Page 35 - IJPS-3-1

Sizhe Liu and Wei Zhang

access perceived support from family members, and perceived support from friends.
Both variables were measured by the sum of two items: (1) how much the respondents
can open up to family or relatives/friends and talk about their worries (1–4), and (2)
how much the respondent can rely on family or relatives/friends if they need help with
serious problems (1–4).
2.3 Analysis

Descriptive statistics that summarize distributions of study variables and bivariate
analyses that summarize the zero-order associations between independent variables
and dependent variables were presented in Table 1. Following this, multivariate logistic
regressions were conducted to analyze the relationship between focal independent
variables and dependent variables, adjusting for the effects of control variables.
Results from logistic regressions were presented in Tables 2–3. For both tables, Model
1 examined the association between immigration-related factors (age at immigration
and English proficiency) and mental health problems (12-month depressive disorder/
suicidal ideation), controlling for all the socio-demographic factors. Model 2 explored
the relationship between individual religious involvement (religious affiliation, service
attendance, and religious coping) and mental health problems. In Model 3, both
immigration-related factors and indicators of religious involvement were incorporated
to see if their effects on mental health problems remain. In the final model (Model 4),
we introduced the interaction terms between religious involvement and immigration-
related factors to examine the stress-buffering effect of religion. Only the significant
interaction terms were reported.
Considering the three stage sampling design of the NLAAS, weighted
analyses were employed in STATA. We first used command “svyset SECLUSTR
[pweight=NLAASWGT], strata (SESTRAT)” before variable management section
in STATA to specify weight and sampling design in the NLAAS. And then we
generated subsample for Asian-American immigrants. Finally we used command “svy,
subpop (Asian)” before every logistic regression model to conduct weighted logistic
regression.

3 Results

3.1 Descriptive Statistics

Among 1,641 Asian-American immigrants, approximately 3.99% reported having
major depressive disorder for the past 12 months, and 5.39% reported that they
had seriously thought about committing suicide. As shown in Table 1, 16.49% of
the respondents immigrated to the U.S. younger than or at age 12. Approximately
41.67% of the respondents indicated limited English proficiency. In terms of religious
involvement, 32.24% of the respondents attended religious services more than once
a week and 26.94% reported having sought comfort from religion frequently when
encountering difficulties. Approximately 41.94% of respondents were Christian;
21.81% of them did not have any religious affiliation, and 36.25% were with other
religious traditions. For control variables, approximately 70% of the respondents were
under 50 years old. More than half of the respondents were female (53.43%) and most
of them were married (74.37%) and employed (63.46%). For social economic status,
65.22% of Asian immigrants reported having at least some college level of education,
84.95% of them received their education outside the U.S. before age 16, and 39.28%
of them reported having an annual income being more than $75,000. More than
65% of the respondents have stayed in the U.S. for over 10 years. Nearly one third
(30.60%) of Asian-American immigrants were Chinese, and 16.30%, 19.69%, 33.41
were Vietnamese, Filipino and other Asian, respectively. The means of two scales,
perceived social support from family and perceived social support from friends, are 5.41
(SD=1.94) and 5.11 (SD=1.85), respectively.

International Journal of Population Studies 2017, Volume 3, Issue 1 29