Social identity, perceived urban neighborhood quality, and physical inactivity: A comparison study of China, Taiwan, and South Korea

Health & Place 41 (2016) 1–10

Contents lists available at ScienceDirect

Health & Place journal homepage: www.elsevier.com/locate/healthplace

Social identity, perceived urban neighborhood quality, and physical inactivity: A comparison study of China, Taiwan, and South Korea Duan-Rung Chen a,n, Yi-Ching Lin b a b

Institute of Health Behaviors and Community Sciences, Center for Population and Gender Studies, National Taiwan University, Taiwan Institute of Health Behaviors and Community Sciences, National Taiwan University, Taiwan

art ic l e i nf o

a b s t r a c t

Article history: Received 21 October 2015 Received in revised form 4 June 2016 Accepted 7 June 2016

Asian countries are currently witnessing unprecedented increase in physical inactivity and subsequent negative health outcomes; however, few cross-country studies documenting this trend exist. This paper presents the findings of a nationally representative sample, based on the East Asian Social Survey in 2011. The study sought to examine the association of social identity factors, such as objective socio-economic position, perceived social status and neighborhood quality with physical inactivity, while controlling for psychosocial and physical health. A sample of 5222 adults living in urban areas across China, Taiwan, and South Korea were surveyed. Methods: Multivariate nested logistic regressions were constructed. Results: Perceived social status was positively associated with physical activity. Gender difference in physical activity was significant, and this difference widened as educational levels increased. Class division in physical activity was also found. Perceived physical and social features of neighborhood such as suitability for walking and jogging, air quality, and help from neighbors were to different degrees associated with physical inactivity. Conclusion: Gender, marital status, education and perceived social status were common factors associated with physical inactivity in East Asian countries. Perceived urban neighborhood quality is particularly important for Chinese people to stay physically active. Cultural-behavioral norms for physical activity associated with gender and social status call for more studies on cultural perspective for health behaviors in cross-cultural contexts. & 2016 Elsevier Ltd. All rights reserved.

Keywords: Physical inactivity Perceived neighborhood quality Subjective social status Socio-economic position

1. Introduction In Asia, urbanization generally carries with it a higher income, shift in work style, and change of residential location. People are exchanging more labor-intensive occupations in rural industries for sedentary jobs in urban areas. Consequently, such a transition has led to increasing physical inactivity. By definition, physical inactivity refers to not having 3 or more days of vigorous activity lasting at least 20 min that makes people sweat or breathe heavier than usual, on a weekly basis (IPAQ Research Committee, 2005). The term is in contrast to sedentary behaviors that involve consistently sitting and low levels of energy expenditure (IPAQ Research Committee, 2005; Guthold et al., 2008). Studies revealed 17–31% prevalence of physical inactivity globally from 2008 to 2012 (WHO, 2009; Hallal et al., 2012). Currently, one-third of adult populations are physically inactive (Pratt et al., 2015). There is considerable concern about physical inactivity and its association with chronic diseases (WHO, 2009) and adverse health n

Corresponding author. E-mail address: [email protected] (D.-R. Chen).

http://dx.doi.org/10.1016/j.healthplace.2016.06.001 1353-8292/& 2016 Elsevier Ltd. All rights reserved.

outcomes, including worldwide mortality (WHO, 2009; Bauman et al., 2012; Pratt et al., 2015). Factors associated with physical inactivity have been examined extensively in developed countries (Kahlmeier et al., 2015; Lee et al., 2012). However, very little information is known in developing countries (Guthold et al., 2008). Therefore, in order to prevent chronic diseases resulting from physical inactivity, particularly in East Asian countries, there is need to fill this knowledge gap. The literature has shown that people at the top of the socioeconomic strata report higher levels of leisure time or moderate to vigorous intense physical activity than those at the bottom (Gidlow et al., 2006; McNeill et al., 2006). Yet, research on the social patterning of physical inactivity in Asian contexts is relatively scarce (Jurj et al., 2007). The association between socioeconomic position (SEP) and physical activity is related to the developmental level of a country. In developing countries, a labor-active lifestyle that requires physical activities at work and commute to work is considered a necessity. For these nations, adopting what is characterized as a healthier Western lifestyle is thought to be a privilege that only the affluent can afford (Gidlow et al., 2006).

2

D.-R. Chen, Y.-C. Lin / Health & Place 41 (2016) 1–10

In China, the wealthy and those with socioeconomic privilege were twice as likely to be physically active in terms of leisure time activity (OR ¼2.1 and 2.7 for men and women, respectively), but less likely to be physically active in work (OR ¼0.1 and 0.1 for men and women, respectively) and commuting-related activities (OR ¼0.4 and 0.5 for men and women, respectively) compared with the less affluent (Bauman et al., 2011). In South Korea, the rate of physical activity is significantly associated with income levels (Kim et al., 2014). While several psychosocial, cognitive, and emotional factors on the individual level have been used to explain why some people are active and others are inactive (Sherwood and Jeffery, 2000), physical activity has been defined as a behavior that is shaped by one's social environment (Li et al., 2005). It is difficult to assess or change the low prevalence of physical activity in Asian countries without consideration of social and cultural norms, neighborhood resources for engaging in physical activity, and environmental constraints such as unsafe neighborhoods. 1.1. Social identity Social identity, including SEP and subjective social status (SSS), is defined as the mechanism of classifying oneself or others into various social categories (Ashforth and Mael, 1989). Research has shown that SSS is consistently and often strongly related to cholesterol levels, obesity, waist circumference, waist-to-hip ratio, overweight status, hypertension, diabetes, and depression (Demakakos et al., 2008; Operario et al., 2004). Perceptions of one's standing within a social hierarchy reflects how one evaluates oneself in terms of group-defining attributes; and, this identity can produce motivation, sense of status and control, preferences, and opportunities that influence health and health behaviors (Shilling, 1991). Perception of one's own low social status may generate status anxiety and consequent psychological distress, which may help explain the relationship between socioeconomic status (SES) and health (Adler et al., 2000; Demakakos et al., 2008). 1.2. Perceived neighborhood quality Several studies have indicated that perceived (subjective) neighborhood characteristics have a higher impact than actual (objective) ones on health outcomes and health behavior (Ross and Mirowsky, 2001). Subjective assessments of neighborhood characteristics substantially reflect an individual's perspectives and appraisal of the quality of his/her neighborhood (Cho et al., 2005). Therefore, subjective assessments of neighborhood quality are used as measures in this study. Adopted from Macintyre et al. (2002) and McNeill et al. (2006), this study identifies four aspects of physical and social environments. First, features of the physical environment of the neighborhood, such as poor air quality, may inhibit physical activity (An and Xiang, 2015; Li et al., 2015; Humphreys et al., 2014). Second, the perceived lack of available and accessible health and municipal services such as recreational facilities can limit opportunities for physical activity (Heath et al., 2012; Sallis et al., 2012). Third, the social environment of the neighborhood such as social support and cultural norms for encouraging or enforcing patterns of social control may place constraints on individual choice (Bauman et al., 2012; Beets et al., 2010). Fourth, the living conditions of neighborhoods can shape where people spend their leisure time and where they exercise (Casper and Harrolle, 2013; Santos et al., 2016). For instance, daily exposure to an unsafe environment can discourage physical activity (Timperio et al., 2015).

