The factor structure of a short acculturation scale for Filipino Americans in an adult U.S.‐born sample

Abstract The influx of non‐European immigrants since 1965 ushered the development and use of acculturation measures in immigrant health studies. A Short Acculturation Scale for Filipino Americans (ASASFA) represents a validated, unidirectional ethnic‐specific measure used with first‐generation FAs. ASASFA's psychometric properties with adult U.S.‐born children—the second generation—remain unexplored. This study determined (a) the factor structure of ASASFA with adult U.S.‐born FAs and (b) the predictors of their acculturation scores. A secondary analysis was conducted on ASASFA data from a mental health survey of 116 U.S.‐born FAs. Exploratory factor and parallel analyses showed a two‐factor solution: language use and preference (Factor 1) and ethnic social relations (Factor 2). Ordinary least squares regression indicated gender and ethnic self‐identification predict Factor 1 scores; self‐identification solely predicts Factor 2 scores. Results demonstrate ASASFA's validity and parsimony, supporting its use in FA health studies when lengthy bidirectional acculturation measures become impractical.

20 million U.S.-born children of the new immigrants have come of age (Pew Research Center, 2013). As U.S. citizens, their acculturation patterns and adaptation lay the foundation for the enduring character of their ethnic communities (Portes & Rumbaut, 2005).
The arrival of the new immigrants spawned the development of a variety of acculturation measures (Balls Organista, Marin, & Chung, 2010), with most of these measures focusing on Latin American acculturation (Gamst, Liang, & Der-Karabetian, 2011). Asians, a heterogeneous racial group, now comprise the fastest growing immigrant population in the United States (Malik, 2015, May 21); nevertheless, a limited number of ethnic-specific acculturation measures exist (Gamst, Liang, & Der-Karbetian, 2011). Moreover, research on the acculturation of their adult U.S.-born children is scarce due to the paucity of psychometrically validated ethnic-specific measures.
Filipinos presently represent the fourth largest immigrant group by country of origin behind Mexico, China, and India (McNamara & Batalova, 2015). Post-1965 Filipino migration features a greater proportion of highly educated professionals such as physicians, nurses, and engineers. The U.S. Census Bureau (2015)  On the other hand, the children of Filipino immigrants who are born in the United States live in two contrasting cultures: the American culture that emphasizes individualism and autonomy, and the Philippine culture that underscores collectivism-the primacy of family and harmonious relations (Nadal, 2011). They experience the American mainstream culture in their communities, schools, and work places, and through the media, and the Philippine culture at home with parents and extended families as well as during the observance of and participation in Filipino cultural events. Therefore, U.S.-born FAs undergo acculturation in two disparate cultures. They navigate and adapt variously to both cultures through the process of acculturation, especially because these two cultures manifest contrasting orientations. Given the relationship between acculturation and health outcomes, and the variability of acculturation across individuals and generations, an appropriate measure of acculturation becomes an invaluable tool when designing culturally-tailored health care interventions for U.S.-born FAs.
A Short Acculturation Scale for Filipino Americans (ASASFA; dela Cruz, Padilla, & Agustin, 2000;dela Cruz, Padilla, & Butts, 1998) represents the only ethnic-specific acculturation measure and hence the most used for FAs.
Since its development and original validation, the studies using ASASFA involved first-generation (Philippine-born) Filipinos (dela Cruz & Galang, 2008;dela Cruz et al., 2013;Ea, 2007;Kataoka-Yahiro, 2010;McAdam, Stotts, Padilla, & Puntillo, 2005;Reyes & Cohen, 2013;Serafica et al., 2013). Very little research on its use with U.S.-born FAs exists. The purpose of this study therefore was to fill this gap in the literature. Specifically, it sought to determine (a) the factor structure of the scale with adult U.S.-born FAs and (b) the predictors of their acculturation scores. ASASFA underwent its initial validation in 1998(dela Cruz et al., 1998. The measure traces its development and adaptation from the Short Acculturation Scale for Hispanics (SASH; Marin, Sabogal, Marin, Otero-Sabogal, & Perez-Stable, 1987). ASASFA has two validated language versions-in English and in Filipino or Tagalog, the Philippine national language. These language versions underwent steps to ensure their content, technical, experiential, semantic, and conceptual equivalence (dela Cruz et al., 2000). Consistent with SASH (Marin et al.,1987), the measures exclude sociodemographics in the scales to maintain their use as correlates of acculturation.

