An Overview of Asian-American Quality-of-Life Predictors
Introduction
As Asian-Americans are considered the fastest growing minority group in the U.S, and emerge as the largest group of new immigration (Cohn, 2015), understanding the progress, setbacks, and their current status might better explain trends within the Asian-American demographic and predict future immigration trends in the U.S. This is especially pertinent as they are one of the most understudied minority groups (Ðoàn et al., 2019). Therefore, we would like our research to answer the question: What is the role of age, income, mental health, length of residency, and English skill in predicting our response variable of interest: Quality of Life (QoL). Our null hypothesis is that these predictors are equal to zero. Our alternative hypothesis suggests that at least one of these four predictors is not equal to zero.
Literature on this topic provides a diverse list of factors that influence Asian-American Quality of Life (AAQoL). A Pew Research Center study found that the majority of Asian immigrants in the U.S. (77%) believe that their standard of living is higher than their parents’, but under half (48%) think their children’s standard of living will be better than theirs (Tian et al., 2024). For U.S.-born Asian adults, these proportions drop to 60% and 29%, respectively. (Tian et al., 2024). This illustrates the importance of birthplace in influencing quality of life. In Hawaii, studies have found that multi-ethnic Asian-Americans reported higher satisfaction with both mental and physical health compared to their mono-ethnic counterparts in some way (Zhang, 2011). One possible explanation for such a discrepancy is that multi-ethnic individuals have to navigate complex social networks, which bring more social benefits and increased fulfillment. Meanwhile, a different study has found that young Asian-Americans are isolated from their families due to their unmarried status (Jang et al., 2021). Additionally, middle-aged Asian-Americans from 40 to 59 years old have the highest odds of being socially isolated due to their limited English proficiency (Jang et al., 2021). This trend is worth noting as social isolation has been considered a health risk factor comparable to smoking 15 cigarettes a day (Holt-Lunstad et al., 2015). Moreover, limited English proficiency also plays a role in Asian-Americans’ lower participation in health services (Jang & Kim, 2018), which could potentially delay and complicate underlying health conditions, especially with avoidance of treatment and regular check-ups. These references complicate an already-complex argument and will be crucial in providing a more objective and all-encompassing assessment of quality of life for our interpretation.
.
Material & Method
We utilize a dataset from “The Final Report on the Asian American Quality of Life”, located on the data.austintexas.gov website, the official City of Austin data portal. The report was submitted in October 2016 under the lead of Dr. Yuri Jang from the University of Texas at Austin School of Social Work. This data derives from large-scale paper surveys distributed to Asian-American residents of Austin, Texas, through 76 survey sessions across 891 survey sites selected based on the Austin Asian Community Resource Database from August to December, 2015. They include city public centers, educational facilities, medical establishments, religious institutions, local businesses (such as restaurants and grocery stores), and other social services providers. Notably, Asian-Americans are defined in the report as “individuals having origins in any of the original peoples of the Far East, Southeast Asia, or the Indian subcontinent” aged 18 and above. It took about 20 min for each participant to complete the questionnaire, and respondents were each paid U.S. $10 for their participation. The dataset included 2609 unique observations and 231 variables covering many diverse topics. Some are picked from existing national and state surveys, including information on demographics, immigration and acculturation, health, emotional wellbeing, special interests, social and community resources, and general living conditions in Austin. They were collected in Texas as this state displays the highest U.S. growth rate for Asian-American populations, with a 72.4% increase from 2000 to 2010 (City of Austin, Texas, 2018). Austin also currently houses an estimated 110,000 to 115,000 Asian residents, second only to Houston, Texas (City of Austin, Texas, 2018). Our dataset highlights an ever-increasing population, and while our findings may not apply to other states with slower growth rates, they will prove pertinent in recognizing patterns, as the proportion of Asian-Americans in the United States is projected to increase from 5.6% to 10% by 2050 (City of Austin, Texas, 2018).