1.3. Cross-cultural comparison In comparing Taiwan with China and South Korea, key cultural and societal assumptions must be addressed. First, the social and economic aspects of Taiwan and China are clearly different due in part to divergences in political ideology, which have led to communist practice in Mainland China and democratic policies in Taiwan. Both, however, are still considered Chinese societies that share long-standing, traditional Confucian philosophies and social norms that affect all aspects of life. The factors affecting health behaviors may be similar between Taiwan and China, and detailed comparisons between the two nations will further deepen understanding of how cultural values serve as an infrastructure allowing for similarities and differences. On the contrary, the process of social and economic development in South Korea in recent years is also similar to that of Taiwan. For example, South Korea and Taiwan are both leading nations in the information technology industry and both have similar levels of gross domestic product (GDP). The South Korean GDP in 2014 was estimated at US $26,991 while in Taiwan the GDP in 2014 was estimated at US $23,582 (IMF, 2015). Both countries have high life expectancy, low infant mortality rate, an average of 12 years of education, and 4 98% literacy rate among citizens above age 15. Similar social development brought similar trends of urbanization and globalization. If such social development serves as crucial infrastructure in these two societies, similar socioenvironmental factors affecting health behaviors can be identified. In light of the above discussion, this study aims to address the extent to which objective SEPs (such as education level and occupation-based social class), SSS, and perceived neighborhood quality are associated with physical inactivity across Taiwan, China, and South Korea, while controlling for overweight status, chronic disease, and psychological distress.

2. Methods 2.1. Survey and sample This study uses data from the East Asian Social Survey (EASS) (Health Module) collected from China and South Korea in 2010 and from Taiwan in 2011 (Noriko et al., 2010). These data consist of a common module, set into a general social survey (GSS)-type questionnaire, which is a nationally representative sample survey from each of the three countries. Samples were selected by multistage stratified probability sampling. Valid response rates were 71.99%, 63%, and 49.7% in China, South Korea, and Taiwan, respectively. Details of EASS data are described at the EASS website (http://eass.info). In-person interviews were conducted to collect the data in each of the study countries. Because there is no weighting conducted in South Korea, and only urban residents were selected for the analysis, no weighting scheme was employed in this study. 2.2. Eligibility The data were derived from individuals aged Z18 years, with 2199, 3866, and 1576 valid samples from Taiwan, China, and South Korea, respectively (Chang et al., 2012). Because of the distinct urban–rural differences in China, samples from rural areas in these three countries were excluded to avoid possible bias (Treiman, 2012). Hence, 5222 adult urban residents aged Z18 years were included in the analysis, with the largest group being from China (n ¼2268), followed by Taiwan (n ¼1659) and South Korea (n ¼1295). However, the study sample had missing responses for some of the variables studied. Education level had the largest

D.-R. Chen, Y.-C. Lin / Health & Place 41 (2016) 1–10

number of missing data (214, 4.1% of the total), followed by physical inactivity (0.8%). We decided to exclude the missing data in the multivariate analyses, because they constituted o 5% in the total sample. 2.3. Human subject approval We used the public version of the data, which is already stripped of personal identifiers before being released for secondary data analyses. In addition, our analyses were conducted in an aggregate manner, thus imposing no potential physical or health risks to the participants. Therefore, no approval from the internal review board was sought.

3

activity: “How often do you engage in vigorous exercise for at least 20 min per day in a way that makes you sweat or breath heavier than usual?” The response options ranged from “daily,” “several times a week,” “several times a month”, “several times a year or less often,” to “never.” In line with the definition of inactive (category 1) in the Short Form International Physical Activity Questionnaire (IPAQ Research Committee, 2005), respondents who answered “several times a month”, “several times a year or less often,” or “never” were recorded as “1”, matching into the physical inactivity level (category 1). The remaining participants with physical activity options as “daily” and “several times a week” were recorded as “0″ to indicate that they had engaged in physical activities including minimally active (category 2) and HEPA (Health Enhancing Physical Activity, category 3).

3. Measures

3.5. Perceived neighborhood quality

3.1. Socioeconomic position

Respondents were asked about the extent of agreement on physical and social dimensions of their perceived neighborhood environment. Neighborhood was defined in this study as an area of 1 square km/1.60 square miles, the equivalent of walking approximately 15 min around the resident's home. The physical dimension of neighborhood quality included two questions related to physical activity. The first question was “Is the neighborhood suitable for walking and jogging?” using a five-point Likert measurement ranging from “strongly agree”, “agree”, “neither agree nor disagree”, “disagree”, and “strongly disagree”. The second question “How severe is the air pollution in the area of your local residence?” was measured by a five-point Likert scale of “very severe”, “severe”, “somewhat severe”, “not so severe”, and “not severe at all”. For determining social dimensions of perceived neighborhood environment, three questions using the aforementioned five-point Likert scale of agreement were asked: “Does your neighborhood have adequate public facilities? (e.g., community center, library, and park)”, “Is your neighborhood safe?”, and “Are your neighbors willing to provide assistance when you are in need?”.

In this study, SEP variables used included years of education and occupation-based social class. 3.2. Occupation-based social class Occupations were assessed by the International Standard Classification of Occupations (ISCO-88) (International Labor Office, 1990). The ISCO-88 system was used to facilitate the international use and comparison of occupational information. The basic criteria used to define the system of classes are “skill level” and “skill specialization” required to perform competently the tasks and duties of occupations. It classifies occupations into the following categories: 10: bourgeois, congressmen, high-level managers; 9: professionals; 8: managers and technicians; 7: clerks; 6: service workers; 5: agricultural workers; 4: forestry, animal husbandry, and fishery workers; 3: skilled workers; 2: machine operators; 1: unskilled laborers; and none. Bourgeois, congressmen, high-level managers, and professionals are considered upper social class; managers and technicians, clerks, and service workers are considered middle social class; agricultural, forestry, animal husbandry and fishery workers, skilled workers, machine operators, and unskilled laborers are considered lower social class; and unemployed or having no occupation are classified as none. The spouse's class was used for housewives or househusbands. 3.3. Subjective social status Respondents rated their SSS using the MacArthur Scale of Subjective Social Status, which assesses perception of social status as they relate to society and which has been found to have adequate test–retest reliability (r ¼0.62) (Adler et al., 2000; Demakakos et al., 2008; Operario et al., 2004). The MacArthur Scale of Subjective Social Status was developed to capture each individual's sense of place on the social ladder, which takes into account standing on multiple dimensions of SES and social position. The SSS ladder provides a summative measure of social status. In an easy pictorial format, the survey presents a “social ladder” and asks individuals to place an “X” on the rung on which they feel they stand. A Likert scale numbered 1–10 with participants rating themselves 1 as lowest and 10 highest was used. Singh-Manoux et al. (2005) suggested that SSS provides a better assessment of a person's future prospects, opportunities, and resources than objective SES. 3.4. Physical inactivity Respondents were asked a single-item measure of physical

4. Control variables Additional variables including gender, age, marital status, chronic disease status (whether one self-reports diabetes, hypertension, heart diseases, respiratory diseases, or other), and overweight status (i.e., body mass index4 25) were controlled in the analysis. Psychological distress was measured by one item in the Short Form Health-related Quality of Life questionnaire (SF12). Respondents were asked if they felt downhearted and depressed during the last 4 weeks with possible responses of “all the time”, “most of the time”, “some of the time”, “a little of the time”, and “none of the time”. Respondents answering with “all of the time” and “most of the time” were coded as “1”, while others were coded as “0”.

5. Statistical analysis Descriptive statistics were used to summarize the sample demographic characteristics and physical inactivity across all three countries. Nested logistic regressions were then applied to discover the association of respondents’ social identity factors and perceived neighborhood quality in relation to physical inactivity across all three countries. Nested regression model comparison test was conducted using the likelihood ratio (LR) test (Hosmer et al., 2013) in STATA 14. The LR test was performed by estimating two models and comparing the fit of one model to the fit of the

4

D.-R. Chen, Y.-C. Lin / Health & Place 41 (2016) 1–10

6. Results

other. The LR test does this by comparing the log likelihoods of two models, and if this difference is statistically significant, then the model with more variables is considered to fit the data significantly better than the restrictive model. Model 1 presented social identity factors, including education, occupation-based social class, and SSS, while controlling for psychological distress, chronic disease, overweight status, marital status, age, and country. In order to make country-level comparisons using logistic regressions (Mood, 2010), multiple interaction terms with country were examined. All interaction terms between social identity factors and country were examined and only significant interaction terms were included. It revealed that gender interacted with education levels and occupation-based social class interacted with country. Both interaction terms were then included in Model 2 with all the variables in Model 1. Perceived neighborhood quality variables were then added to Model 3 with all variables in Model 2. Perceived neighborhood suitability for walking and jogging, neighborhood social assistance, and air quality were found interacted with country, and then were included with all other variables in Model 4 (final Model). Each interaction term was examined by using the LR test to examine the improvement of model fit with interaction term over the fit of the model without interaction term (Hosmer et al., 2013).

