TIME-USE AND HEALTH STATUS: AN ANALYSIS OF TIME-USE ASSIMILATION AND - - PDF document

1

TIME-USE AND HEALTH STATUS: AN ANALYSIS OF TIME-USE ASSIMILATION AND SELF-RATED HEALTH AMONG MEXICAN IMMIGRANTS IN THE U.S. Introduction Disparities in health across race/ethnicity and socioeconomic status (SES) have been documented and studied in the U.S. (Isaac, 2013). The importance of these dimensions of disparities was stated in the Healthy People 2020 program, where disparities are defined as any type of health difference that is interrelated to social, economic, and/or environmental disadvantage, recognizing the importance of social determinants of health outcomes (Office of Disease Prevention and Health Promotion, n.d.). The social determinants of health regulate access to sufficient social and economic resources and their role in health outcomes (The Secretary’s Advisory Committee on Health Promotion and Disease Prevention Objectives for 2020, 2008), resulting in what Link & Phelan (1995) identified as the fundamental cause of diseases. This perspective establishes that population in the lower SES categories have fewer resources to develop their human capital, are less competitive in the labor market, less employable, and their wages tend to be lower, which directly affects their health status (LaVeist, 2005). Some other mechanisms by which SES affects health are explained by the social causation

• theory. This approach explains the relationship between SES and health outcomes through the

differential exposure to risk factors, the attenuation of the impact of health risks, preferences for healthier lifestyles, and the availability of resources for pursuing those healthier preferences (LaVeist, 2005; Rogers, Hummer, & Krueger, 2005).

2

Failing to control for SES might result in the wrong interpretation of the relationships found between race/ethnicity and health. For instance, some health disparities have been explained by the longstanding differences in SES between racial and ethnic groups and the enduring experience

• f racial discrimination of the least advantaged group (Williams & Collins, 2013a). However,

failing to control for the correct SES dimension in the analyses of racial/ethnic disparities in health could result in the overestimation of the effect that race/ethnicity has and fostering emphasis in biological differences (Adler & Rehkopf, 2008). According to Galobardes, Shaw, Lawlor, Smith and Lynch (2006), one single indicator of SES will not capture the experiences of all the groups in a society, given that the social circumstances among different ethnic groups can be different and specific stratification for one group in a particular SES indicator could not capture the experiences of other groups. For this reason, including new dimensions of SES in this analysis will allow for a better identification of the specific mechanisms resulting in differences in health outcomes. The role of dimensions of SES on health Educational attainment is among the most commonly used indicators of SES in the study of mortality and health disparities. Most studies have concluded that higher levels of educational attainment are associated with lower mortality risk explained by the role of education in increasing earning power, access to resources, improved individual agency, developing social connections and promoting healthy behaviors and healthier life trajectories (Elo, 2009; Hummer & Lariscy, 2011). Other commonly used indicators of SES are income, wealth and poverty. The underlying principle attached to these indicators is that health is related to them by material deprivation,

3

restrictions in social participation and lack of control over one’s life (Marmot, 2002). For instance, income can be used for accessing health care, better food, housing and education, as well as protecting individuals from financial stress (Adler & Newman, 2002; Elo, 2009). All these indicators for different dimensions of SES have been found of fundamental importance for explaining racial/ethnic differences in health outcomes. An important resource that has not been explored in detail in the study of health outcomes is time. In the following section I examine the role that time allocation has for health outcomes. Time-use and health Time allocation has been studied from the economic perspective in explaining the decision- making processes in households to maximize their utility (Becker, 1965; Hamermesh & Pfann, 2005). According to Becker’s original framework on time allocation, time has to be divided between income producing activities, household reproduction, and time consuming activities (Becker, 1965). Time allocation has also been used in the gender literature to identify imbalances in workload between men and women (Anxo et al., 2011; Razavi, 2011; Sayer, 2005) and to measure trends in free time behaviors (Robinson & Godbey, 1997). However, links between time- use and health outcomes remains relatively unexplored. Research in health promoting behaviors usually highlights the importance of exercising and

• ther leisure activities in reducing obesity and improving health (Adler & Newman, 2002; Adler

& Stewart, 2010; Elo, 2009). The participation in activities promoting social integration and engagement in social institutions — such as church and clubs — has been suggested as an important connection with health improvement (Adler & Newman, 2002). For instance, lack of social interaction is related to low self-rated health and increased mortality among older adults

4

(Luo, Hawkley, Waite, & Cacioppo, 2012). Also, religious attendance has been found to provide protective effects on mortality and health (Dupre, Franzese, & Parrado, 2006; Ellison & Levin, 1998; Hummer, Rogers, Nam, & Ellison, 1999). This relationship is likely to operate through the facilitation of coping mechanisms and social support, as well as control over unhealthy behaviors (Ellison & Levin, 1998; George, Ellison, & Larson, 2002; George, Larson, Koenig, & McCullough, 2000; Rogers et al., 2005). Poverty of time and health Poverty is a dimension that is highly related to health outcomes. New perspectives on the measure of poverty suggest that indicators should go beyond the material perspective, including not having leisure activities or not participating in social activities or accounting for unfair division between work and leisure (Harvey & Mukhopadhyay, 2007; Marmot, Friel, Bell, Houweling, & Taylor, 2008). The mechanisms linking time allocation to leisure, exercise and social engagement assume that the population has time available to spend in these activities, but that is not always the

• case. The way in which individuals and households allocate their time and the constraints they face

have implications for the ability to overcome poverty (Bardasi & Wodon, 2006). Poverty of time refers to the lack of time for leisure, rest and social activities, after the time in work, housework and other activities necessary for household survival are taken into account (Bardasi & Wodon, 2006). Time poverty, as well as income poverty, constrains the access to the necessary resources needed to achieve a basic standard of living. Some individuals and households can compensate for the lack of time for household reproduction by market purchases — such as domestic workers, food, and day care— with income from paid work (Zacharias, Antronopoulos,

5

& Masterson, 2012), while some others will have monetary income below poverty thresholds, limiting their ability to purchase household services (Kalenkoski & Mahrick, 2012). Not having sufficient time available or the freedom to allocate time to the activities the individuals want to conduct has direct implications for their well-being. For instance, Spinney and Millward (2010) found that income and time poverty represent constraints to engagement in physical activities, but the effect of time poverty was higher than that of income in explaining the intensity of exercise. Also, time spent in leisure has been found to be more beneficial for women than for men, while time in paid work resulted in greater gains in health for men (Bird & Fremont, 1991), although general demands from paid work and family obligations are usually associated with increasing high risk of illness reports (Krantz, Berntsson, & Lundberg, 2005). Time poverty also has implications for the mechanisms for dealing with diseases. Access to primary care among minorities is limited by the availability of total time, waiting times and transportation to the physician’s office (Joyce & Stewart, 1999; Lara, Gamboa, Kahramanian, Morales, & Hayes Bautista, 2013; Mechanic, 2002). As a consequence, minority populations can delay or forgo timely treatment, resulting in adverse effects on their health status. Hispanic paradox and the role of poverty of time The Hispanic population in the U.S. does not seem to follow the traditional pattern of SES- health outcomes, with a better health status than their SES should predict (Williams & Collins, 2013). This apparent Hispanic epidemiological paradox has been documented and explored, with the evidence suggesting that Hispanics, predominantly of Mexican origin, have health outcomes closer to that of whites than to the outcomes for blacks (Markides & Coreil, 1986; Markides & Eschbach, 2005).