We are dealing with a relatively robust dataset that has been well-constructed. As a result, the data was mess-free, and we did not need to clean anything. Our only problem was the presence of NAs in our observed variables. Due to our large sample size of 2565 observations in total after removing 267 observations that contain NAs from our variables of interest, we decided it was appropriate to simply remove them, as we have a sufficient number of datapoints left. Our summary statistics can be seen in the tables below. Table 1 focuses on categorical variables by showing the proportion of the levels per variable. For example, household income in USD annually has been divided into 2 groups: under $70,000 and above that, as 42% of the sample have income of over $70,000. Mental health is also reported with proportions, including 5 groupings from ‘Poor’ to ‘Excellent’. On the other hand, our numerical summary statistic is calculated in Table 2 with the mean and standard deviation, plus the number of missing values (NAs) we removed from the dataset per variable (column).
Since quality of life is a numeric response variable, and there are multiple predictors presented, we believe it is reasonable to use the Multiple Linear Regression (MLR) model. Initially, we composed a full model with 12 variables as shown in Table 1, and built an MLR model utilizing all of them. Although this model explained the most variability in the dataset (Adjusted R-squared = 0.285), there were concerns about multicollinearity. To achieve a simpler, more straightforward model in mind, we first selected the best suited three subset that do not contain confounding variables based on significance (by a two-sided t-test) and relevancy (regarding the context of the survey). Then, limit them to only one promising subset including ‘Age’, ‘Income’, ‘Present Mental Health’, ‘Duration of Residency’, and ‘English Difficulties’. However, we suspected that ‘Income’ might not be as significant in the subset, so a nested F-test was performed for a model with ‘Income’ and one without. Consequently, our test results indicated that the model with ‘Income’ is statistically significant (p < 0.001). Therefore, we decided to examine quality of life with age, income, duration of residency, present mental health, and English speaking difficulties as the final predictors. In our initial Exploratory Data Analysis, the QoL score is skewed left, but since we wanted to make it skewed right and take a log transformation, we decided to transform QoL score and create a new variable called \(log_{quality}\) which can be calculated using this following formula: \(log_{quality} = log(11-QoL)\). After finding the estimated coefficients from the MLR model, we applied the formula \(1/(e^{-estimate})\) to retransform the QoL scores.
Results
| Variables | Missing values | N = 2,5651 |
|---|---|---|
| Ethnicity | 0 | |
| Asian Indian | 568 (22%) | |
| Chinese | 630 (25%) | |
| Filipino | 262 (10%) | |
| Korean | 467 (18%) | |
| Other | 143 (5.6%) | |
| Vietnamese | 495 (19%) | |
| Marital Status | 13 | |
| Living with a partner | 102 (4.0%) | |
| Married | 1,704 (67%) | |
| Other | 29 (1.1%) | |
| Single | 717 (28%) | |
| Income (US dollars) | 187 | |
| under70k | 1,390 (58%) | |
| over70k | 988 (42%) | |
| Present Mental Health | 9 | |
| Excellent | 630 (25%) | |
| Fair | 190 (7.4%) | |
| Good | 720 (28%) | |
| Poor | 28 (1.1%) | |
| Very Good | 988 (39%) | |
| English Difficulties | 31 | |
| Much | 541 (21%) | |
| Not at all | 765 (30%) | |
| Not much | 723 (29%) | |
| Very much | 505 (20%) | |
| 1 n (%) | ||
| Variables | Missing values | N = 2,5651 |
|---|---|---|
| Quality of life (Scale of 1-10) | 0 | 7.67 ± 1.63 |
| Age (years) | 4 | 43 ± 17 |
| Household size (persons) | 11 | 3.29 ± 1.47 |
| Duration of residency (years) | 36 | 16 ± 13 |
| Education completed (years) | 31 | 15.11 ± 2.41 |
| 1 Mean ± SD | ||
[1] “kableExtra” “knitr_kable”
| Coefficients | p-value | Lower Bound | Upper Bound | |
|---|---|---|---|---|
| Age | 0.998** | 0.007 | 0.997 | 0.999 |
| Income over $70,000 | 1.189*** | 0 | 1.143 | 1.236 |
| Mental health (Fair) | 0.516*** | 0 | 0.474 | 0.561 |
| Mental health (Good) | 0.62*** | 0 | 0.587 | 0.654 |
| Mental health (Poor) | 0.519*** | 0 | 0.43 | 0.626 |
| Mental health (Very Good) | 0.739*** | 0 | 0.704 | 0.776 |
| Duration of residency | 1.006*** | 0 | 1.004 | 1.008 |
| English difficulties (Not at all) | 1.167*** | 0 | 1.101 | 1.236 |
| English difficulties (Not much) | 1.011 | 0.696 | 0.957 | 1.068 |
| English difficulties (Very much) | 1.05 | 0.115 | 0.988 | 1.116 |
| Num.Obs. | 2258 | |||
| R2 Adj. | 0.265 |
In general, our data shows many significant coefficients with each of the explanatory variables in conjunction with each other (p<0.05). Holding all other variables constant, for each additional year added in age, AAQoL decreases 0.998 (0.997;0.999) times on average, while for each increment of a year of residency, AAQoL increases 1.006 (1.004;1.008) times. Our categorical variables show a similar picture with an overall positive correlation with QoL. Particularly with other variables accounted for, Asian-American whose annual household income from $70,000 and above have 1.189(1.143;1.236) times higher QoL compared to the lowest income group of $0 to $10000 on average. It is also worth noting that Asian-Americans with a “Very Good” mental health state have a QoL that is 0.739 (0.704;0.776) times lower than the people with “Excellent” mental health. Additionally, the QoL of Asian-Americans who considered their English speaking difficulty as “Not at all” was 1.167 (1.101;1.236) times larger than those who reported their difficulty as “Much”, in respect to other variables.