6.1. Sample demographics The study sample across three countries was diverse with respect to age, education background, perceived neighborhood environment, occupation-based social class, SSS, and physical inactivity. Table 1 presents the descriptive findings on sample characteristics. The majority of participants in the three countries were married. Their average age ranged from 42 to 49 years. The gender makeup for the total sample was 52% women and 48% men. China had the lowest average years of education (10.18 years), followed by Taiwan (12.43 years) and South Korea (12.78 years). The mean values of the participants’ SSS were roughly in the middle: 4.16, 5.0, and 4.67 for China, Taiwan, and South Korea, respectively. 6.2. Nested logistic regressions Table 2 presents the adjusted log adds of social identity factors including education, occupation-based social class, and subjective status, and perceived neighborhood quality on the likelihood of being physically inactive with pooled data while controlling for psychological distress, chronic disease, overweight status, marital status, age, and country. As shown in Table 2, from Model 1 to Model 4, the associations of social identity factors such as education, marital status, and SSS with the likelihood of physical inactivity were consistent across all three countries. Women were 41% more likely to be physically

Table 1 The characteristics of study sample.

Perceived neighborhoood quality (Percent of strongly agree/agree) Suitability for exercise Adequate public facilities Neighborhood safety Willing to provide assistance Air pollution Gender Women Men Age (means, SD) Year of education Marital status Single Married Divorced/seperated Widow Occupation-base social class Bourgeouis, congressmen, high-level managers, professionals (upper class) Low-level managers, Technicians, Clerks (middle class) Skilled and non-skilled workers (lower class) Unemployed/no occupation Subjective social status (1–10) Physical activity Daily Several times a week Several times a month Several times a year or less often Never Chronic diseases (Yes) Overweight status (BMI Z 25) Psychological distress (Yes) n

p o 0.05. nn

po 0.01. p o0.001.

nnn

Total (N¼ 5222)

China (N ¼2268)

Taiwan (N¼1659)

South Korea (N¼1295)

N

%

N

%

N

%

N

%

Sig

3859 3250 3483 3291 1851

74.03% 62.34% 66.88% 63.70% 35.49%

1361 905 1536 1593 858

60.19% 40.01% 67.90% 70.55% 37.88%

1493 1564 1238 1300 491

90.05% 94.33% 74.85% 80.50% 29.63%

1005 781 709 398 502

77.67% 60.40% 54.88% 30.78% 38.79%

nnn

2717 2505 43.08 11.54

52.03% 47.97% 14.37 4.21

1077 1191 44.06 10.18

47.49% 52.51% 14.06 4.15

806 853 42.73 12.43

48.58% 51.42% 15.08 4.27

622 673 41.83 12.78

48.03% 51.97% 13.85 3.46

1223 3540 228 218

23.48% 67.96% 4.38% 4.19%

322 1762 86 93

14.23% 77.86% 3.80% 4.11%

558 940 82 74

33.74% 56.83% 4.96% 4.47%

343 838 60 51

26.55% 64.86% 4.64% 3.95%

510

9.77%

217

9.57%

189

11.39%

104

8.03%

1505 821 2386 4.55

28.82% 15.72% 45.69% 1.65

476 310 1265 4.16

20.99% 13.67% 55.78% 1.64

608 304 558 5.00

36.65% 18.32% 33.63% 1.54

421 207 563 4.67

32.51% 15.98% 43.47% 1.65

988 1321 760 689 1421 1452 1339 405

19.08% 25.51% 14.67% 13.30% 27.44% 27.90% 25.90% 7.76%

438 402 244 286 868 650 557 204

19.57% 17.96% 10.90% 12.78% 38.78% 28.66% 24.59% 9.44%

345 490 272 287 261 471 505 71

20.85% 29.61% 16.44% 17.34% 15.77% 28.68% 31.39% 4.30%

205 429 244 116 292 331 277 120

15.94% 33.36% 18.97% 9.02% 22.71% 25.58% 21.39% 9.17%

nnn nnn nnn nnn

nnn nnn nnn

nnn

nnn nnn

nn nnn nnn

D.-R. Chen, Y.-C. Lin / Health & Place 41 (2016) 1–10

5

Table 2 Logistic regression of social identity and perceived urban neighborhood quality on physical inactivity (log odds and 95%CI). Model 1

Model 2

Log odds

95% CI

1.393

0.979, 1.807

Model 3

Log odds

95% CI

nnn

1.392

0.928, 1.856

 0.023  0.029,  0.016  0.189  0.338, 0.041 0.175 0.113, 0.238 0.022  0.118, 0.163

Model 4

Log odds

95% CI

nnn

3.123

2.477, 3.770

nnn

 0.021  0.198 0.163 0.031

 0.028,  0.015  0.348,  0.047 0.100, 0.227  0.111, 0.173

Log odds

95% CI

nnn

4.22

3.375, 5.063

nnn

nnn

 0.021  0.194 0.168 0.032

 0.027, 0.014  0.345  0.043 0.104, 0.232  0.110, 0.175

nnn

nnn

Intercept Controls Age Chronic disease Psychological distress Overweight

 0.022  0.028,  0.016  0.195  0.343,  0.046 0.178 0.116, 0.241  0.0003  0.140, 0.139

nnn

Country (ref: China) Taiwan South Korea

 0.266  0.332

 0.417, 0.115  0.489,  0.175

nn nnn

0.015  0.021

 0.394, 0.425  0.461, 0.419

0.128 0.005

 0.291, 0.547  0.446. 0.455

 1.92  1.965

 3.078, 0.761  3.067,  0.862

0.343

0.219, 0.467

nnn

 0.018

 0.244, 0.208

0.007

 0.222, 0.235

0.009

 0.220, 0.239

 0.357

 0.522,  0.191

nnn

 0.547

 0.782,  0.313

nnn

 0.52

 0.758,  0.283

nnn

 0.526

 0.765,  0.287

nnn

 0.435  0.538

 0.640,  0.229  0.743,  0.333

nnn

 0.623  0.900

 0.898,  0.347  1.164,  0.637

nnn

 0.613  0.867

 0.892,  0.334  1.133, 0.600

nnn

 0.608  0.859

 0.888,  0.328  1.127  0.591

nnn

0.380

0.068, 0.693

n

0.372

0.056, 0.687

n

0.376

0.059, 0.694

n

0.388 0.719  0.104

0.015, 0.760 0.397, 1.041  0.143,  0.064

n

0.391 0.015, 0.767 0.756 0.431, 1.081  0.094  0.134,  0.054

n

0.397 0.193, 0.775 0.738 0.412, 1.064  0.097  0.138,  0.057

n

0.331 0.412 0.363

0.138, 0.524 0.074, 0.751  0.044, 0.770

nn

0.364 0.349 0.401

0.168, 0.560 0.007, 0.690  0.011, 0.813

nnn

0.364 0.366 0.402

0.168, 0.591 0.226, 0.709  0.009, 0.814

nnn

0.147 0.563 0.012

 0.177, 0.471 0.207, 0.919  0.303, 0.328

0.148 0.585 0.009

 0.180, 0.475 0.224, 0.946  0.310, 0.329

0.147 0.606 0.005

 0.184, 0.479 0.240, 0.972  0.319, 0.329

 0.141  0.267  0.522  0.594  0.359  0.177

 0.614, 0.333  0.767, 0.234  1.040  0.004  1.143,  0.046  0.827, 0.110  0.703, 0.349