6

Recent analyses have provided some explanation for the Hispanic paradox. On the one hand, the paradox is predominantly observable among foreign-born populations, which has been linked to healthy migrant selectivity, salmon bias effect1 and low quality of information for this population (Palloni & Arias, 2004; Palloni & Morenoff, 2001). The healthy migrant effect and the salmon bias refer to the fact that most labor migrants are positively selected in terms of health status and they go back — or stop migrating — when their health status declines (Acevedo-Garcia & Bates, 2008; Palloni & Morenoff, 2001). Evidence suggest that, even when these factors account for some of the favorable outcomes among immigrant populations, they do not account entirely for the paradox (Abraído-Lanza, Dohrenwed, Ng-Mak, & Turner, 1999; Martinez, Aguayo-Tellez, & Rangel-Gonzalez, 2015). Another set of explanations are related to the protective cultural effect among immigrants. According to this perspective, cultural practices provide immigrants with some form of health protection (Lara et al., 2013). Acculturation, as defined by Gordon (1964), refers to the adoption

• f the cultural patterns of the host society such as language, diet, norms of conduct, among others.

As a consequence of this, distinctions between immigrant and native populations start to disappear and the cultural protection enjoyed by immigrant groups vanishes. By acculturating, immigrant groups move towards behaviors that are more common in the mainstream group with positive or negative consequences. In this respect, a review of literature by Lara et al. (2013) showed that, even when the evidence for health outcomes was mixed, the general trend was for negative effects

• f acculturation on health behaviors.

1 First highlighted by Pablos-Méndez (1994) and measured by Abraído-Lanza et al. (1999) salmon bias effect refers

to foreign-born Hispanics returning to their home country after temporary unemployment, retirement or serious illness, resulting in foreign deaths that are not tabulated in U.S. statistics.

7

Other concerns regarding the study of health among the Hispanic population focus on the way in which health status is being measured. In particular, the use of self-rated health does not seem to support the health advantage of the Hispanic population (Markides & Eschbach, 2005). For instance, Cho and colleagues (Cho, Frisbie, Hummer, & Rogers, 2004) found that Hispanics were more likely than non-Hispanic whites to rate their health status as poor, even when the actual health outcome was better. Language of the interview explains some of the difference, with Hispanics responding to interviews in Spanish reporting their health status worse than Hispanics who responded in English (Viruell-Fuentes, Morenoff, Williams, & House, 2011) given the different connotations that the categories for rating health have in Spanish when compared to English (Angel & Guarnaccia, 1989; Idler & Benyamini, 1997). Regardless of these findings, self-rated health might reflect the actual general health or well- being of the population. Reporting good health can be interpreted as a reflection of the health risk that population perceives and their interpretation of their status, which overcomes the differences in language and cultural pessimism (Markides, Rudkin, Angel, & Espino, 1997). The role of time allocated to different activities among the Hispanic population has not been taken into account. Research has found that Mexican immigrant populations in the United States tend to spend less time than native groups in leisure and socialization activities, and that Mexican-

• rigin immigrants are particularly affected (Hamermesh & Trejo, 2013). Furthermore, in studies

where physical activities are taken into account, the differences better health outcomes hold for Mexican-origin populations when compared to native groups, net of socioeconomic status (Winkleby & Cubbin, 2004). Poverty of time as a fundamental cause of health disparities

8

According to Link and Phelan (1995), social conditions involve a person’s relationships with

• ther people within the social and economic structures of society and are determinant for the onset

and evolution of illness. As a consequence, social conditions are fundamental to understand the differences in health status across groups in the society. Therefore, these social conditions can be combined with the demographic characteristics of the population to understand differences in health status across groups. Poverty of time, as well as other dimensions of poverty, results in reduced access to the necessary resources to achieve a satisfactory basic standard of living. However, poverty of time does not operate in the same way as income poverty. On the one hand, some individuals can

• vercome the lack of time by using their income to purchase market substitutes such as [healthy]

prepared food, domestic workers, among other necessities, while other individuals will lack the necessary income to overcome this shortage of time (Kalenkoski & Mahrick, 2012; Zacharias, 2011; Zacharias et al., 2012). Also, lacking enough free time is associated with fewer hours for socializing, leisure and religious participation, activities that can improve perceptions of quality of life and well-being (Adler & Newman, 2002; Spinney & Millward, 2010). On the contrary, having time for leisure and participation in social and altruistic activities can also improve the perception

• f mental and physical health and the assessment of one’s health status (Jackson, 2000; Lum &

Lightfoot, 2005; Piliavin & Siegl, 2007). Model for poverty of time, leisure and social activities and self-rated health Figure 1 displays the model guiding the present analysis. I include the time poverty dimension to the traditional measures of socioeconomic status such as education, income, poverty and employment, among the fundamental causes related to disparities in health outcomes in order

9

to better understand self-rated health patterns among the Mexican-origin population in the U.S. I follow a comparative perspective with other racial/ethnic groups, which is part of the basic demographic characteristics. Then, I include the participation in specific activities that have been found to have a direct impact on the likelihood of having good perceived health and to promote healthy behaviors such as leisure activities and participation in social events. Specifically, this dimension explores the impact of time in physical activities and religious participation on self- rated health status. Figure 1 Model explaining differences in time allocated to housework and care, poverty of time and self-rated health across racial/ethnic groups and gender. Research questions and hypotheses The aim of this chapter is to identify the effects of time allocation in self-rated health status among racial/ethnic and gender lines in the U.S. As was the case in previous chapters, particular interest in paid to the acculturation effects experienced by the Mexican-origin populations in the

• country. For this, the following research questions will lead the analysis:

10

• 1. How does poverty of time affect self-rated health, net of other dimensions of socioeconomic

status?

• 2. Does time spent in leisure and social activities increases the likelihood of reporting better

health, net of other socioeconomic indicators? 2.1 Does time spent in religious and physical activities improve self-rated health status?

• 3. Do the effects of time poverty, leisure and social engagement on self-rated health outcomes vary

by race and ethnicity? I expect the effect of time poverty to be negatively associated with self-rated health

• utcomes by accounting for this characteristic as a dimension of SES now being controlled in the
• models. Also, I expect time spent in leisure and social participation, particularly in religious and

physical activities, to be positively associated with perceived good health. I expect the effects of time poverty to differ across the different racial/ethnic groups included in the analysis; however, I expect the effects of specific activities such as physical activities and religious involvement, to remain in the same direction. Data and methods Data source and measures Data for the present analysis comes from the American Time Use Survey (ATUS) obtained from the ATUS-X extract builder (Hofferth, Flood, & Sobek, 2013). The goal of the ATUS is to measure how people divide their time among different activities, during a designated period of 24 hours, supporting estimations of time allocation for individuals aged 15 and over across population subgroups and over time (Bureau of Labor Statistics, 2014).

