Elsevier

Health & Place

Volume 17, Issue 1, January 2011, Pages 78-85
Health & Place

BMI differences in 1st and 2nd generation immigrants of Asian and European origin to Australia

https://doi.org/10.1016/j.healthplace.2010.08.016Get rights and content

Abstract

We estimate assimilation of immigrants’ body mass index (BMI) to the host population of Australia over one generation, conducting separate analyses for immigrants from 7 regions of Europe and Asia. We use quantile regressions to allow for differing impact of generational status across 19 quantiles of BMI from under-weight to morbidly obese individuals. We find that 1st generation South European immigrants have higher, and South and East Asian immigrants have lower BMI than Australians, but have assimilated to the BMI of their hosts in the 2nd generation. There are no or only small BMI differences between Australians and 1st and 2nd generation immigrants from East Europe, North-West Europe, Middle East and Pacific regions. We conclude that both upward and downward assimilation in some immigrant groups is most likely caused by factors which can change over one generation (such as acculturation), and not factors which would take longer to change (such as genetics). Our results suggest that public health policies targeting the lifestyles of well educated Asian immigrants may be effective in preventing BMI increase in this subgroup.

Introduction

It is well documented that immigrants to high income countries tend to have lower body mass index (BMI) than the host population on arrival, but that those differences erode over time (Antecol and Bedard, 2006). This assimilation effect has been observed for different groups of immigrants to countries with high levels of immigration, including for Hispanic, Asian and other ethnic group immigrants to the USA (Kaushal, 2009, Sanchez-Vaznaugh et al., 2008, Bates et al., 2008, Antecol and Bedard, 2006, Kaplan et al., 2004, Gordon-Larsen et al., 2003), Canada (McDonald and Kennedy, 2005), various immigrant groups to Germany (Sander, 2008) and African immigrants to Australia (Renzaho et al., 2008).1 We extend this literature by estimating differences in BMI between 1st and 2nd generation immigrants from different regions of Asia and Europe and the host population in Australia. Estimates of the impact of generational status are generated across the whole range of BMI rather than at the average, using quantile regression methods.

Although there are a number of studies that measure and compare assimilation effects across different immigration groups, there is still a lack of evidence on why this phenomenon occurs. It seems that groups with higher education levels experience a smaller increase in BMI and/or assimilate at a slower rate than groups with lower education levels (Sanchez-Vaznaugh et al., 2008, Kaushal, 2009). Also, maintaining traditions of the home country (Renzaho et al., 2008) and greater integration into ethnic group networks (McDonald and Kennedy, 2005) seems to slow the process of adjustment to higher BMI levels. These latter two factors are sometimes summarized under the term acculturation, the process which describes assimilation to the lifestyles, behaviours and beliefs of the host country (Thomson and Hoffman-Goetz, 2009b, Thomson and Hoffman-Goetz, 2009a). A systematic review by Thomson and Hoffman-Goetz (2009b) found considerable variation in how acculturation is defined and measured, and the debate in the literature on this concept is ongoing (see the debate by Carter-Pokras and Bethune, 2009, Alegria, 2009, Thomson and Hoffman-Goetz, 2009a). However, Ryder et al. (2000) demonstrate how acculturation measures can work in distinguishing heritage values (culture of birth) from host values (predominant mainstream culture), and we will return to this issue in the discussion.

Apart from acculturation, there are some important long-term factors which are likely to impact on assimilation rates, but are difficult to measure. One of those factors is genetics. Genetic information is rarely collected in survey data, and can only be approximated by race or place of birth. The effects of several variables relating to place of birth and ethnicity have not yet been fully examined, although their relationship to the change in rates of obesity may be very important (Sanchez-Vaznaugh et al., 2008). In summary, analysing assimilation is hampered by the difficulty of defining and measuring important long and short term causal factors. For the purpose of this paper we assume that assimilation is influenced by observable factors, such as education and income, and by two types of unobserved factors; time-variant factors which can change over the course of one generation, and time-invariant factors which cannot. Examples of unobserved time-variant factors are acculturation and the social and physical environment; an important example of a time-invariant factor is genetics. Although we cannot observe those factors, we can still test their relative importance by analysing BMI values of immigrants and their children as they settle and adapt to life in their new country.