 0.111  0.275  0.531  0.616  0.359  0.133

 0.590, 0.367  0.783, 0.233  1.055,  0.007  1.173,  0.059  0.833, 0.114  0.668, 0.402

 0.115  0.270  0.550  0.633  0.346  0.140

 0.595, 0.364  0.780, 0.240  1.076,  0.236  1.193,  0.737  0.821, 0.129  0.676, 0.396

 0.209

 0.272,  0.146

nnn

 0.300

 0.385,  0.214

nnn

 0.133  0.104  0.064  0.027

 0.212,  0.054  0.171, 0.036  0.127,  0.0004  0.086,  0.033

nnn

 0.228  0.229  0.059  0.032

 0.346,  0.111  0.346,  0.112  0.123, 0.004  0.091, 0.028

nnn

0.255

0.101, 0.410

nnn

0.138

0.006, 0.269

n

0.087 0.261 0.204 0.162 26.28 (6)

 0.099, 0.273 0.070, 0.453 0.037, 0.370 0.005, 0.319 p ¼0.0002

Social Identity Variables Gender (ref: Men) Women Education (ref: no formal education) Elementary/Junior high school High school College and above

n nnn

nnn

Interaction with Women Women  Elementary/Junior high school Women  High school Women  College and above Subjective Social Status (1–10)

 0.106

 0.145,  0.067

nnn

Marital Status (ref: Single) Married Divorced/seperated Widow

0.296 0.381 0.284

0.105, 0.486 0.044, 0.719  0.122, 0.690

nn

Objective socio-economic position (ref: upper class) Middle class 0.041 Lower class 0.246 No occupation  0.184

 0.165, 0.247 0.007, 0.486  0.402, 0.033

Interaction with Country Middle class  Taiwan Middle  South Korea Lower class  Taiwan Lower class  South Korea No occupation  Taiwan No occupation  South Korea

n

n

Perceived Neighborhood Quality Suitabitity for walking & jogging Air quality Social Assistance Community safety Access to public facilities Interaction with Country Suitabitity for walking & jogging  Taiwan Suitabitity for walking & jogging  South Korea Air quality  Taiwan Air quality  South Korea Social Assistance  Taiwan Social Assistance  South Korea The Likelihood Ratio (LR) Test (Chi-square, df)

302.61 (17)

p o0.00001

a: Model 1 was compared to null model (no predictors). n

po 0.05. p o0.01. nnn p o0.001. nn

29.50 (9)

p ¼ 0.0005

n nnn

nnn

nnn nnn

n

nn

n n

93.77 (5)

p o 0.00001

nn nnn

nnn

nnn nnn

n

nn

n n

nn n

n nnn

nnn

nnn

nnn nnn

n

nnn

n n

nnn

nn n n

6

D.-R. Chen, Y.-C. Lin / Health & Place 41 (2016) 1–10

inactive than men across all three countries (log odds ¼0.343, OR (odds ratio) ¼1.41 in Model 1). Furthermore, the interaction between gender and education was found statistically significant. The gradient effect of education on physical inactivity was particularly salient for men. Men with education level of college or above college were notably less likely to be physically inactive than those with no formal education (log odds ¼ 0.538 to 0.859 from Model 1 to Model 4), followed by men with high school diploma (log odds ¼  0.435 to  0.608 from Model 1 to Model 4), and men with elementary or junior high school diploma (log odds ¼  0.357 to 0.526 from Model 1 to Model 4). Most importantly, gender difference widened as educational levels increased. The gender difference in the likelihood of physical inactivity was found particularly evident among individuals with college education or above (gender difference in log odds is from 0.719 to 0.738 from Model 2 to Model 4), followed by those with high school diploma (log odds is from 0.388 to 0.397) and those with elementary/junior high school (log odds is from 0.38 to 0.376). The results suggested that men and women may have very different behavioral norm toward physical inactivity, and this finding warrants more research on the cultural–behavioral understanding of health behaviors between genders (see Model 2 to Model 4). Individuals with one-unit increase of the sense of social status were about 13% less likely to be physically inactive (log odds is 0.097 (OR ¼0.87) in Model 4 (final model). The associations of occupation-based social class with physical inactivity were different across all three countries. Indicated in Model 4 (final model), as compared to upper social class, lower class was about 83% more likely to be physically inactive in China (log odds is 0.606 (OR¼ 1.83); yet, the likelihood of physical inactivity decreased to 6% more likely for lower class for Taiwanese (log odds is 0.056 (  0.55þ 0.606 ¼0.056), OR¼ 1.06), yet, on the contrary, it was about 3% less likely for lower class to be physically inactive in South Korea (the log odds is 0.056 (  0.633þ 0.606 ¼ 0.027), OR ¼0.97). It is suggested that the difference in the likelihood of physical inactivity between upper and lower classes was much larger in China than those in Taiwan and South Korea. In Model 3, five perceived neighborhood quality variables were examined with the variables in Model 2. The LR test showed that adding five neighborhood quality variables in Model 3 significantly improved the model fit over Model 2 (Chi-square¼93.77, df ¼5, p ¼0.0002). The result indicated that perceived neighborhood quality such as provision of social assistance, suitability for walking and jogging, community safety, and air quality significantly decreased the likelihood of physical inactivity across all three countries. In Model 4, the interaction effects between each neighborhood quality variables and country were examined. Using the LR test, adding each of the three perceived neighborhood quality interaction terms was found to significantly improve model fit: suitability for walkability and jogging (Chi-square ¼ 13.44, p¼0.001), air quality (Chi-square¼5.87, p¼0.05), and social assistance (Chi-square ¼8.08, p¼0.02). Three interaction terms were then included in Model 4 with all variables in Model 3. The LR test showed that adding three interaction terms in Model 4 significantly improved the model fit over Model 3 (Chi-square ¼26.64, df¼6, p¼ 0.0002). Significant country-level differences in the extent of the association between perceived neighborhood quality and physical inactivity were found in Model 4. First, the impact of perceived suitability for walking and jogging on physical inactivity was significantly different between Chinese and Taiwanese, and between Chinese and South Koreans. The likelihood of reducing physical inactivity by one-unit increase of perceived suitability for walking and jogging (a five-point Likert scale) was highest for Chinese (log odds¼ 0.30, OR¼0.74) than for South Koreans (log odds¼  0.039 ( 0.30þ0.138¼ 0.162), OR ¼0.85) and Taiwanese (log odds¼  0.045( 0.30þ0.255), OR¼0.96). Second,

the association of perceived air quality with physical inactivity was significantly different only between Chinese and South Koreans. For Chinese, one-unit increase of perceived air quality could reduce about 20% of the likelihood of physical inactivity (log odds¼  0.228, OR¼0.80), and yet it only had minimal impact on South Koreans (log odds¼0.033 ( 0.228þ 0.261¼ 0.033, OR¼ 1.03). Third, the association of perceived neighborhood social assistance with physical inactivity was significantly different between Chinese and Taiwanese (p¼ 0.02), and between Chinese and South Koreans (p¼0.04). For Chinese, one-unit increase of perceived social assistance from neighbors significantly reduced 20% of the likelihood of physical inactivity (log odds¼  0.229, OR¼0.80), yet the impact diminished for Taiwanese (log odds¼  0.025 (0.229þ0.204¼  0.025), OR¼ 0.98) and South Koreans (log odds¼  0.063 ( 0.229þ0.162¼  0.067), OR ¼0.94). Country differences in the likelihood of physical inactivity were significant after all confounding variables were controlled in Model 4. South Koreans were 86% less likely (log odds ¼  1.965, OR¼0.14) and Taiwanese were 85% less likely (log odds ¼  1.92, OR¼0.15) to be physically inactive than Chinese. This result revealed that South Korean and Taiwanese were more physically active than Chinese.