11

From 2006 to 2013, with the exception of 20092, the ATUS asked the respondents to assess their general health status in five categories from poor to excellent. The dependent variable for the present analysis is self-rated health status. The variable was dummy coded as good if the respondent rated his or her health as good, very good or excellent; and poor if the responded selected fair or poor health status. After eliminating the missing values for the outcome variable in the dataset, a total of 78,566 cases were available for the analysis. Race/ethnicity, time poverty and time in leisure and social activities are the key independent variables for this analysis. Race/ethnicity is a categorical variable identifying non- Hispanic whites (reference group), non-Hispanic blacks, Mexican, and Mexican-Americans populations in the sample. Time poverty is a dummy variable constructed based on the average time spent in activities related to leisure and socialization from the total population in the ATUS. A person was time poor if he or she was in the lowest quartile of the general distribution, the population in any of the other three quartiles of the distribution was classified as not time poor (reference group). Total time for leisure and social activities are continuous variables obtained from the time-diary by adding the time the respondent spent on general leisure and social activities. Other variables are included in the analyses to control for characteristics that are likely to influence the outcome. Among the controls included are sex and age. Sex is included as a dummy variable (male as reference). Age group is included as a categorical variable in 10-year intervals from 15-24 (reference) to 65 and above. Educational attainment identifies if the person has less than high school, high school diploma (reference category), or some education after high school. Marital status is a variable that identifies if the person is married (reference), single, or was

2 For convenience I refer to the period 2006-2013, even when 2009 is not included.

12

divorced, widowed, or separated. Labor force participation is also included in three categories: the person was employed (reference), unemployed, or not in the labor force. The final variable corresponds to poverty status. The poverty variable was calculated based on the poverty thresholds for 2010 that take into account family income reported in the survey and household size (U.S. Census Bureau, n.d.). Missing information about household income was imputed with an algorithm based on sex, age, region, and household size. In a second stage in the analysis, the poverty of time variable is replaced by time allocated to leisure, religious and volunteer activities. Leisure activities are divided between passive leisure,

• r time for relaxation, watching TV and rest, and active leisure, or time practicing sports,

exercising or working out. Also, I add controls for the total number of activities throughout a day and the average duration of the activity. Analytical approach The first part of the analysis consists of a bivariate analysis with significance tests for significant differences in health status by the characteristics being used as independent and control

• variables. The second part includes a multivariate analysis based on nested logistic regression

models to measure the effect of the independent variables on self-rated health status while controlling for other characteristics. Finally, a set of stratified models is included in order to

• bserve the different effects for racial/ethnic groups. Therefore, the general expression for the

models is: !"#$% &' = )* + ),-,

. ,/0

with wi =1 representing reporting poor health status, and bk the effects for the control variables.

13

Results Bivariate results Figure 1 displays the percentages of each racial/ethnic group reporting poor health status accompanied by the 95% confidence interval. Overall, 16% of the population rate their health as poor, but significant differences can be observed across racial/ethnic lines. Non-Hispanic whites have the lowest percentage of poor health reports, with 14% of the population; followed by immigrants from countries different from Mexico with 15%; however, the differences are not statistically significant, since the confidence intervals for these two groups overlap. About 18% of Mexican-Americans reported poor health status, followed by 23% of non-Hispanic blacks and almost 28% of Mexican immigrants; all these percentages were statistically different. Figure 2 Percentage of the population indicating poor health by race/ethnicity, U.S. 2006-2013

14

These percentages provide preliminary evidence for the existence of significant differences in self-rated health status, with the proportion of Mexican immigrants reporting poor health being twice the proportion observed among non-Hispanic whites. Some preliminary conclusions can be drawn from the bivariate analysis from Table 1. Reports of poor health were more common among women (16.5%) than among men (14.9%), and the difference is significant. Poor health reports, as expected, increase with age going from 8.6% among the youngest group to 27% among the oldest one. Marital status also displays significant differences in reports of poor health, with single individuals having the smallest values (12.2%), followed by married (14.6%), separated or divorced (23.3%), and widows (30.8%). Higher levels

• f education are related to reductions in the percentages reporting poor health; while 26.6% of

those without high school diploma report poor health status, this percentage is reduced to nearly 20% among those who graduated from high school and even more to 10.3% among those with some higher education. Poor health status reports are more common among unemployed individuals (16.4%) than among those who are employed (10%), but way more common among those out of the labor force (29.3%). Reports of poor health are also different by poverty status; among those living above the poverty threshold, only 14% reported poor health, half of the reports among those living in poverty (29.2%). Time poverty presented the opposite patterns of income poverty, the percentage of those who reported poor health was lower among those in time poverty (11.8%) than those who are not time poor (17.1%). Engagement in another type of activities also seems to be associated with differences in self-reported health status. About 13.7% of those who do not engage in leisure activities report their health to be poor, while the percentage is almost 16% among those who do participate in

• leisure. Participation is sports and exercise results in a larger difference. While 17% of those who

15

do not practice these activities reported poor or bad health, this percentage is only 9.7% among those who engage in the activity. Religious involvement does not seem to result in significant differences in health status reports. Finally, those who engage in volunteer activities appear to have a better perception of their health, since 10.7% of them reported poor health, compared to 16% among those who did not engage in these activities. Overall, this preliminary analysis indicates the existence of significant differences across racial/ethnic groups, gender, socioeconomic status and participation in different activities, which points to the convenience of the analysis using multivariate techniques.

Table 1 Self-rated health status by demographic and socioeconomic characteristics, U.S. 2006-2013.

Est.

95% C.I.

Est.

95% C.I.

Est.

95% C.I.

Est.

95% C.I.

Total 15.77 (15.47 - 16.08) 84.23 (83.92 - 84.53) Employment status Race/ethnicity Employed 10.02 (9.70 - 10.35) 89.98 (89.65 - 90.30) Non-Hispanic white 13.74 (13.37 - 14.12) 86.26 (85.88 - 86.63) Unemployed 16.36 (14.93 - 17.89) 83.64 (82.11 - 85.07) Mexican immigrant 27.71 (25.81 - 29.70) 72.29 (70.30 - 74.19) Not in the labor force 29.31 (28.56 - 30.07) 70.69 (69.93 - 71.44) Mexican-American 18.24 (16.53 - 20.08) 81.76 (79.92 - 83.47) Poverty Non-Hispanic black 23.02 (22.05 - 24.03) 76.98 (75.97 - 77.95) Not in poverty 14.02 (13.70 - 14.34) 85.98 (85.66 - 86.30) Other immigrant 15.23 (14.25 - 16.27) 84.77 (83.73 - 85.75) In poverty 29.17 (28.06 - 30.30) 70.83 (69.70 - 71.94) Gender Poverty of time Male 14.85 (14.41 - 15.30) 85.15 (84.70 - 85.59) Not time poor 17.06 (16.68 - 17.45) 82.94 (82.55 - 83.32) Female 16.65 (15.47 - 16.08) 83.35 (82.89 - 83.80) Time poor 11.84 (11.29 - 12.40) 88.16 (87.60 - 88.71) Age Leisure activities 15-24 8.63 (7.90 - 9.42) 91.37 (90.58 - 92.10) No 13.67 (12.15 - 15.34) 86.33 (84.66 - 87.85) 25-34 10.21 (9.54 - 10.92) 89.79 (89.08 - 90.46) Yes 15.86 (15.54 - 16.19) 84.14 (83.81 - 84.46) 35-44 12.64 (11.97 - 13.35) 87.36 (86.65 - 88.03) Sport activities 45-54 16.44 (15.69 - 17.21) 83.56 (82.79 - 84.31) No 17.20 (16.85 - 17.56) 82.80 (82.44 - 83.15) 55-64 21.43 (20.49 - 22.41) 78.57 (77.59 - 79.50) Yes 9.74 (9.13 - 10.39) 90.26 (89.61 - 90.87) 65 and older 27.06 (26.14 - 28.00) 72.94 (72.00 - 73.86) Religious activities Marital status No 15.71 (15.38 - 16.05) 84.29 (83.95 - 84.62) Single 12.21 (11.61 - 12.85) 87.79 (87.15 - 88.39) Yes 16.40 (15.41 - 17.45) 83.60 (82.55 - 84. 59) Married 14.62 (14.21 - 15.04) 85.38 (84.96 - 85.79) Volunteer activities Separated/divorced 23.32 (22.42 - 24.24) 76.68 (75.76 - 77.58) No 16.14 (15.82 - 16.47) 83.86 (83.53 - 84.18) Widow 30.82 (29.39 - 32.29) 69.18 (67.71 - 70.61) Yes 10.67 (9.66 - 11.77) 89.33 (88.23 - 90.34) Education Less than high school 26.26 (25.29 - 27.67) 73.74 (72.74 - 74.71) High school 19.68 (19.03 - 20.34) 80.32 (79.66 - 80.97) More than high school 10.36 (10.03 - 10.70) 89.64 (89.30 - 89.97)