Discussion
The above interpretation of our coefficients is within our expectation based on established trends of higher income, better mental health, lengthening of stay and more cultural assimilation correlate to a higher degree of life satisfaction for everyone, including Asian-Americans. While aging demonstrated a negative correlation to AAQoL, the difference is not really large, and thus less meaningful when compared to the rest of the explanatory variables. Since age is in opposition with duration of residency, which means that the longer people live in a place, they have higher life satisfaction as they get used to the new location but might have lower life satisfaction due to aging, making these two variable coefficients cancel out. Our final model does not contain any confounding variables as we have eliminated them beforehand.
Contextually, we do not have a chance to compare variables that are related to our literature like ethnicity, marital status or household size in our final model. Thus, the relationship of the predictors we have might be strictly interpreted for correlation and not for causal relationship. While all states of mental health are related to QoL in our model, it is interesting that the supposedly worst mental health state of “Poor” has a smaller difference to “Excellent” mental health. These small discrepancy could be accounted for by the noteworthy overlapping between “Poor” state of mental health’s QoL scores to “Fair” and “Good” mental health states, with a small QoL score overlap to “Excellent” and “Very Good” in the Present Mental Health boxplot in Figure 1. The AAQoL report has noted that there is a possible lack of awareness of mental health problems among the populations as more than one third survey participants believed that depression is a sign of weakness, and almost one half have misconception of antidepressant, they think that it is addictive. The internalization of the model minority myth that all Asian-American is wealthy, highly educated and problem free (Yi et al., 2016) could exacerbate the predisposing mental health stigma of Asian-Americans which stem from Asian cultural emphasis on Asian cultural values like emotional restrain, shame avoidance, and saving face (Shea & Yeh, 2008). These results are making the case for a nuanced approach of providing Asian-Americans, especially immigrants the anticipated help they need.
However, one limitation from our model is variability with the income groups, where the over $70,000 group is disproportionately represented of almost half the sample. Self-assessed mental health might be underestimated due to stigma while English difficulties could be overestimated from actual capability. While Austin serves as an excellent sample of the Asian population’s life experience, it might overrepresent the new immigration population with much higher income or average income of the American middle-class. Future study could further survey other cities in different states, such as California that also experience an increase of Asian immigration and also expand our research to different racial groups and ethnicities. The same AAQoL survey could also be re-introduced as it has been almost 10 years since this report and new insight could be made between the first and the second future report.
Contextually, we do not have a chance to compare variables that are related to our literature like ethnicity, marital status or household size in our final model. Thus, the relationship of the predictors we have might be strictly interpreted for correlation and not for causal relationship. While all states of mental health are related to QoL in our model, it is interesting that the supposedly worst mental health state of “Poor” has a smaller difference to “Excellent” mental health. These small discrepancy could be accounted for by the noteworthy overlapping between “Poor” state of mental health’s QoL scores to “Fair” and “Good” mental health states, with a small QoL score overlap to “Excellent” and “Very Good” in the Present Mental Health boxplot in Figure 1. The AAQoL report has noted that there is a possible lack of awareness of mental health problems among the populations as more than one third survey participants believed that depression is a sign of weakness, and almost one half have misconception of antidepressant, they think that it is addictive. The internalization of the model minority myth that all Asian-American is wealthy, highly educated and problem free (Yi et al., 2016) could exacerbate the predisposing mental health stigma of Asian-Americans which stem from Asian cultural emphasis on Asian cultural values like emotional restrain, shame avoidance, and saving face (Shea & Yeh, 2008). These results are making the case for a nuanced approach of providing Asian-Americans, especially immigrants the anticipated help they need.