7. Discussion Increasing research has confirmed that physical activity is determined by individual, social, and environmental factors (McNeill et al., 2006). However, the majority of these studies have been conducted in Western countries. This study aims to fill this knowledge gap in East Asian countries, by evaluating the impact of social identity and perceived environmental factors on physical inactivity in China, Taiwan, and South Korea. Several interesting results have been found. First, the South Koreans and Taiwanese tend to be more physically active than the Chinese. In China, about 51% of the adult urban residents reported doing exercise only several times a year or less often or never, yet 33% of Taiwanese and 32% of South Korean gave this response. The prevalence of physical inactivity seems to correspond with the level of socioeconomic development in each country. South Korea has the highest level of GDP, followed by Taiwan and China. Fig. 1 showed the adjusted predicted probability of physical inactivity across three countries from Model 4. As shown, Chinese reported highest average probability of physical inactivity after controlling for all confounding variables (mean¼0.61, SD ¼0.15), followed by South Koreans (mean ¼0.50, SD ¼0.13) and Taiwanese (mean¼0.49, SD ¼0.11). South Koreans and Taiwanese reported similar average predicted probability of physical inactivity, yet the variance is larger in South Korea than that in Taiwan. Several studies have indicated that the dominant type of daily energy expenditure would be different at different stages of economic development. Higher levels of leisure time or moderate to vigorous physical activity are typically found in nations with higher socioeconomic levels than those with lower socioeconomic levels (Gidlow et al., 2006; Kim et al., 2004). In comparison, the International Prevalence Study on Physical Activity (IPPA) indicated that China had the highest activity levels (57.7% activity prevalence rate) and Taiwan had the lowest activity levels ( 42.3% activity prevalence rate) (Bauman et al., 2009), while South Korea was not included in the study. This discrepancy may result from the fact that, in China, a higher proportion of physical activity is defined as being related to work or transportation (Bauman et al., 2009, 2011). Second, with regard to the impact of perceived neighborhood quality, physical features (suitability for walking and jogging and air pollution) and social feature (help from neighbors) exerted

D.-R. Chen, Y.-C. Lin / Health & Place 41 (2016) 1–10

7

Fig. 1. Predicted probability of physical in activity.

notable influence in physical inactivity (see Table 2), and significant country differences were found. Desirable environmental factors have been shown to be protective against physical inactivity (US Department of Health and Human Services, 2009). Recent reviews of the literature (Ding et al., 2013; Sallis et al., 2012; Timperio et al., 2015; Van Dyck et al., 2013) have also shown positive associations between physical activity and neighborhood quality, such as access to facilities, safety, and aesthetics. Neighborhoods with high walkability are often associated with high residential density, accessibility of facilities, and safety. Residents of these neighborhoods have been more likely to report being physically active than others (Saelens et al., 2003). The impact of outdoor environment, including help from neighbors, air quality, and suitability for walking and jogging for physical activity, was especially notable in China. A well-connected and designed street network is often found in modern urbanized areas (Kim et al., 2012), such as those in South Korea and Taiwan. In comparison, areas with lack of relatively better spatial access and proximity to public space or lack of infrastructure including trails, sidewalks, bicycle lanes, and facilities may generate higher impact and this may have inhibited participation in physical activities in China (Kim et al., 2012; Sloan et al., 2013). In addition, as China suffers serious air pollution, it is not surprising that perceived air quality in neighborhoods seriously reduced the chance of physical activity in China, as compared to those in Taiwan and South Korea (Li et al., 2015). Highly developed countries are currently challenged by rapid urbanization, which results in higher population density, growing crime rates, resource limitation, and other aggregated consequences that could jeopardize urban safety, an important factor affecting individual health behavior (Li et al., 2012). In order to avoid areas perceived to be less safe, people may alter their behavior by constraining their physical activity (Foster and GilesCorti, 2008). Neighborhood safety thus plays a decisive role in

promoting physical activity and underlines a fundamental requirement for doing exercise (McNeill et al., 2006). However, the results from previous studies on the association between safety and physical activity are inconsistent. Some studies indicated that neighborhood safety is positively related to being physically active (Rind and Jones, 2014; Timperio et al., 2015). On the contrary, other studies have shown the reverse (Hu et al., 2013; Van Dyck et al., 2013). In this study, we found that urban safety was unrelated to physical activity in the final model that had controlled for country differences. Most of the urban neighborhoods are generally safe in East Asian countries and may generate no important correlation with physical activity. Third, only a few Asian studies (Bauman et al., 2009, 2011; Chen et al., 2011; Frerichs et al., 2014; Hallal et al., 2012) have investigated the association between SES and physical activity. One study found that in Taiwan, those with higher education levels and SES were positively associated with physical activity (Chen et al., 2011), and men tend to exercise more than women (Chen et al., 2011). In addition, men in China with higher income levels had higher odds of doing regular exercise (Lee et al., 2007). In this study, the findings showed clear gradient pattern in the relationship between education levels and physical inactivity. Men with college education or above tend to be more physically active than those with no formal education. The results are also consistent with studies conducted in the United States (Yi et al., 2015) and Germany (Finger et al., 2012). In addition, gender difference was particularly evident among study participants with college education or above, and this difference widened significantly with increase in educational levels. Women generally tended to be more physically inactive than men, and this is consistent with previous studies (Eyler et al., 2002; Haase et al., 2004). However, the odds of physical inactivity for women in Asia appear to be unique compared with that of women from other ethnicities (Sternfeld et al., 1999). The gender

8

D.-R. Chen, Y.-C. Lin / Health & Place 41 (2016) 1–10

gap in the likelihood of physical activity is strongest between men and women with highest education level, and this gender gap remains significant across different levels of education. It is suggested that the major perceived barrier of participating in vigorous–moderate physical activity is due to social expectation of gender roles in Asian countries (Babakus and Thompson, 2012). While lack of time, safe place, and fatigue is recognized as obstacles to leisure time physical activity for women of other ethnicities (Eyler et al., 2002; King et al., 2000), Asian women are expected to fulfill gender role expectations by being responsible for domestic and caretaking duties. Therefore, taking leisure time to do exercise may be seen as a waste of time or selfish act (Babakus and Thompson, 2012). In addition, Asian women are particularly prone to stay indoors, because a fair complexion is considered an important element in constructing female beauty (Frith et al., 2004). Compliance to culturally desirable ideal of a sedentary lifestyle may also expose women to physical inactivity. Effective intervention would be needed to enhance the levels of physical activity in leisure time for Asian women. This study shows the potential cultural constraints of Asian women and provides a unique perspective and valuable standpoint for future promotion. Fourth, a positive gradient was found between SEP and physical activity in studies conducted in the United States (Crespo et al., 2000), Australia (Chau et al., 2012), and Scotland (Popham and Mitchell, 2007). Our findings are consistent with previous studies conducted in Western societies showing that people with lower SEP tend to be more physically inactive. However, the difference between upper and lower social classes was higher in China, followed by Taiwan and South Korea. It revealed that the class-based inequality of physical inactivity decreased when the prevalence of physical activity increased in these Asian countries. As indicated in previous studies, blue-collar workers may compensate for their heavy physical labor by engaging in less exercise (Chau et al., 2012; Finger et al., 2012). It may highlight that there exists a conceptual gap in knowledge of the difference between healthy physical activity and labor activity and that healthy physical exercise during leisure time is still needed for this group. This finding also supports the fundamental cause theory of health inequalities proposed by Phelan and his associates (Phelan et al., 2010) that, when knowledge of disease prevention has diffused in society, the social gradient of health is likely to diminish. Recognizing this gap may be critical in effective health promotion programs and in reducing health inequalities across social classes. Finally, the measurement of SEP may not be sufficient to explore the mechanism behind physical activity in East Asian contexts. On the contrary, SSS, defined as a perceived social status in society after assessing one's social position relative to others, capturing cognitive perceptions of inequality (Garbarski, 2010) proved to be a significant predictor of physical inactivity for the three Asian countries. Specifically, one's self-perception may affect decisions on adopting certain health behaviors in order to set oneself apart from other social groups or to be associated with ingroup norms (Dressler, 1998; Pampel et al., 2010). For example, sedentary activities such as watching television may be stigmatized in higher-SES households (Pampel et al., 2010). The link between social subjective status and physical activity warrants future exploration of the psychosocial pathways of why people of low social status stay physically inactive (Ball and Crawford, 2010; Rind and Jones, 2014). The rate of time preference may also play a key role here. If people are less willing to trade current utility for potential future health benefits, then we expect they will engage in more sedentary leisure pursuits despite the future adverse consequences (Komlos et al., 2004). High-status groups often have positive time preference for the future. Therefore, it is more likely to observe health behaviors such as physical activity in high-status groups. On the contrary, low-status groups