Source: 2006-2013 American Time Use Survey (ATUS).

Health status Poor Good Health status Poor Good Characteristic Characteristic

165

17

Multivariate models Table 2 displays the odd-ratios from five nested models for the likelihood of reporting poor health status. Model I controls only for the race/ethnicity variable; Model II adds controls for demographic characteristics; Model III controls for socioeconomic status; Model IV controls for engagement in specific activities; and, model V includes controls for the total number of activities and the average duration of each activity. The odd ratios represent the likelihood of the outcome to present given the independent variable; values above 1 mean the covariate is related to increased

• dds of observing the outcome, while values below 1 indicate the covariate is negatively associated

with the outcome. Compared to non-Hispanic whites — the reference category—, all other race/ethnicities are associated with higher odds to report poor health, particularly Mexican immigrants and non- Hispanic blacks. These effects are even stronger after adding controls for demographic

• characteristics. Gender is associated with differences in health perceptions, with women being

more likely than men to report poor health status. Age, as expected, is positively associated with poor health reports, which means that older people are more likely than young people to report their health status as poor. Married populations are less likely to report poor health than single individuals, but the opposite is observed among ever married individuals. Controls for socioeconomic status moderates the race/ethnicity effect, but it continues to be significant. However, these controls account for gender differences. Among the socioeconomic controls, higher education is associated with smaller odds to report poor health, while unemployment and being out of the labor force are associated with increases in the likelihood of poor health reports when compared to employed individuals. Poverty status increases the odds of reporting poor health, but time poverty does not have a significant effect.

18

Table 2 Odds ratios for the likelihood of reporting poor health, U.S. 2006-2013. (continued)

OR

Sig.

OR

Sig.

OR

Sig.

OR

Sig.

OR

Sig.

Race/ethnicity Non-Hispanic white† Mexican immigrant 2.4067 *** 3.4804 *** 1.9001 *** 1.9261 *** 1.8929 *** Mexican-American 1.4007 *** 2.0548 *** 1.5624 *** 1.5579 *** 1.5737 *** Non-Hispanic black 1.8777 *** 1.9256 *** 1.5396 *** 1.5337 *** 1.4889 *** Other immigrants 1.1282 ** 1.2846 *** 1.1661 ** 1.1749 ** 1.1487 ** Sex Males† Females 1.0568 * 0.9490 0.9394 * 0.9923 Age group 15 - 24† 25 - 34 1.3755 *** 2.4450 *** 2.2638 *** 2.2644 *** 35 - 44 1.9004 *** 3.3883 *** 3.1624 *** 3.1517 *** 45 - 54 2.7268 *** 4.7019 *** 4.3848 *** 4.2904 *** 55 - 64 3.9504 *** 5.7152 *** 5.3580 *** 5.2167 *** 65 and above 5.3431 *** 4.3795 *** 4.1482 *** 4.0312 *** Marital status Single† Married 0.6990 *** 0.7532 *** 0.7641 *** 0.7872 *** Separated/Divorced 1.1694 ** 1.1735 ** 1.1633 ** 1.1802 ** Widowed 1.0982 0.9516 0.9458 0.9633 Educational attainment Less than high school† High school 0.6783 *** 0.6749 *** 0.6866 *** More than high school 0.3965 *** 0.4172 *** 0.4337 *** Employment status Employed† Unemployed 1.6545 *** 1.6577 *** 1.6625 *** Not in the labor force 2.9575 *** 3.0168 *** 2.9956 *** Model V Variable Model IV Model III Model II Model I

19

Table 2 (continued) The fourth model, including participation in different activities as a control variable, does not display substantial changes in the other control variables, with the exception of gender, where women are now less likely than men to report poor health, and the time poverty variable, which is negatively associated with poor health status. In this model, passive leisure is not related to health status; participation in physical activities is associated with fewer odds to report poor health status. Also, involvement in religious and volunteering activities reduces the odds of reporting poor health status. The final model includes controls for the total number of activities reported on a regular day and the average duration of each activity. In this model, the effects of the main independent

OR

Sig.

OR

Sig.

OR

Sig.

OR

Sig.

OR

Sig.

Poverty status Not in poverty† In poverty 1.6649 *** 1.6293 *** 1.6070 *** Time poverty Not in poverty† In poverty 0.9921 0.8831 ** 0.8930 ** Participation in specific activities1 Passive leisure 0.9061 1.0161 Active leisure 0.5553 *** 0.5834 *** Religious 0.8042 *** 0.8549 ** Volunteering 0.6593 *** 0.7077 *** Total number of activities 0.9922 * Average duration of activity 1.0023 *** Intercept 0.1593 *** 0.0694 *** 0.0603 *** 0.0804 *** 0.0638 *** Pseudo-R2 0.0595 0.0595 0.1206 0.1280 0.1305 AIC 66407 66407 62108 61593 61419

Source: 2003-2014 American Time Use Survey (ATUS).

† Reference category. 1 Reference category: no participation.

Significance levels: *0.05 **0.01 ***0.001

Model V Variable Model IV Model III Model II Model I

20

variable and controls were similar those observed in model IV. Regarding the new control variables, total number of activities is negatively associated with reports of poor health, but the average duration of the activities is associated with increasing perceptions of poor health. To better understand this seemingly contradictory relationship between the last two control variables I included an interaction model between the two of them. Table 3 displays the results for the stratified models by race and ethnicity based on model V from the series of nested models discussed above. The effect of gender was only significant among non-Hispanic whites, with women having 15% higher odds of reporting good health when compared to males. The effect of age, on the contrary, was similar among all racial and ethnic

• groups. Being married was significant only among non-Hispanic whites and immigrants from other

countries, who have odds 20% and 27% lower than the odds of single individuals to report poor

• health. The effect of being separated or divorced was only significant for non-Hispanic whites,

and the effect of widowhood was significant only among Mexican immigrants. Educational attainment was negatively associated with the likelihood of reporting poor health; however, among non-Hispanic blacks, this effect was only observed for those with higher

• education. Being unemployed or not in the labor force is associated with higher odds of reporting

poor health except among Mexican immigrants. Poverty was significant for all race/ethnicities and displayed the expected behavior, where those living below the poverty threshold being more likely to report poor health than those living above the poverty threshold. Time poverty was only significant for non-Hispanic whites and immigrants from countries different from Mexico. The direction of the effect was negative, which indicates that being time poor reduces the likelihood of reporting poor health status.