ASASFA
In the initial study, principal components analysis revealed three dimensions being measured by the scale in each language version: language use and preference at home, work, and with friends; language use and preference in media-TV, film, and radio; and preference for ethnic social relations. Each language version obtained an overall Cronbach's alpha coefficient of .85, indicating high internal consistency among the items (dela Cruz et al., 1998(dela Cruz et al., , 2000. Studies that have used this measure have shown appropriate coefficient alphas as evidence of its reliability and have been consistent with the theoretical literature (Ayers, Atkins, & Lee, 2010).
The original validation study showed that ethnic identity, food preference, educational level, length of U.S. residence, family income, and age on arrival to the United States were associated with the acculturation level of firstgeneration FAs. Stepwise multiple regression indicated that ethnic identification serves as the primary factor associated with their acculturation level (dela Cruz et al., 1998(dela Cruz et al., , 2000. ASASFA represents a unidirectional but multidimensional model of acculturation (Flannery, Riese, & Yu, 2001). This model posits that acculturation occurs in a continuum-starting with total immersion in the heritage culture, progressing to biculturalism, and finally assimilation into the host culture (Abe-Kim, Okazaki, & Goto, 2001;Zhang & Tsai, 2014).
Another acculturation model, the bidirectional model, contrasts with the unidirectional model. The bidirectional model proposes that the simultaneous acquiring or adhering to a new culture and maintaining the heritage culture represent two independent processes (Abe- Kim et al., 2001;Zhang & Tsai, 2014). Accordingly, U.S.-born FAs concurrently learn and selectively acquire and adapt the customs, behaviors, and attitudes of their heritage culture along with the American mainstream culture.
Despite the increasing use of bidirectional measures, a unidirectional measure such as ASASFA provides an economical and parsimonious tool (Flannery, Reise, & Yu, 2001) for health community surveys that include acculturation as a variable. Furthermore, different models may apply to different subgroups (Zhang & Tsai, 2014). In a previous study, the unidimensional model applied more to Chinese Americans who came to the United States after adolescence, while the bidirectional model applied more to Chinese Americans who immigrated prior to adolescence (Tsai, Ying, & Lee, 2000;Zhang & Tsai, 2014).

Research design and sample
A secondary analysis was conducted on the acculturation data collected from 116 U.S.-born FAs, as part of a survey that investigated the factors that relate to and predict the mental health of Philippine-born and U.S.-born FAs (Vindua, 2010). The sample was recruited from local organizations in the Greater Los Angeles Area. Study participants met at a designated location, either one on one or in small groups of 5-10 individuals, where they completed the survey questionnaires. The study received the approval of the university's institutional review board.

Measures
This study used two measures from the original survey as follows.

ASASFA
ASASFA assessed language use and preference at work and home and with friends; and language use and preference for media. The measure uses a scale of 1 = only Philippine languages, 2 = more Philippine languages than English, 3 = both equally, 4 = more English than Philippine languages, and 5 = only English. To assess preference for ethnic social relations, the measure uses a scale of 1 = all Filipinos, 2 = more Filipinos than Americans, 3 = about half and half, 4 = more Americans than Filipinos, and 5 = All Americans (dela Cruz et al., 1998(dela Cruz et al., , 2000. Based on the value assigned to the response, the total acculturation score ranged from 12 to 60; the mean score for the total scale and subscales ranged from 1 to 5. Lower scores reflect lower acculturation levels toward the American culture and middle scores signal biculturalism, while higher scores indicate higher acculturation level toward the American culture. This study on U.S.-born FAs focuses only on the English version of the scale.

Data analysis
Descriptive statistics were used to examine the sociodemographic characteristics of the study participants and their responses to ASASFA, using SPSS (version 22). In addition, we computed the Kaiser-Meyer-Olkin measure of sampling adequacy and Bartlett's test of sphericity to determine the factorability of the data (Young & Pearce, 2013). We employed exploratory factor analysis (EFA) to explore the scale's factor structure and to generate the loading plot, using JMP (pro version 13).
To select the number of factors in EFA, we initially viewed the inflection point in the scree plot; in addition, we adopted the loading plot approach and parallel analysis for verification. A loading plot is a graphical technique to depict the direction, proximity, and grouping pattern of the variables (the scale items). The relationship between variables and factors can be expressed as weighted linear equations, which can be re-expressed as vectors in the loading plot (Jacoby, 1998;SAS Institute, 2016;Yan & Kang, 2003).
As the name implies, EFA is exploratory in nature, and thus it is error-prone (Osborne, 2014). As a remedy, many authors recommend using parallel analysis to verify the factor structure (Buja & Eyuboglu, 1992;Glorfeld, 1995;Ledesma & Valero-Mora, 2007). Parallel analysis is a form of resampling, such that the existing sample is regarded as a pseudo-population (Yu, Osborn-Popp, DiGangi, & Jannasch-Pennell, 2007). The algorithm generates a set of random data correlation matrices by resampling from the sample, and then the average eigenvalues and the 95th percentile eigenvalues are computed. The underlying logic is: the extracted factors must substantively outperform the random factors by chance alone. We used SAS (version 9.4) as the programming environment for running the parallel analysis macros developed by O'Connor (2000). In this study 1,000 re-samples were used in parallel analysis.
Additionally, the relationships between acculturation and certain sociodemographic variables were examined after the factor structure was ascertained. Ordinary least squares (OLS) regression was employed to examine which of these demographic variables-age, gender, educational level, occupational status, household income, religious preference, food preference, and self-identification-are predictors of the acculturation subscales. Marital status was excluded because most participants were single. Likewise, occupation was omitted because 44.8% of the sample left this item unanswered; these study participants were looking for work or studying full time.
To avoid estimating too many parameters and yielding an unstable model due to lack of degrees of freedom, we combined the following demographic variables.