may produce social stress and discouragement for prospects, and the resulting negative affect may leave people prone to physical inactivity (Adler et al., 2000; Operario et al., 2004). How SSS influences physical activity has not yet been described for Asian populations to date, making this study particularly important in light of recent trends of increased physical inactivity worldwide. This study also adds to the current literature showing that psychological stress predicts fewer tendencies to exercise (StultsKolehmainen and Sinha, 2014), and that married and divorced people are less likely to report physical activity (Ball and Crawford, 2010). Responsibilities at home or emotional strain because of marital dissolve may lead to reductions in physical activity (Allender et al., 2008; Ball and Crawford, 2010). Thus, this study shows that psychological stress contributing to physical inactivity is a cross-cultural phenomenon. Further studies on this subject could be useful in determining whether this trend continues in developing countries. Future health promotion interventions could target this trend through stress management training using exercise or through changing social norms by helping heads of households realize that exercise is a responsibility, which not only benefits one's health but also sets an example of good health behavior within the household. 7.1. Limitations Consistent findings across the three nations can be generalized across the Asian population. However, the design of this study is cross-sectional, hence causality cannot be determined. Furthermore, the physical activity measure was self-reported and was of limited precision for measuring specific activity levels. In terms of the measure of physical activity, although the criterion used in this study was adapted from the Short Form IPAQ designed to access physical activity across four domains, this single-item measure in this study is most similar to validated physical activity questionnaire items that survey moderate-to-vigorous activities beyond work and active transportation (Bauman et al., 2009). This single-item measure was also used in other studies (Milton et al., 2013; Wanner et al., 2013). Furthermore, this measure was found to be validated, and it even performed better than longer questionnaires. However, given the potential misclassification and low diagnosis capacity of physical activity levels (Zwolinsky et al., 2015), the concerns of utilizing single-item measure cannot be completely ignored. Future studies are encouraged to better design culturally tailored health interventions to promote physical activity.

References Adler, N.E., Epel, E.S., Castellazzo, G., Ickovics, J.R., 2000. Relationship of subjective and objective social status with psychological and physiological functioning: preliminary data in healthy, White women. Health Psychol. 19 (6), 586–592. http://dx.doi.org/10.1037/0278-6133.19.6.586. Allender, S., Hutchinson, L., Foster, C., 2008. Life-change events and participation in physical activity: a systematic review. Health Promot. Int. 23 (2), 160–172. http: //dx.doi.org/10.1093/heapro/dan012. An, R., Xiang, X., 2015. Ambient fine particulate matter air pollution and leisuretime physical inactivity among US adults. Public Health 129 (12), 1637–1644. http://dx.doi.org/10.1016/j.puhe.2015.07.017. Ashforth, B.E., Mael, F., 1989. Social identity theory and the organization. Acad. Manag. Rev. 14 (1), 20–39. http://dx.doi.org/10.5465/AMR.1989.4278999. Babakus, W.S., Thompson, J.L., 2012. Physical activity among South Asian women: a systematic, mixed-methods review. Int. J. Behav. Nutr. Phys. Act. 9, 150–168. Ball, K., Crawford, D., 2010. The role of socio-cultural factors in the obesity epidemic. Obes. Epidemiol. Aetiol. Public Health, 105–118. Bauman, A.E., Reis, R.S., Sallis, J.F., Wells, J.C., Loos, R.J.F., Martin, B.W., 2012. Correlates of physical activity: why are some people physically active and others not? Lancet 380 (9838), 258–271. http://dx.doi.org/10.1016/s0140-6736(12) 60735-1. Bauman, A.E., Ma, G., Cuevas, F., Omar, Z., Waqanivalu, T., Phongsavan, P., Keke, K.,