21

Participation in passive leisure did not display a significant effect, but time in active leisure was negatively associated with poor health, but only among non-Hispanic whites, non-Hispanic blacks and immigrants from countries different from Mexico. As expected, engagement in this activity resulted in fewer odds of reporting poor health status. Participation in religious activities, as well as volunteering, were associated with lower odds to report poor health status among non- Hispanic white and black populations. While participation in specific activities may result in changes in the perception of one’s health status, time in the activity will determine the magnitude of the effect. For that reason, a new series of models were fitted for the likelihood of reporting poor health, but replacing the participation variables by the time spent in that activity. To avoid confusion between participation in the activity and the intensity of the involvement, the new models restricted the sample to those who reported engaging in the activity. Hence, the first model is restricted to those who engaged in passive leisure and the variable for participation in passive leisure was replaced for minutes spent in these activities, all the other activity-specific variables in this model remained the same, indicating participation or not. The second model is restricted to those who engaged in active leisure and controls for time in this activity while controlling for involvement in the other three groups of activity. The third model is restricted to those who engaged in religious activities, and the fourth model to those individuals that reported spending some time performing volunteering activities.

22

Table 3 Odds ratios for the likelihood of reporting poor health from the stratified models by race/ethnicity, U.S. 2006-2013.

OR

Sig.

OR

Sig.

OR

Sig.

OR

Sig.

OR

Sig.

Age group 15 - 24† 25 - 34 2.2114 *** 1.3935 1.9604 ** 2.5227 *** 3.1639 *** 35 - 44 3.3878 *** 1.6476 * 2.4451 ** 2.7928 *** 4.3966 *** 45 - 54 4.2848 *** 2.6734 *** 3.6752 *** 3.8631 *** 6.4211 *** 55 - 64 5.0350 *** 3.6148 *** 4.9266 *** 5.1864 *** 6.9345 *** 65 and above 3.5127 *** 3.9932 *** 5.5451 *** 3.7978 *** 7.4443 *** Sex Males† Females 0.9060 * 1.2806 1.2128 1.1954 * 1.2744 * Marital status Single

†

Married 0.7990 ** 0.8559 0.9536 0.8611 0.7277 * Separated/Divorced 1.2204 ** 0.8820 1.3789 1.1438 1.1031 Widowed 0.9817 2.3034 * 0.5474 0.9343 0.9009 Educational attainment Less than high school

†

High school 0.6688 *** 0.5533 *** 0.6080 ** 0.9494 0.6077 *** More than high school 0.4254 *** 0.4649 *** 0.4091 *** 0.5824 *** 0.3616 *** Employment status Employed

†

Unemployed 2.1019 *** 0.7062 1.9060 * 1.4849 ** 1.1204 Not in the labor force 3.7246 *** 1.2508 1.7180 ** 3.3114 *** 1.9283 *** Poverty status Not in poverty† In poverty 1.7470 *** 1.4911 ** 1.5692 ** 1.3442 *** 1.7950 *** Time poverty Not in poverty† In poverty 0.8448 ** 0.9958 1.0819 1.1297 0.7394 ** Participation in specific activities

1

Pasive leisure 1.0364 1.2923 0.9262 0.9725 0.8357 Active leisure 0.5128 *** 1.0621 0.7120 0.6047 *** 0.7506 * Religious 0.8621 * 0.9047 0.9381 0.7893 * 0.8446 Volunteering 0.6945 *** 1.2181 0.7942 0.6146 ** 0.7765 Total number of activities 0.9907 * 0.9818 0.9846 1.0081 0.9908 Average duration of activity 1.0021 * 1.0022 1.0019 1.0027 * 1.0027 Intercept 0.0650 *** 0.1916 ** 0.1400 *** 0.0517 *** 0.0740 *** Pseudo-R2 0.1350 0.0796 0.1021 0.1258 0.1167 AIC 38747 4211 3249 8284 6625

Source: 2006-2013 American Time Use Survey (ATUS).

†

Reference category. Significance levels: *0.05 **0.01 ***0.001

Variable Non-Hispanic white Non-Hispanic black Mexican immigrant Other immigrants Mexican- American

23

Figure 3 displays the marginal effects of time in each specific group of activities by racial/ethnic identification. The marginal effects indicate the gains in the probability of reporting poor health. Therefore, among the four activities only passive leisure is associated with increasing probability of reporting poor health. Moreover, the magnitude of the marginal effects for each additional minute spent in passive leisure are significantly higher than the effects of the other

• activities. 3 Notably, as time in passive leisure increases, the probability of reporting poor health

increases considerable, and this effect is observed for all racial/ethnic groups. On the other hand, active leisure and volunteering activities have a strong effect in reducing the probability of reporting poor health, particularly for Mexican immigrants. Figure 3 Marginal effects on the probability of reporting poor health by group-specific activity and racial/ethnic identification, U.S. 2006-2013.

3 The graph for passive leisure is at a different scale than the other graphs.

24

Acculturation and health status In order to account for the effect of acculturation, I fitted three additional models for Mexican-origin populations controlling for language in which the interview was conducted. In the first of these models, I replaced the race/ethnicity variable by a new one containing five categories: non-Hispanic white, Mexican-American, Mexican immigrant with ten years or more in the U.S., Mexican immigrant with five to ten years in the country, and Mexican immigrants with less than five years in the country. In the second model I include a variable indicating if the interview was conducted in Spanish, since language use is one dimension of acculturation (Cabassa, 2003; Marin & Sabogal, 1987) and has been linked to differences in self-rated health (Deyo, Diehl, Hazuda, & Stern, 1985; Viruell-Fuentes et al., 2011). In the third model, I include an interaction term between the populations in this section of the analysis and language of the interview. Table 4 displays the results of the models of acculturation. Compared to non-Hispanic whites, all Mexican-origin populations have higher odds to report poor health. These odds vary from 1.5 times among Mexican-Americans and immigrants that have been in the U.S. 5-10 years to 2.6 among the most recent immigrants. When the control for the language of the interview is included, the effect for immigrants who have been in the country from 5 to 10 years is not significant, and the effects for the other groups are reduced. In this tone, those who reported answering the interview in Spanish have odds to report poor health 62% higher than those answering in English. To better understand the effect of language and time in the U.S. as mediators for self-rated health status, the marginal effects for these two variables were estimated. These effects are displayed in figure 4.

Table 4 Odds ratios for the likelihood of reporting poor health from the acculturation models, U.S. 2006-2013.

OR

Sig.

OR

Sig.

OR

Sig.

OR

Sig.