Preliminary analysis
While computing the descriptive statistics of the responses to the ASASFA, we found that the respondents uniformly answered "All English" to item AC7, which asked, "In what language(s) are the radio programs you usually listen to?" Because of this lack of variability, we deleted this item in all subsequent statistical computations.
The sample's mean total acculturation score registered at 3.91 (SD = .44). The mean scores for the items on lan-

EFA
EFA suggested a two-factor solution, based on the inflection point of the scree plot (Figure 1). Note. SD = standard deviation; SSI = Supplemental Security Income; WIC = Special Supplemental Program for Women, Infants, and Children. Note. ASASFA = A Short Acculturation Scale for Filipino Americans. Boldface indicates highest factor loadings.

F I G U R E 2 Loading plot of vectors of the variables for Factor 1 and Factor 2
The loading plot (Figure 2) provides further evidence of ASASFA's factor structure. The vectors that represent the variables form two clusters. The proximity of all language use and preference (AC) items is close as are all ethnic social relations (SR) items. Although the two sets are not perfectly orthogonal (90 degrees), they are far apart from each other, and thus the two constructs should be considered distinct.
Parallel analysis provides additional evidence. Table 3 shows the original eigenvalues, the means of resampled eigenvalues, and the 95 th percentile of resampled eigenvalues. A two-factor solution based on the original eigenvalues can substantively outperform the random results.

Language use and preference scores
OLS regression indicates that gender (p = 0.0016) and self-identification (p < .0001) were crucial predictors of language scores (Table 4). The diamond plots show that males have a significantly higher score than females after two female outliers were removed ( Figure 3) and Figure 4 compares the self-identification groups in relation to language use and preference scores. All the diamond plots do not overlap, indicating that all three self-identification groups significantly differ from each other. The CI information on gender can be found at Table 5 and on self-identification at

Ethnic social relations scores
Another OLS regression model shows that only self-identification (Table 7) was a crucial predictor of social ethnic relation scores (p = .0047). After eight outliers were removed, the diamond plot, as displayed in Figure 5, indicated that there was no significant difference between "almost 50/50 Filipino and American" and "very Filipino and more Filipino than American," but the social ethnic relation score of "very American and more American than Filipino" is significantly higher than the other two groups. The CI information is shown in Table 8.

DISCUSSION
This study examined the factor structure of ASASFA in adult U.S.-born FAs as well as the sociodemographic predictors of their acculturation scores. EFA revealed that the scale measured two factors-language use and preference and ethnic social relations, with the deletion of one item related to the language of radio programs listened to. Parallel F I G U R E 3 Diamond plots comparing females and males in relation to language use and preference scores F I G U R E 4 Diamond plots comparing self-identification groups in relation to language use and preference scores  Confidence intervals for self-identification groups in relation to language use and preference scores

Factor structure of ASASFA
The factor structure of ASASFA in adult U.S.-born FAs dovetails with the two most common dimensions measured by major ethnic-specific acculturation measures-language use and preference, and preference in social ethnic relations (Celenk & Van de Vijer, 2011;Matsudaira, 2006;Zane & Mak, 2003). Being born of parents speaking a language at home that differs from the mainstream American language used outside of the home, U.S.-born FAs reveal a shift to English language use and preference as a key aspect of their acculturation to American society. At the same time, they disclosed TA B L E 8 Confidence intervals for self-identification groups in relation to social ethnic relations scores With the deletion of the item related to the language of radio programs listened to, ASASFA for U.S.-born FAs now comprises 11 items; this short scale falls within the mean of 11 items of most acculturation measures (Celenk & Van de Vijer, 2011). The total scale reliability (.82) and dimensional reliabilities of language use and preference (.86) and ethnic social relations (.81) all meet the minimum standard of .80 (Nunnally & Bernstein, 1994). The factor structure and reliabilities of ASASFA indicate the robustness of its psychometric properties, suggesting its utility as a parsimonious acculturation measure for U.S.-born FAs.