D.-R. Chen, Y.-C. Lin / Health & Place 41 (2016) 1–10

Bhushan, A., 2011. Cross-national comparisons of socioeconomic differences in the prevalence of leisure-time and occupational physical activity, and active commuting in six Asia-Pacific countries. J. Epidemiol. Community Health 65 (1), 35–43. http://dx.doi.org/10.1136/jech.2008.086710. Bauman, A.E., Bull, F., Chey, T., Craig, C.L., Ainsworth, B.E., Sallis, J.F., Bowles, H.R., Hagströmer, M., Sjöström, M., Pratt, M., 2009. International prevalence study on physical activity: results from 20 countries. Int. J. Behav. Nutr. Phys. Act. 6 (1), 21–32. http://dx.doi.org/10.1186/1479-5868-6-21. Beets, M.W., Cardinal, B.J., Alderman, B.L., 2010. Parental social support and the physical activity–related behaviors of youth: a review. Health Educ. Behav., 1–24. http://dx.doi.org/10.1177/1090198110363884. Casper, J.M., Harrolle, M.G., 2013. Perceptions of constraints to leisure time physical activity among Latinos in Wake County, North Carolina. Am. J. Health Promot. 27 (3), 139–142. http://dx.doi.org/10.4278/ajhp.110401-ARB-145. Chang, Y-H., Tu, S.H., Liao, P.S., 2012. Taiwan Social Change Survey (Round 6,Year 3). Institution of Sociology, Academia Sinica. Available from Survey Research Data Archive, Center for Survey Research, Research Center for Humanities and Social Sciences, Academia Sinica. Doi: 10.6141/TW-SRDA-C00223_1-1. Chau, J.Y., van der Ploeg, H.P., Merom, D., Chey, T., Bauman, A.E., 2012. Cross-sectional associations between occupational and leisure-time sitting, physical activity and obesity in working adults. Prev. Med. 54 (3), 195–200. http://dx.doi. org/10.1016/j.ypmed.2011.12.020. Chen, Y.J., Huang, Y.H., Lu, F.H., Wu, J.S., Lin, L.L., Chang, C.J., Yang, Y.C., 2011. The correlates of leisure time physical activity among an adults population from southern Taiwan. BMC Public Health 11 (1), 1–9. http://dx.doi.org/10.1186/ 1471-2458-11-427. Cho, Y., Park, G.-S., Echevarria-Cruz, S., 2005. Perceived neighborhood characteristics and the health of adult Koreans. Soc. Sci. Med. 60 (6), 1285–1297. http: //dx.doi.org/10.1016/j.socscimed.2004.06.054. Crespo, C.J., Smit, E., Andersen, R.E., Carter-Pokras, O., Ainsworth, B.E., 2000. Race/ ethnicity, social class and their relation to physical inactivity during leisure time: results from the Third National Health and Nutrition Examination Survey, 1988–1994. Am. J. Prev. Med. 18 (1), 46–53. http://dx.doi.org/10.1016/ S0749-3797(99)00105-1. Demakakos, P., Nazroo, J., Breeze, E., Marmot, M., 2008. Socioeconomic status and health: the role of subjective social status. Soc. Sci. Med. 67 (2), 330–340. http: //dx.doi.org/10.1016/j.socscimed.2008.03.038. Ding, D., Adams, M.A., Sallis, J.F., Norman, G.J., Hovell, M.F., Chambers, C.D., Hofstetter, C.R., Bowles, H.R., Hagströmer, M., Craig, C.L., 2013. Perceived neighborhood environment and physical activity in 11 countries: do associations differ by country. Int. J. Behav. Nutr. Phys. Act. 10 (1), 1–11. http://dx.doi.org/ 10.1186/1479-5868-10-57. Dressler, W.W., Bindon, J.R., Neggers, Y.H., 1998. Culture, socioeconomic status, and coronary heart disease risk factors in an African American community. J. Behav. Med. 21 (6), 527–544. Eyler, A.E., Wilcox, S., Matson-Koffman, D., Evenson, K.R., Sanderson, B., Thompson, J., Wilbur, J., Rohm-Young, D., 2002. Correlates of physical activity among women from diverse racial/ethnic groups. J. Women’s Health Gend.-based Med. 11 (3), 239–253. http://dx.doi.org/10.1089/152460902753668448. Finger, J.D., Tylleskär, T., Lampert, T., Mensink, G.B.M., 2012. Physical activity patterns and socioeconomic position: the German National Health Interview and Examination Survey 1998 (GNHIES98). BMC Public Health 12 (1), 1079. http: //dx.doi.org/10.1186/1471-2458-12-1079. Foster, S., Giles-Corti, B., 2008. The built environment, neighborhood crime and constrained physical activity: an exploration of inconsistent findings. Prev. Med. 47 (3), 241–251. http://dx.doi.org/10.1016/j.ypmed.2008.03.017. Frerichs, L., Huang, T.T.K., Chen, D.R., 2014. Associations of subjective social status with physical activity and body mass index across four asian countries. J. Obes. 2014, 1–11. http://dx.doi.org/10.1155/2014/710602. Frith, K.T., Cheng, H., Shaw, P., 2004. Race and beauty: a comparison of Asian and Western models in women’s magazine advertisements. Sex. Roles 50 (1–2), 53–61. Garbarski, D., 2010. Perceived social position and health: Is there a reciprocal relationship? Social. Sci. Med. 70 (5), 692–699. http://dx.doi.org/10.1016/j. socscimed.2009.11.007. Gidlow, C., Johnston, L.H., Crone, D., Ellis, N., James, D., 2006. A systematic review of the relationship between socio-economic position and physical activity. Health Educ. J. 65 (4), 338–367. http://dx.doi.org/10.1177/0017896906069378. Guthold, R., Ono, T., Strong, K.L., Chatterji, S., Morabia, A., 2008. Worldwide variability in physical inactivity: a 51-country survey. Am. J. Prev. Med. 34 (6), 486–494. http://dx.doi.org/10.1016/j.amepre.2008.02.013. Haase, A., Steptoe, A., Sallis, J.F., Wardle, J., 2004. Leisure-time physical activity in university students from 23 countries: associations with health beliefs, risk awareness, and national economic development. Prev. Med. 39 (1), 182–190. http://dx.doi.org/10.1006/pmed.1998.0470. Hallal, P.C., Andersen, L.B., Bull, F.C., Guthold, R., Haskell, W., Ekelund, U., Group, L.P. A.S.W., 2012. Global physical activity levels: surveillance progress, pitfalls, and prospects. Lancet 380 (9838), 247–257. http://dx.doi.org/10.1016/S0140-6736 (12)60646-1. Heath, G.W., Parra, D.C., Sarmiento, O.L., Andersen, L.B., Owen, N., Goenka, S., Montes, F., Brownson, R.C., Group, L.P.A.S.W., 2012. Evidence-based intervention in physical activity: lessons from around the world. Lancet 380 (9838), 272–281. http://dx.doi.org/10.1016/S0140-6736(12)60816-2. Hu, G.C., Chien, K.L., Hsieh, S.F., Chen, C.Y., Tsai, W.H., Su, T.C., 2013. Occupational versus leisure-time physical activity in reducing cardiovascular risks and mortality among ethnic Chinese adults in Taiwan. Asia-Pac. J. Public Health,

9

1–10. http://dx.doi.org/10.1177/1010539512471966. Humphreys, B.R., McLeod, L., Ruseski, J.E., 2014. Physical activity and health outcomes: evidence from Canada. Health Econ. 23 (1), 33–54. Hosmer Jr., D.W., Lemeshow, S., Sturdivant, R.X., 2013. Applied Logistic Regression, 3rd ed. International Labour Office, 1990. International Standard Classification of occupations: ISCO-88. International Monetary Fund (IMF), 2015, October 05. World Economic and Financial Surveys. World Economic Outlook Database. Retrieved from: 〈https:// www.imf.org/external/pubs/ft/weo/2012/01/weodata/weorept.aspx?pr. x¼ 72&pr.y¼ 6&sy ¼ 2010&ey¼ 2017&scsm ¼1&ssd ¼1&sort¼ country&ds ¼ .& br¼ 1&c ¼ 542%2C528&s¼ NGDPDPC&grp¼ 0&a ¼ 〉. IPAQ Research Committee, 2005. Guidelines for Data processing and Analysis of the International Physical Activity Questionnaire (IPAQ)–Short and Long Forms. Retrieved from: 〈www.ipaq.ki.se〉. Jurj, A.L., Wen, W., Gao, Y.-T., Matthews, C.E., Yang, G., Li, H.-L., Zheng, W., Shu, X.-O., 2007. Patterns and correlates of physical activity: a cross-sectional study in urban Chinese women. BMC Public Health 7 (1), 213–224. http://dx.doi.org/ 10.1186/1471–2458-7-213. Kahlmeier, S., Wijnhoven, T.M., Alpiger, P., Schweizer, C., Breda, J., Martin, B.W., 2015. National physical activity recommendations: systematic overview and analysis of the situation in European countries. BMC Public Health 15 (1), 133–147. Kim, S., Symons, M., Popkin, B.M., 2004. Contrasting socioeconomic profiles related to healthier lifestyles in China and the United States. Am. J. Epidemiol. 159 (2), 184–191. http://dx.doi.org/10.1093/aje/kwh006. Kim, S., Kwon, Y., Park, Y., 2014. Association between physical activity and healthrelated quality of life in Korean: the Korea National Health and Nutrition Examination Survey IV. Korean J. Fam. Med. 35 (3), 152. http://dx.doi.org/10.4082/ kjfm.2014.35.3.152. Kim, Y.S., Park, Y.S., Allegrante, J.P., Marks, R., Ok, H., Cho, K.O., Garber, C.E., 2012. Relationship between physical activity and general mental health. Prev. Med. 55 (5), 458–463. http://dx.doi.org/10.1016/j.ypmed.2012.08.021. King, A.C., Castro, C., Wilcox, S., Eyler, A.A., Sallis, J.F., Brownson, R.C., 2000. Personal and environmental factors associated with physical inactivity among different racial-ethnic groups of U.S. middle-aged and older-aged women. Health Psychol. 19 (4), 354–364. http://dx.doi.org/10.1037/0278-6133.19.4.354. Komlos, J., Smith, P.K., Bogin, B., 2004. Obesity and the rate of time preference: is there a connection? J. Biosoc. Sci. 36 (02), 209–219. Lee, I.-M., Shiroma, E.J., Lobelo, F., Puska, P., Blair, S.N., Katzmarzyk, P.T., Group, L.P.A. S.W., 2012. Effect of physical inactivity on major non-communicable diseases worldwide: an analysis of burden of disease and life expectancy. Lancet 380 (9838), 219–229. Lee, S.-A., Xu, W.H., Zheng, W.E.I., Li, H., Ya Ng, G., Xiang, Y.-B., Shu, X.-O., 2007. Physical activity patterns and their correlates among chinese men in shanghai. Med. Sci. Sport. Exerc. 39 (10), 1700–1707. http://dx.doi.org/10.1249/ mss.0b013e3181238a52. Li, F., Fisher, K.J., Brownson, R.C., Bosworth, M., 2005. Multilevel modelling of built environment characteristics related to neighbourhood walking activity in older adults. J. Epidemiol. Community Health 59 (7), 558–564. http://dx.doi.org/ 10.1136/jech.2004.028399. Li, F., Liu, Y., Lü, J., Liang, L., Harmer, P., 2015. Ambient air pollution in China poses a multifaceted health threat to outdoor physical activity. J. Epidemiol. Community Health 69 (3), 201–204. Li, J., Liu, Q., Sang, Y., 2012. Several issues about urbanization and urban safety. Procedia Eng. 43, 615–621. http://dx.doi.org/10.1016/j.proeng.2012.08.108. Macintyre, S., Ellaway, A., Cummins, S., 2002. Place effects on health: how can we conceptualise, operationalise and measure them? Social. Sci. Med. 55 (1), 125–139. http://dx.doi.org/10.1016/s0277-9536(01)00214-3. McNeill, L.H., Kreuter, M.W., Subramanian, S.V., 2006. Social environment and physical activity: a review of concepts and evidence. Social. Sci. Med. 63 (4), 1011–1022. http://dx.doi.org/10.1016/j.socscimed.2006.03.012. Milton, K., Clemes, S., Bull, F., 2013. Can a single question provide an accurate measure of physical activity? Br. J. Sports Med. 47 (1), 44–48. http://dx.doi.org/ 10.1136/bjsports-2011-090899. Mood, Carina, 2010. Logistic regression: why we cannot do what we think we can do, and what we can do about it. Eur. Sociol. Rev. 26 (1), 67–82. Noriko, Iwai (JGSS), Lulu, Li (CGSS), Sang-Wook, Kim (KGSS), and Ying-Hwa, Chang (TSCS), 2010. East Asian Social Survey (EASS), Cross-National Survey Data Sets: Health and Society in East Asia, 2010. ICPSR34608-v2. Ann Arbor, MI: Interuniversity Consortium for Political and Social Research/Seoul, Korea: EASSDA (Distributors), 2014-05-01. http://DOI.org/10.3886/ICPSR34608.v2. Operario, D., Adler, N.E., Williams, D.R., 2004. Subjective social status: reliability and predictive utility for global health. Psychol. Health 19 (2), 237–246. http: //dx.doi.org/10.1080/08870440310001638098. Pampel, F.C., Krueger, P.M., Denney, J.T., 2010. Socioeconomic Disparities in Health Behaviors. Annu. Rev. Sociol. 36 (1), 349–370. http://dx.doi.org/10.1146/annurev.soc.012809.102529. Phelan, J.C., Link, B.G., Tehranifar, P., 2010. Social conditions as fundamental causes of health inequalities theory, evidence, and policy implications. J. Health Social. Behav. 51 (1 suppl), S28–S40. Popham, F., Mitchell, R., 2007. Relation of employment status to socioeconomic position and physical activity types. Prev. Med. 45 (2–3), 182–188. http://dx.doi. org/10.1016/j.ypmed.2007.06.012. Pratt, M., Perez, L.G., Goenka, S., Brownson, R.C., Bauman, A., Sarmiento, O.L., Hallal, P.C., 2015. Can population levels of physical activity be increased? Global