Race/ethnicity Employment status Non-Hispanic white† Employed† Mexican-American 1.5324 *** 1.4385 *** Unemployed 1.8637 *** 1.8754 *** Mexican immigrant, 10+ years 1.8520 *** 1.3351 ** Not in the labor force 3.1791 *** 3.1888 *** Mexican immigrant, 5-10 years 1.5401 *** 1.0594 Poverty status Mexican immigrant, < 5 years 2.6471 *** 1.7835 * Not in poverty† Age group In poverty 1.6791 *** 1.6535 *** 15 - 24† Time poverty 25 - 34 2.1231 *** 2.0982 *** Not in poverty† 35 - 44 3.0905 *** 3.0631 *** In poverty 0.8825 ** 0.8818 ** 45 - 54 4.1465 *** 4.1112 *** Participation in specific activities1 55 - 64 5.0050 *** 4.9657 *** Pasive leisure 1.0702 1.0743 65 and above 3.7342 *** 3.7062 *** Active leisure 0.5588 *** 0.5590 *** Sex Religious 0.8671 * 0.8675 * Males† Volunteering 0.7218 *** 0.7204 *** Females 0.9265 * 0.9245 * Total number of activities 0.9907 * 0.9911 * Marital status Average duration of activity 1.0023 ** 1.0023 ** Single† Language of the interview Married 0.8003 *** 0.7928 *** Other than Spanish Separated/Divorced 1.2066 ** 1.2064 ** Spanish 1.6241 *** Widowed 0.9949 0.9900 Intercept 0.0665 *** 0.0653 *** Educational attainment Pseudo-R2 0.1313 0.1321 Less than high school† AIC 46394 46354 High school 0.6555 *** 0.6724 *** More than high school 0.4196 *** 0.4311 ***

Source: 2006-2013 American Time Use Survey (ATUS).

†

Reference category. Significance levels: *0.05 **0.01 ***0.001

Model I Model II Variable Model I Model II Variable

174

Consistently across groups, individuals responding the survey in Spanish show higher effects for reporting poor health outcomes when compared to those answering in English. Conversely, those who have been the most time in the U.S. have a smaller effect. Mexican- Americans have the smallest difference in the marginal effects by language of the interview while the biggest difference is observed among the most recent immigrants. Figure 4. Marginal effects of language of the interview for the probability to report poor health, U.S. 2006-2013. Discussion and conclusions The present analysis of the factors associated with the self-rated health status of the population and the effects of time constraints highlight a complex association among these

• factors. All demographic and socioeconomic characteristics affected self-rated health the

way they were expected with older people being more likely to report poor health than young people, and with small effects by gender, and with traditional controls such as educational attainment, poverty and employment following the expected path where those in relative

disadvantage being more likely to rate their health as poor. Of central interest in the present analysis were the differences across racial/ethnic groups. When compared to non-Hispanic whites, all other ethnic groups were more likely to perceive their health to be poor, even after adding controls for other characteristics of the population. As previous research has found with self-rated health status, Mexican-origin populations rate their health poorly and worse than other populations, in particular when compared to non-Hispanic whites, even when their health outcomes are no different or even better (Cho et al., 2004; Franzini & Fernandez-Esquer, 2004; Markides & Eschbach, 2005). This does not mean that reports of health status are not to be used. According to Ferraro & Farmer (1999), self-rated health status constitutes a better measure of general health among minority populations than among non-Hispanic whites even after controlling for contextual variables. The first research question was related to the effect of time poverty on self-rated

• health. I expected to find a negative association between self-rated health and poverty of

time; this was not the case. Poverty of time is the result of people spending more time than the average population in paid and unpaid work and having little time for leisure and social

• activities. Two possible explanations can be used for this behavior. On the one hand, more

time spent working means less time for leisure and social activities; a strong negative correlation exists between employment and poor health (Minton, Pickett, & Dorling, 2012); also, some unemployed population may exaggerate their health conditions to justify their nonparticipation in the labor force (Kreider, 1999). Moreover, socioeconomic conditions linked to employment are well-known moderators of illness and mortality risks by reducing exposure to hazards, but also increasing the resources available to face an eventual disease

(Rogers et al., 2005). On the other hand, time poverty might not operate in a direct way to affect self-rated health status but rather through stress associated with engagement in different activities (Zuzanek, Robinson, & Iwasaki, 1998) Participation in specific activities did show significant effects; particularly, participation in active leisure — in the form of exercise — as well as religious and volunteering activities were associated with reductions in poor health reports. Similarly, the total number of activities performed throughout a day resulted in significant reductions in the likelihood of reporting having poor health. These findings are supported by previous research indicating the benefits of physical activities for health status (Haskell et al., 2007; Kalenkoski & Mahrick, 2012; Pate et al., 1995). Also, religious participation has been linked to the existence of social controls and the promotion and maintenance of good health (Ellison & Levin, 1998; Hummer et al., 1999). Finally, engaging in formal volunteering activities has been found to reinforce satisfaction with one’s status by comparing one’s health status with the groups being assisted (Borgonovi, 2008; Lum & Lightfoot, 2005; Piliavin & Siegl, 2007). The last research questions referred to the differential effect of time poverty and leisure and social activities by racial/ethnic group. Poverty of time was only significant among non-Hispanic whites and immigrants from countries different from Mexico and had a negative effect on the likelihood of reporting poor health. Participation in active leisure was negatively associated with poor self-rated health among non-Hispanic whites, non-Hispanic blacks and other immigrants. Finally, time in religious activities and volunteering reduced the likelihood of reporting poor health for non-Hispanic whites and non-Hispanic blacks. Finally, when dealing with immigrant populations, the effect of acculturation should be taken into account. In the present analysis, I approach this matter by analyzing Mexican-

• rigin populations and language of the interview. In line with previous research, time in the

U.S., particularly among the Spanish-speakers, was associated with declines in the likelihood

• f reporting poor health as a consequence of the process of cultural assimilation (Lara et al.,

2013; Morales, Lara, Kingston, Valdez, & Escarce, 2002), but also as a consequence of differences in the perception of health based on language differences (Viruell-Fuentes et al., 2011). It is necessary to acknowledge some limitations of this analysis. First, the diary form

• nly asked about activities conducted the day before of the interview while health status

represents a more long-term state. Also, the effect of specific activities might only be significant if conducted on specific days of the week, such as church attendance and participation in sports during the weekends. Furthermore, the use of exercising facilities is usually limited by the context of accessibility. Minority populations tend to live in poorer areas, with high crime rates, and lack of recreational facilities, which discourages engaging in physical activities (Williams & Collins, 2013a; Williams & Jackson, 2005). Future research should take into account the characteristics of the physical environment and context in which the different populations live together with time constraints. A final consideration is related to the direction of the causal effect. Evidence suggests that socioeconomic conditions are unlikely to affect health directly, but they shape life conditions that influence health (Adler & Rehkopf, 2008). Socioeconomic conditions affect health outcomes by differential access to the necessary resources to promote the maintenance of good health (Elo, 2009; Hummer & Lariscy, 2011; LaVeist, 2005; Rogers et al., 2005). In the present analysis, time poverty was included as a new dimension of the socioeconomic spectrum. Following the fundamental

cause perspective (Link & Phelan, 1995), poverty of time represents a new socioeconomic dimension and controlling for this characteristic should account for some of the disparities in health status. Time poverty did not have a direct effect on self-rated health, except for non- Hispanic whites; however, time spent in specific activities did affect health. Specifically, time in exercise and time in social activities, such religious and volunteering activities, were found to improve self-rated health.

