Taiwan and Korea (Huang, 2008; Kim & Kim, 2011;

Lee, Chae, Wilbur, Miller, Lee, & Jin, 2014). The seven subscales included in the ASSIS are: Perceived Discrimination, Homesickness, Perceived Hate, Fear, Stress Due to Change/Culture Shock, Guilt, and Miscellaneous.

Each item is rated on a 5-point Likert scale ranging from 1 (strongly disagree), to 5 (strongly agree). A total score, ranging from 36 to 180, is the sum of all seven factors, with higher scores indicative of greater acculturative stress perceived by the participants.

A sample item is “I feel intimidated to participate in social activities.”

Previous studies revealed that coefficient alphas for ASSIS ranged from .92 to .94 among international students (Constantine et al., 2004; Wei et al, 2007; Wei, Tsai, Chao,

Du, & Lin, 2012). In this study, an overall Cronbach’s alpha of .95 was obtained.

Moreover, each subscale was also examined: Perceived Discrimination (α = .88);

Homesickness (α = .65); Perceived Hate (α = .77); Fear (α = .67); Stress Due to Change/Culture Shock (α = .69); Guilt (α = .77); and Miscellaneous (α = .79). The result indicated that all seven subscales have acceptable or good reliability.

In addition, the construct validity of the ASSIS was evidenced by applying principal components factor analysis (Sandhu & Asrabadi, 1994). The result shows that the factor loading for each items is >.6, which is higher than the .55 threshold for a good factor (Comrey & Lee 1992) and so indicates acceptable validity. Furthermore, the validity of REMS was also supported by a negative association with social connectedness among international students (Yeh & Inose, 2003) and a positive association with depressive symptoms among international students (Constantine et al., 2004) and Chinese international students specifically (Wei et al., 2007).

The Center for Epidemiological Studies Depression Scale–Revised (CESD–R;

Eaton et al, 2004; Appendix E). The original CESD (Radloff, 1977) was developed at the Center for Epidemiological Studies, a division of the National Institute of Mental Health in the U.S. Because the CES-D was created prior to the third revision of the American Psychiatric Association’s (1980) Diagnostic and Statistical Manual of Mental Disorders (DSM-III), it does not include symptoms in the area of anhedonia, psychomotor retardation/agitation, or suicidal ideation. Additionally, eight items from the original CES-D no longer relates to the current definition of major depression (Eaton et al, 2004;

p. 365). Therefore, the updated version of CES-D, CESD-R, was chosen to measure depressive symptoms in this study. The language use of CESD-R has been updated, items that no longer reflect the fourth edition of DSM (DSM-IV) definition of depression were removed, and items assessing anhedonia, psychomotor retardation/agitation, and

suicidality were added (Eaton et al., 2004).

Like its predecessor, the CESD-R is a 20-item self-report scale that assesses frequency of depressive symptoms in the general population. As part of the revision, each item of CESD-R is rated on a 5-point Likert scale ranging from 0 (Not at all or less than one day per week) to 4 (Nearly every day for 2 weeks). A sample item is “I wished I were dead.”

Total scores, ranging from 0 to 80, is calculated by summing the responses to all items.

Higher scores indicate greater levels of depressive symptomology. Scores of 16 or higher imply a potential risk for clinically significant depression. However, research has indicated that the traditional CES-D cut-score, which also applies to the CESD-R (Eaton et al., 2004), lacks specificity (Santor, Zuroff, Ramsay, Cervantes, & Palacios, 1995).

Relatively, the revision is slightly more conservative in its estimates of depression (Eaton et al., 2004).

The psychometric soundness of the original CES-D scale has been confirmed in previous studies (e.g., Constantine et al., 2004; Radloff, 1977; Wei et al., 2007; Wei et al., 2008) and the updated CESD-R has been reported to have similar characteristics. For example, Eaton et al. (2004) reported Cronbach’s alphas of .96 with a combined inpatient/outpatient population, .93 with a telephone-based survey to rural populations, and .92 with a web-based survey. In addition, the CESD and CESD-R scores obtained from Eaton et al.’s study (2004) were highly correlated, with Pearson Correlation coefficient ranging from .88 to .93, indicating that the relatively good validity of the original CES-D was retained by the revision. In a validation study of CESD-R, Van Dam and Earleywine (2011) also found evidence of high internal consistency for both a large community (α = 0.923) and a smaller student (α = 0.928) sample, with a confirmatory factor analysis supporting a unidimensional factor structure. In this study, the internal consistency was examined using the Cronbach’s alpha, and the result (α = 0.94) indicated

high reliability.

According to Van Dam and Earleywine’s study (2011), support for the convergent validity of CESD-R demonstrated by a positive association with other instruments of emotional distress, such as the State-Trait Inventory for Cognitive and Somatic Anxiety (STICSA; r = .74), the Schizotypal Personality Questionnaire–Brief (SPQ-B; r = .44), and the Negative Affect subscale of the Positive and Negative Affect Scale (PANAS-NA;

r = .58). In addition, divergent validity of CESD-R was demonstrated by a negative

correlation with the Positive Affect subscale of the PANAS (PANAS-PA; r =−0.26).

Therefore, it is indicated that the CESD-R is an accurate and valid measure of depression (Van Dam and Earleywine, 2011).

Demographic questionnaire (Appendix F). A brief questionnaire was used to obtain relevant demographic information including nationality, race/ethnicity, gender, age, native language, relationship status, length of residence, residence (which city), education level, program of study, and self-reported Mandarin proficiency.

Self-reported Mandarin proficiency was measured by the combined score from the following three questions, which were rated on a 5-point Likert scale:

(1) What is your current level of fluency in Mandarin?

(2) How comfortable do you feel communicating in Mandarin? and (3) How often do you communicate in Mandarin?

This method of assessing language proficiency has been documented by previous research (Constantine el al, 2004; Yeh & Inose, 2003). It was revised by changing the assessed language from English to Mandarin for an international student research in Taiwan (Huang, 2008). In this study, Cronbach’s alpha was assessed for the 3-items as 0.94.

The last question of this survey is an open-ended question which allowed

participants an opportunity to state their feedback regarding the questionnaire. Open-ended questions give research participants a chance to respond in their own words and in their own way, and to add anything that may not have been addressed within the questionnaire structure (Barker, Pistrang, & Elliot, 2002).

According to Barker et al, (2002, p. 98), three advantages of open-ended questions include (1) enable the researcher to study complex experiences, (2) respondents are able to qualify or explain their answers, and also have the opportunity to express ambivalent or contradictory feelings, and (3) respondents are free to answer as they wish, using their own spontaneous language.