This paper estimates differences in BMI between the Australian host population and immigrants by generational status. We conduct separate analyses for 1st and 2nd generation immigrants and immigrants originating from 7 geographical regions (East Europe, South Europe, North-West Europe, East Asia, South Asia, the Middle East and Pacific). Australia is an interesting test case, as there is a long history of immigration, traditionally from Europe, but more recently from Asian countries (5.9% of our study population are Asian, this compares to the Asian–American population of 4.3%, and the Canadian–Asian population of around 5%).2 There are a number of studies investigating differences in BMI by generational status for Hispanic and Asian immigrants to the USA (Bates et al., 2008, Khan et al., 1997, Popkin and Richard Udry, 1998, Liu et al., 2009). These studies analyse assimilation effects by comparing BMI values for first, second and sometimes even third generation immigrants. Bates et al. (2008) find that generational status is associated with increased BMI and obesity among Latinos and Asian Americans; Khan et al. (1997) find that generational status is associated with BMI only for Mexican Americans, but not for Cuban American nor Puerto Rican immigrants; Popkin and Udry (1998) find that Asian–American and Hispanic adolescents born in the U.S. are more than twice as likely to be obese as first generation residents; Liu et al. (2009) find that generational status is not a significant correlate of obesity among Hispanic adolescents, whereas Gordon-Larsen et al. (2003) find a striking difference in the BMI of foreign and US-born Hispanic adolescent immigrants, attributed to differences in lifestyle.

Section snippets

Data and methods

This study uses pooled data from three years (2003–2005) of the Victorian Population Health Survey (VPHS), a representative survey of adults (18 and over) in the Australian state of Victoria (for more information see Department of Human Services, 2005). The VPHS is an annual telephone survey that includes extensive information on health, demographic and socioeconomic indicators. Our sample consists of 6376 males and 9407 females, after exclusion of respondents with missing information on one or

Results

Results comparing Natives and the 7 immigrant groups are displayed in Table 2. Due to space restrictions, coefficient estimates and associated standard errors are displayed for the median quantile only.

Coefficient estimates are quite similar across the subsamples, and generally confirm the existing literature (for a review see Wang and Beydoun, 2007): BMI increases by 0.3 points with every life year, but at a decreasing rate. Females are lighter than males, and divorced and never married

Discussion

We find that immigrants assimilate to the bodyweight of the host population over one generation and across all weight categories. QR results show that assimilation occurs for all BMI groups, from normal-weight to morbidly obese, and both for immigrant groups who are larger than Natives in the first generation (South Europeans), and those lighter than Natives in the first generation (East and South Asians). Our results also show that QR results present more information than OLS results which are

Conclusions

We estimate direction and extent of assimilation of immigrants’ BMI to the host population of Australia over one generation. We conduct separate analyses for immigrants from 7 distinct geographical regions, and we use quantile regression (QR) methods to allow for varying impact of assimilation across different weight categories. In short, while standard approaches on the relation between generational status and BMI can answer the question “on average, how does generational status influence

References (26)

  • Lisa M. Bates et al.

    Immigration and generational trends in body mass index and obesity in the United States: results of the National Latino and Asian American Survey, 2002–2003

    American Journal of Public Health

    (2008)
  • Cawley, J. 2000. Body weight and women's labor market outcomes. NBER working paper...
  • Department of Human Services. 2005. Victorian Population Health Survey. Melbourne,...
  • Cited by (24)

    • Ethnicity, length of residence, and prospective trends in body mass index in a national sample of Australian adults (2006–2014)

      2018, Annals of Epidemiology
      Citation Excerpt :

      Acculturation has often been used to explain obesity progression in immigrants, with acculturation typically defined as the process of individual adaptation to the host country's lifestyle, environment, and culture [15]. Cross-sectional studies have shown that proxy measures for acculturation, including generational status (second and subsequent immigrant generations) [7,16,17], younger age at migration [3,18], and longer residence in the host country [3,10,19,20], are associated with higher body mass index (BMI), overweight, or obesity. Acculturation, however, has been criticized as an overly-simplistic concept when based on individual cultural change [15,21–23], and theorists have asserted the need to consider social determinants of health [22,24] and the interplay with other power dimensions such as gender and class [21,22].

    • Immigrants and immigrant health

      2015, Handbook of the Economics of International Migration
    • Influence of neighbourhood ethnic density, diet and physical activity on ethnic differences in weight status: A study of 214,807 adults in Australia

      2013, Social Science and Medicine
      Citation Excerpt :

      Meanwhile, protective effects of ethnic density also emerged for the English. Previous work has reported differences in BMI within ethnic groups and placed some of the explanation on differences in behavioural norms, such as diet and physical activity (Anikeeva et al., 2010; Antecol & Bedard, 2006; Bates et al., 2008; Gray et al., 2007; Haenszel & Kurihara, 1968; Hauck et al., 2011; Kouris-Blazos et al., 1999, 1996; Marmot & Syme, 1976). Our findings confirm that a range of differences in weight status between ethnic and country of birth groups exist and appear to persist even controlling for these conventional risk factors.

    View all citing articles on Scopus
    View full text