REFERENCES Abraído-Lanza, A., Dohrenwed, B., Ng-Mak, D., & Turner, B. (1999). The Latino Mortality Paradox: A Test of the “Salmon Bias” and Healthy Immigrant Hypotheses. American Journal of Public Health, 89(10), 1543–1548. https://doi.org/10.2105/AJPH.89.10.1543 Acevedo-Garcia, D., & Bates, L. M. (2008). Latino Health Paradoxes: Empirical Evidence, Explanations, Future Research, and Implications. In H. Rodríguez, R. Sáenz, & C. Menjívar (Eds.), Latinas/os in the United States: Changing the Face of América (pp. 101–113). Springer. Adler, N. E., & Newman, K. (2002). Socioeconomic Disparities In Health: Pathways And Policies. Health Affairs, 21(2), 60–76. https://doi.org/10.1377/hlthaff.21.2.60 Adler, N. E., & Rehkopf, D. H. (2008). U.S. Disparities in Health: Descriptions, Causes, and

• Mechanisms. Annual Review of Public Health, 29(1), 235–252.

https://doi.org/10.1146/annurev.publhealth.29.020907.090852 Adler, N. E., & Stewart, J. (2010). Health disparities across the lifespan: Meaning, methods, and

• mechanisms. Annals of the New York Academy of Sciences, 1186(1), 5–23.

https://doi.org/10.1111/j.1749-6632.2009.05337.x Angel, R., & Guarnaccia, P. J. (1989). Mind, body, and culture: somatization among Hispanics. Social Science and Medicine, 28(12), 1229–1238. https://doi.org/10.1016/0277- 9536(89)90341-9 Anxo, D., Mencarini, L., Pailhé, A., Solaz, A., Tanturri, M. L., & Flood, L. (2011). Gender differences in time use over the life course in France, Italy, Sweden, and the US. Feminist Economics, 17(3), 159–195. Bardasi, E., & Wodon, Q. (2006). Measuring Time Poverty and analyzing Its Determinants: Concepts and Application to Guinea. Becker, G. (1965). A Theory of the Allocation of Time. The Economic Journal, 75(299), 493–517. Bird, C., & Fremont, A. (1991). Gender, time use, and health. Journal of Health and Social Behavior, 32(2), 114–129. Borgonovi, F. (2008). Doing well by doing good: the relationship between formal volunteering and self-reported health and happiness. Social Science & Medicine, 66(11), 2321–2334.

Bureau of Labor Statistics. (2014). American Time Use Survey User’s Guide: Understanding ATUS 2003 to 2013. Retrieved from http://www.bls.gov/tus/atususersguide.pdf Cabassa, L. (2003). Measuring acculturation: Where we are and where we need to go. Hispanic Journal of Behavioral Sciences, 25(2), 127–146. Retrieved from http://hjb.sagepub.com/content/25/2/127.short Cho, Y., Frisbie, W. P., Hummer, R. A., & Rogers, R. G. (2004). Nativity, duration of residence, and the health of Hispanic adults in the United States. International Migration Review, 38(1), 184–211. Deyo, R. A., Diehl, A. K., Hazuda, H., & Stern, M. P. (1985). A Simple Language-based Acculturation Scale for Mexican Americans: Validation and Application to Health Care Research. American Journal of Public Health, 75(1), 51–55. Dupre, M., Franzese, A., & Parrado, E. (2006). Religious attendance and mortality: Implications for the black-white mortality crossover. Demography, 43(1), 141–164. Retrieved from http://link.springer.com/article/10.1353/dem.2006.0004 Ellison, C., & Levin, J. S. (1998). The religion-health connection: evidence, theory, and future

• directions. Health Education & Behavior, 25(6), 700–720.

Elo, I. (2009). Social class differentials in health and mortality: patterns and explanations in comparative perspective. Annual Review of Sociology, 35, 553–572. Ferraro, K. F., & Farmer, M. M. (1999). Utility of health data from social surveys: is there a gold standard for measuring morbidity? American Sociological Review, 64(2), 303–315. Franzini, L., & Fernandez-Esquer, M. E. (2004). Socioeconomic, cultural, and personal influences on health outcomes in low income Mexican-origin individuals in Texas. Social Science & Medicine, 59(8), 1629–1646. Galobardes, B., Shaw, M., Lawlor, D. A., Smith, G. D., & Lynch, J. (2006). Indicators of socioeconomic position. In J. M. O. & J. S. Kaufman (Ed.), Methods in social epidemiology (pp. 47–85). San Francisco: Jossey-Bass. George, L. K., Ellison, C., & Larson, D. B. (2002). Explaining the relationships between religious involvement and health. Psychological Inquiry, 13(3), 190–200. George, L. K., Larson, D. B., Koenig, H. G., & McCullough, M. E. (2000). Spirituality and health: what we know, what we need to know. Journal of Social and Clinical Psychology, 19(1), 102–116. Gordon, M. (1964). Assimilation in American Life: The Role of Race, Religion, and National Origins. New York: Oxford University Press. Hamermesh, D., & Pfann, G. (2005). Time-use data in economics. European Economic Review, 49(1), 1–7. Hamermesh, D., & Trejo, S. (2013). How do immigrants spend their time? The process of

• assimilation. Journal of Population Economics, 26(2), 507–530.

Harvey, A., & Mukhopadhyay, A. K. (2007). When twenty-four hours is not enough: time poverty of working parents. Social Indicators Research, 82, 57–77.

Haskell, W., Lee, I.-M., Pate, R., Powell, K., Blair, S., Franklin, B., … Bauman, A. (2007). Physical Activity and Public Health: Updated Recommendation for Adults From the American College

• f Sports Medicine and the American Hearth Association. Circulation, 116(9), 1081–1093.

Hofferth, S., Flood, S., & Sobek, M. (2013). American Time Use Survey Data Extract System: Version 2.4 [Machine-readable database]. Maryland Population Research Center, University of Maryland, College Park, Maryland, and Minnesota Population Center, University of Minnesota, Minneapolis, Minnesota. Retrieved from www.atusdata.org Hummer, R. A., & Lariscy, J. T. (2011). Educational attainment and adult mortality. In R.G. Rogers & E.M.Crimmins (Ed.), International handbook of adult mortality (pp. 241–261). New York: Springer. Hummer, R. A., Rogers, R., Nam, C., & Ellison, C. (1999). Religious involvement and US adult

• mortality. Demography, 36(2), 273–285.

Idler, E. L., & Benyamini, Y. (1997). Self-rated health and mortality: a review of twenty-seven community studies. Journal of Health and Social Behavior, 38(1), 21–37. Isaac, L. (2013). Defining health and health care disparities and examining disparities across the life

• span. In T. A. LaVeist & L. Isaac (Eds.), Race, Ethnicity, and Health: A Public Health Reader (pp.