10

D.-R. Chen, Y.-C. Lin / Health & Place 41 (2016) 1–10

evidence and experience. Progress. Cardiovasc. Dis. 57 (4), 356–367. http://dx. doi.org/10.1016/j.pcad.2014.09.002. Rind, E., Jones, A., 2014. Declining physical activity and the socio-cultural context of the geography of industrial restructuring: a novel conceptual framework. J. Phys. Act. Health 11 (4), 683–692. http://dx.doi.org/10.1123/jpah.2012-0173. Ross, C.E., Mirowsky, J., 2001. Neighborhood disadvantage, disorder, and health. J. Health Social. Behav. 42 (3), 258. http://dx.doi.org/10.2307/3090214. Saelens, B.E., Sallis, J.F., Black, J.B., Chen, D., 2003. Neighborhood-based differences in physical activity: an environment scale evaluation. Am. J. Public Health 93 (9), 1552–1558. http://dx.doi.org/10.2105/ajph.93.9.1552. Sallis, J.F., Floyd, M.F., Rodríguez, D.A., Saelens, B.E., 2012. Role of built environments in physical activity, obesity, and cardiovascular disease. Circulation 125 (5), 729–737. http://dx.doi.org/10.1161/CIRCULATIONAHA.110.969022. Santos, I., Ball, K., Crawford, D., Teixeira, P.J., 2016. Motivation and barriers for leisure-time physical activity in socioeconomically disadvantaged women. PLoS One 11 (1), 1–14. http://dx.doi.org/10.1371/journal.pone.0147735. Sherwood, N.E., Jeffery, R.W., 2000. The behavioral determinants of exercise: implications for physical activity interventions. Annu. Rev. Nutr. 20 (1), 21–44. http://dx.doi.org/10.1146/annurev.nutr.20.1.21. Shilling, C., 1991. Educating the body: Physical capital and the production of social inequalities. Sociology 25 (4), 653–672. http://dx.doi.org/10.1177/ 0038038591025004006. Singh-Manoux, A., Marmot, M.G., Adler, N.E., 2005. Does subjective social status predict health and change in health status better than objective status? Psychosom. Med. 67 (6), 855–861. http://dx.doi.org/10.1097/01. psy.0000188434.52941.a0. Sloan, R.A., Sawada, S.S., Girdano, D., Liu, Y.T., Biddle, S.J.H., Blair, S.N., 2013. Associations of sedentary behavior and physical activity with psychological distress: a cross-sectional study from Singapore. BMC Public Health 13 (1), 1–8. http: //dx.doi.org/10.1186/1471-2458-13-885. Sternfeld, B., Ainsworth, B.E., Quesenberry, C.P., 1999. Physical activity patterns in a diverse population of women. Prev. Med. 28 (3), 313–323.

Stults-Kolehmainen, M.A., Sinha, R., 2014. The effects of stress on physical activity and exercise. Sports Med. 44 (1), 81–121. http://dx.doi.org/10.1007/ s40279-013-0090-5. Timperio, A., Veitch, J., Carver, A., 2015. Safety in numbers: does perceived safety mediate associations between the neighborhood social environment and physical activity among women living in disadvantaged neighborhoods? Prev. Med. 74, 49–54. http://dx.doi.org/10.1016/j.ypmed.2015.02.012. Treiman, D.J., 2012. The “difference between heaven and earth”: urban–rural disparities in well-being in China. Res. Social. Strat. Mobil. 30 (1), 33–47. http://dx. doi.org/10.1016/j.rssm.2011.10.001. US Department of Health and Human Services, 2009. Healthy People 2010: Understanding and Improving Health. US Government Printing Office, Washington, DC. Van Dyck, D., Cerin, E., Conway, T.L., De Bourdeaudhuij, I., Owen, N., Kerr, J., Cardon, G., Frank, L.D., Saelens, B.E., Sallis, J.F., 2013. Perceived neighborhood environmental attributes associated with adults’ leisure-time physical activity: findings from Belgium, Australia and the USA. Health Place 19, 59–68. http://dx.doi.org/ 10.1016/j.healthplace.2012.09.017. Wanner, M., Probst-Hensch, N., Kriemler, S., Meier, F., Bauman, A., Martin, B.W., 2013. What physical activity surveillance needs: validity of a single-item questionnaire. Br. J. Sport. Med., 1–8. http://dx.doi.org/10.1136/ bjsports-2012-092122. WHO, 2009. Global Health Risks: Mortality and Burden of Disease Attributable to Selected Major Risks. World Health Organization, Switzerland. Yi, S.S., Roberts, C., Lightstone, A.S., Shih, M., Trinh-Shevrin, C., 2015. Disparities in meeting physical activity guidelines for Asian Americans in two metropolitan areas in the United States. Ann. Epidemiol. 25 (9), 656–660. http://dx.doi.org/ 10.1016/j.annepidem.2015.05.002. Zwolinsky, S., McKenna, J., Pringle, A., Widdop, P., Griffiths, C., 2015. Physical activity assessment for public health: efficacious use of the single-item measure. Public Health 129 (12), 1630–1636. http://dx.doi.org/10.1016/j. puhe.2015.07.015.