Sociodemographic predictors of acculturation scores 4.2.1 Language use and preference
Using OLS regression, our study revealed that U.S.-born FA males indicated a preference for and use of the English language compared to U.S.-born FA females. This finding can be attributed to the practice of the traditional Filipino immigrant family: to keep a more sheltered life for daughters than sons. Consequently, daughters are expected to stay more at home with members of the family to imbibe their role in maintaining family ties and cultural traditions-activities that foster the use and maintenance of the heritage language (Espiritu, 2001). Our results concur with similar findings in an earlier study that revealed U.S.-born female students were more competent in the heritage language of their immigrant parents (Portes & Hao, 1998).
In addition, data showed that all U.S.-born FAs clearly use and prefer the English language, highlighting the reality of its instrumental use in navigating day-to-day life in the United States. But the strength and intensity of this acculturative attitude and behavior varied according to their disclosed ethnic identity. The incremental increase in the English language use and preference mean scores, from those who claimed their identity as Filipinos, to 50/50 Filipinos and Americans, and to Americans, support the strength of this attitude and behavior. Likewise, the significant differences in the scores between these ethnic identification groups reinforce their language inclinations. These results align with previous studies that found a strong link between ethnic identification and language use and preference-that these aspects of acculturation entwine with each other (Phinney, Romero, Nava, & Huang, 2001;Portes & Schauffer, 1994).

Ethnic social relations
OLS regression further revealed the significant role ethnic identity plays among adult U.S.-born FAs' preference for ethnic social relations. On the whole, based on their ethnic identity, their ethnic social relations mean scores were lower along the acculturation scale continuum, compared to their language use and preference mean scores. The ethnic social relations mean scores gradually increased from those identifying themselves as Filipinos to those who identified themselves as Americans.
As in language use and preference for English, significant differences appeared in the preferences for ethnic social relations based on professed ethnic self-identity. As indicated by the diamond plots in Figure 5, those who identified themselves very Filipino and more Filipino than American as well as those who classified themselves as almost 50/50 Filipino and American preferred co-Filipino Americans in social relations more than those who categorized themselves as more American than Filipino and very American. Thus, our study results capture the within-group acculturation differences among U.S.-born FAs based on their ethnic self-identification.
However, even the higher social ethnic relations mean scores of those who claimed more American identity still maintain ethnic relations with Filipino Americans, alluding to the enduring Filipino familial and co-ethnic ties that permeate their lives, suggesting that ethnic identity depends on social context (Nagel, 1994;Schimmele & Wu, 2015).  The second limitation stems from the study's use of a cross-sectional research design, using a self-report questionnaire. Respondents could have answered the ASASFA questions in a socially desirable manner. Also, this research design precludes causality. However, the results of studies on the children of post-1965 immigrants consistently support the study's findings on their acculturation levels and the factors that predict their language use and preference and preference for ethnic social relations (Espiritu & Wolf, 2001;Portes & Rambaut, 2006). Future studies aiming to capture a more comprehensive scope of the acculturation of U.S.-born FAs must include additional measures of cultural values, beliefs, practices, and ethnic identity. Furthermore, to apprehend the dynamic process of acculturation, a longitudinal design including a qualitative approach can discover the contextual aspects shaping the process, and the sources and amount of acculturative change in U.S.-born adults (Fuligni, 2001;Ozer, 2013).

Strengths and limitations and Directions for future Research
Third, the study did not assess the acculturation level of the sample's parents. It is possible, in particular, that more acculturated immigrant parents may have encouraged their children to use the English language rather than the heritage language at home to heighten the children's academic achievement and progression.
The fourth limitation relates to the use of only one question related to self-identification, limiting the capture of the multidimensionality of the concept of ethnic identity that includes not only self-identification but also group attachment, in-group attitudes, and ethnic involvement (Schimmele & Wu, 2015). These dimensions reflect not only selfascription but also ascription by others-and they emerge as a result of intergroup relations (Schimmele & Wu, 2015).
To capture these dimensions and the changeableness of ethnic identity across contexts would require a longer psychometrically valid measure.

Conclusion
This study shows that ASAFA demonstrates acceptable reliabilities and construct validity by Cronbach's alpha, EFA, and parallel analysis. ASASFA captures the within group differences among U.S.-born FA adults in two common aspects of acculturation: language use and preference and preference in ethnic social relations. Hence, as an 11-item acculturation scale, it provides a parsimonious tool for future studies focusing on FA health that treat acculturation as a variable.
The availability of a valid and parsimonious tool supports its use when lengthy bidirectional acculturation measures become impractical in studies of immigrant health.