11–31). San Francisco. Jackson, E. (2000). Will Research on Leisure Constraints Still Be Relevant in the Twenty-first Century? Journal of Leisure Research, 32(1), 62–68. Joyce, M., & Stewart, J. (1999). What can we learn from time-use data. Monthly Lab. Rev., 122, 3– 6. Kalenkoski, C. M., & Mahrick, K. S. (2012). How does time poverty affect behavior? A look at eating and physical activity. Applied Economic Perspectives and Policy, 1–17. Krantz, G., Berntsson, L., & Lundberg, U. (2005). Total workload, work stress and perceived symptoms in Swedish male and female white-collar employees. The European Journal of Public Health, 15(2), 209–214. Kreider, B. (1999). Latent Work Disability and Reporting Bias. The Journal of Human Resources, 34(4), 734–769. Lara, M., Gamboa, C., Kahramanian, M. I., Morales, L. S., & Hayes Bautista, D. E. (2013). Acculturation and Latino health in the United States: a review of the literature and its sociopolitical context. In T. A. LaVeist & L. Isaac (Eds.), Race, Ethnicity, and Health: A Public Health Reader (pp. 215–252). San Francisco: Jossey-Bass. LaVeist, T. A. (2005). Minority Populations and Health: An Introduction to Health Disparities in the United States. San Francisco: Jossey-Bass. Link, B., & Phelan, J. (1995). Social conditions as fundamental causes of disease. Journal of Health and Social Behavior, 80–94. Lum, T. Y., & Lightfoot, E. (2005). The effects of volunteering on the physical and mental health of

• lder people. Research on Aging, 27(1), 31–55.

Luo, Y., Hawkley, L., Waite, L., & Cacioppo, J. (2012). Loneliness, health, and mortality in old age: A

national longitudinal study. Social Science & Medicine, 74(6), 907–914. Retrieved from http://www.sciencedirect.com/science/article/pii/S0277953612000275 Marin, G., & Sabogal, F. (1987). Development of a Short Acculturation Scale for Hispanics. Hispanic Journal of Behavioral Sciences, 9(2), 183–205. Markides, K. S., & Coreil, J. (1986). The health of Hispanics in the southwestern United States: an epidemiologic paradox. Public Health Reports, 101(3), 253. Markides, K. S., & Eschbach, K. (2005). Aging, migration, and mortality: current status of research

• n the Hispanic Paradox. The Journals of Gerontology Series B: Psychological Sciences and

Social Sciences, 60(Special issue 2), S68–S75. Markides, K. S., Rudkin, L., Angel, R., & Espino, D. V. (1997). Health status of Hispanic elderly. In B.

• S. & L. G. Martin (Ed.), Racial and ethnic differences in the health of older Americans (pp.

285–300). National Academy Press. Marmot, M. (2002). The influence of income on health: views of an epidemiologist. Health Affairs, 21(2), 31–46. Marmot, M., Friel, S., Bell, R., Houweling, T. A. J., & Taylor, S. (2008). Closing the gap in a generation: health equity through action on the social determinants of health. Lancet, 372, 1661–1669. Martinez, J. N., Aguayo-Tellez, E., & Rangel-Gonzalez, E. (2015). Explaining the Mexican-American Health Paradox Using Selectivity Effects. International Migration Review, 49(4), 878–906. Mechanic, D. (2002). Disadvantage, inequality, and social policy. Health Affairs, 21(2), 48–59. Minton, J. W., Pickett, K. E., & Dorling, D. (2012). Health, employment, and economic change, 1973-2009: repeated cross sectional study. British Medical Journal, 344(7856), e2316. Morales, L., Lara, M., Kingston, R., Valdez, R., & Escarce, J. (2002). Socioeconomic, cultural and behavioral factors affecting Hispanic health outcomes. Journal of Health Care for the Poor and Underserved, 13(4), 477–503. Office of Disease Prevention and Health Promotion. (n.d.). Disparities. Retrieved from http://www.healthypeople.gov/2020/about/foundation-health-measures/Disparities Pablos-Méndez, A. (1994). Mortality among Hispanics: A Letter to the Editor. Journal of the American Medical Association, 271(16), 1237–1238. Palloni, A., & Arias, E. (2004). Paradox Lost: Explaining the Hispanic Adult Mortality Advantage. Demography, 41(3), 385–415. https://doi.org/10.1353/dem.2004.0024 Palloni, A., & Morenoff, J. (2001). Interpreting the paradoxical in the Hispanic paradox. Annals of the New York Academy of Sciences, 954(1), 140–174. Retrieved from http://onlinelibrary.wiley.com/doi/10.1111/j.1749-6632.2001.tb02751.x/full Pate, R., Pratt, M., Blair, S., Haskell, W., Mancera, C., Bouchard, C., … Wilmore, J. (1995). Physical Activity and Health: A Recommendation from the Centers of Disease Control and Prevention and the American College of Sports Medicine. Journal of the American Medical Association, 273(5), 402–407.

Piliavin, J. A., & Siegl, E. (2007). Health benefits of volunteering in the Wisconsin Longitudinal

• Study. Journal of Health and Social Behavior, 48(4), 450–464.

Razavi, S. (2011). World development report 2012: gender equality and development. Washington, D.C.: The World Bank. Retrieved from https://siteresources.worldbank.org/INTWDR2012/Resources/7778105- 1299699968583/7786210-1315936222006/Complete-Report.pdf Robinson, J., & Godbey, G. (1997). Time for life: the surprising ways Americans use their time. (2nd ed.). University Park, PA: Penn State Press. Retrieved from http://www.cabdirect.org/abstracts/19971811225.html Rogers, R. G., Hummer, R. A., & Krueger, P. (2005). Adult mortality. Retrieved from http://link.springer.com/chapter/10.1007/0-387-23106-4_11 Sayer, L. C. (2005). Gender, time and inequality: Trends in women’s and men’s paid work, unpaid work and free time. Social Forces, 84(1), 285–303. Spinney, J., & Millward, H. (2010). Time and money: a new look at poverty and the barriers to physical activity in Canada. Social Indicators Research, 99(2), 341–356. The Secretary’s Advisory Committee on Health Promotion and Disease Prevention Objectives for

• 2020. (2008). Recommendations for the Framework and Format of Healthy People 2020.

U.S. Census Bureau. (n.d.). Poverty thresholds by size of family and number of children, 2010. Retrieved from http://www.census.gov/hhes/www/poverty/data/threshld/ Viruell-Fuentes, E. A., Morenoff, J. D., Williams, D. R., & House, J. S. (2011). Language of Interview, self-rated health, and the other Latino health puzzle. Research and Practice, 101(7), 1306– 1313. Williams, D. R., & Collins, C. (2013a). Racial residential segregation: a fundamental cause of racial disparities in health. In T. A. LaVeist & L. Isaac (Eds.), Race, Ethnicity, and Health: A Public Health Reader (pp. 331–354). San Francisco: Jossey-Bass. Williams, D. R., & Collins, C. (2013b). U.S. Socioeconomic and racial differences in health: patterns and explanations. In T. A. LaVeist & L. Isaac (Eds.), Race, Ethnicity, and Health: A Public Health Reader (pp. 375–418). San Francisco: Jossey-Bass. Williams, D. R., & Jackson, P. B. (2005). Social sources of racial disparities in health. Health Affairs, 24(2), 325–334. Winkleby, M. A., & Cubbin, C. (2004). Changing patterns in health behaviors and risk factors related to chronic diseases, 1990-2000. American Journal of Health Promotion, 19(1), 19–27. Zacharias, A. (2011). The Measurement of Time and Income Poverty (No. 690). New York, NY. Zacharias, A., Antronopoulos, R., & Masterson, T. (2012). Why time deficits matter: implications for the measurement of poverty. In World Economic Association Conference “The political economy of economic metrics.” Zuzanek, J., Robinson, J. P., & Iwasaki, Y. (1998). The Relationship Between Stress, Health, and Physically Active Leisure as a Function of Life-Cycle. Leisure Sciences, 20(4), 253–275.