However, one limitation from our model is variability with the income groups, where the over $70,000 group is disproportionately represented of almost half the sample. Self-assessed mental health might be underestimated due to stigma while English difficulties could be overestimated from actual capability. While Austin serves as an excellent sample of the Asian population’s life experience, it might overrepresent the new immigration population with much higher income or average income of the American middle-class. Future study could further survey other cities in different states, such as California that also experience an increase of Asian immigration and also expand our research to different racial groups and ethnicities. The same AAQoL survey could also be re-introduced as it has been almost 10 years since this report and new insight could be made between the first and the second future report.
Bibliography
City of Austin, Texas. (2018). Final Report of the Asian American Quality of Life (AAQoL). In Austintexas.gov. https://data.austintexas.gov/dataset/Final-Report-of-the-Asian-American-Quality-of-Life/hc5t-p62z/about_data
Cohn, D. (2015, October 5). Future immigration will change the face of America by 2065. Pew Research Center. https://www.pewresearch.org/short-reads/2015/10/05/future-immigration-will-change-the-face-of-america-by-2065/
Ðoàn, L. N., Takata, Y., Sakuma, K.-L. K., & Irvin, V. L. (2019). Trends in Clinical Research Including Asian American, Native Hawaiian, and Pacific Islander Participants Funded by the US National Institutes of Health, 1992 to 2018. JAMA Network Open, 2(7), e197432–e197432. https://doi.org/10.1001/jamanetworkopen.2019.7432
Holt-Lunstad, J., Smith, T. B., Baker, M., Harris, T., & Stephenson, D. (2015). Loneliness and Social Isolation as Risk Factors for Mortality. Perspectives on Psychological Science, 10(2), 227–237. https://doi.org/10.1177/1745691614568352
Jang, Y., & Kim, M. T. (2018). Limited English Proficiency and Health Service Use in Asian Americans. Journal of Immigrant and Minority Health, 21(2), 264–270. https://doi.org/10.1007/s10903-018-0763-0
Jang, Y., Park, J., Choi, E. Y., Cho, Y. J., Park, N. S., & Chiriboga, D. A. (2021). Social isolation in Asian Americans: risks associated with socio-demographic, health, and immigration factors. Ethnicity and Health, 27(6), 1428–1441. https://doi.org/10.1080/13557858.2021.1881765
Peterson Foundation. (2025, March 12). Income and Wealth in the United States: An Overview of Recent Data. Peterson Foundation. https://www.pgpf.org/article/income-and-wealth-in-the-united-states-an-overview-of-recent-data/
Shea, M., & Yeh, C. (2008). Asian American Students’ Cultural Values, Stigma, and Relational Self-construal: Correlates of Attitudes Toward Professional Help Seeking. Journal of Mental Health Counseling, 30(2), 157–172. https://doi.org/10.17744/mehc.30.2.g662g5l2r1352198
Tian, Z., Im, C., Mukherjee, S., & Budiman, A. (2024, October 9). 2. Asian American immigrants’ views of quality of life in the U.S. Pew Research Center. https://www.pewresearch.org/race-and-ethnicity/2024/10/09/asian-american-immigrants-views-of-quality-of-life-in-the-u-s/
Yi, S. S., Kwon, S. C., Sacks, R., & Trinh-Shevrin, C. (2016). Commentary: Persistence and Health-Related Consequences of the Model Minority Stereotype for Asian Americans. Ethnicity & Disease, 26(1), 133. https://doi.org/10.18865/ed.26.1.133
Zhang, W. (2011). Health Disparities and Relational Well-Being between Multi- and Mono-Ethnic Asian Americans. Social Indicators Research, 110(2), 735–750. https://doi.org/10.1007/s11205-011-9956-9