Acculturation and Dietary Change in Mexican-American Immigrant Women

Our study explored the process of acculturation among Mexican women living in southwest U.S., and the consequences regarding dietary risk factors associated to health. The cross-sectional study included face to face interviews with 150 migrant women and 150 non-migrant women. Interviews consisted of two non-consecutive 24-hour recalls, and data on anthropometry, acculturation, dietary change and lifestyle. Multiple regression analysis showed that consumption of calories from saturated fat and body mass index (BMI) were significantly higher in migrant women compared to non-migrant women, even after adjusting for other lifestyle and diet related variables. Overall, acculturation seems to be associated with more access to food rather than with a change in consumption of dietary risk components. Furthermore, acculturation was associated positively with socioeconomic status, indicating interplay of socioeconomic and cultural variables related to eating behavior in Mexican immigrant women. Positive association of BMI with acculturation and of acculturation with socioeconomic status suggests that health risk factors among Mexican immigrant women follow similar trends of those of women in their born country.


Introduction
We are experiencing an era where the shift in focus from communicable to chronic diseases is challenging not only the current perspectives of health education, but also the priorities of health care.However, results from interventions to change health-related behaviors that considered lifestyle as short-term practices have been quite unsuccessful.Behaviors resulting from different lifestyles are understood not as isolated acts controlled by the individual, but as acts that are socially conditioned, culturally embedded, and economically constrained.(Green and Kreuter, 1991;CSDH, 2008) [1] [2].
Concurring with the above perspective, being sensitive to how individuals define and understand health according to their cultural orientation will facilitate the health promoter's work.The more a health promoter knows about the beliefs, attitudes, traditions and practices of a given cultural group and how they change within a new cultural setting, the more his/her activities will improve and become effective.Health facilitators that have an understanding and respect for the community's ways could be more easily involved in a program intended to improve its health (González et al., 1991;Cyril et al., 2015) [3] [4].At a national level, it is also imperative to be aware of the needs that different sub-groups in a society can have given their economic and cultural differences.
According to the document Healthy Border 2020: A Prevention & Health Promotion Initiative (2010) [5], the main risk factors for obesity and diabetes among the population of the U.S. and Mexico border communities are physical inactivity, poor diet (high caloric intake), poverty, genes (non-modifiable determinants), lack of breastfeeding, and education/lack of information.Most of this determinants have been documented higher in Hispanics populations compared to White populations (Willey et al., 2012; Siega-Riz et al., 2014) [6] [7].
Not only Hispanics are one of the largest ethnic minority groups in the U.S. (Cortes-Bergoderi et al., 2013)  [8], but the number of Hispanics is increasing almost four times as fast as the rest of the population, totaling 47,435,002 million according to the U.S Census Bureau, (2010) [9].Nationally, Mexicans are the largest Hispanic origin group, making up 64.6% of all Hispanics (Brown and López, 2013) [10].The largest concentrations of Mexican Americans are in the Western states, notably California and Texas, with others growing rapidly, namely in Arizona and New Mexico.California, Texas, New México and Arizona have 32%, 33%, 30% and 27% of Mexican origin population respectively (Brown and López, 2013) [10].
Regional information on Mexican Americans' health has been concentrated in those states with the largest numbers of Mexican Americans, such as California and Texas.There is, however, a growing interest in exploring other places in which the Mexican population is increasing.It is also imperative to understand the differences in the lifestyle and living conditions of Mexican Americans as a result of different historic development of their communities (Robinson, 1998;Roberts, 1995) [11] [12].In this context, it is also worth exploring the challenges and strategies related to health and nutrition that Mexican-Americans are facing in those places, as well as the geographic, socioeconomic, demographic, and cultural characteristics of the environment that shape their lifestyles (Ayala et al., 2008) [13].
Examining the data related to health status among Mexican Americans, it is notable that even though this group belongs to a minority group often classified as low income, it has a health prognosis that is related to variables somewhat different from other low income minority groups.Several studies and review papers, have explored the relationship between socioeconomic status (mainly through income, education and occupation) and health risk factors (i.e.obesity, smoking) (Sobal and Stunkard, 1989;Belcher, et al., 1993 [8] [14]- [21], and concluded that not only is socioeconomic status related to the presence of more health risk factors, but other variables are also involved in the process of dietary behavior.Acculturation has been frequently associated to health risk factors and dietary behavior among Latino and Hispanic populations in the U.S., however several studies proposed that such association may be mediated by other variables such a place of origin, and residence, access to and use of health care facilities, awareness of the diet-health relationship, the importance of physical exercise, and adaptation to new environments (Ayala et al., 2008; Pérez-Escamilla R, 2011) [13] [22].
Within the aforementioned, it is the aim of this study to further the understanding of the present social, cultural, and economic characteristics that shape the lifestyle, and as a consequence, the dietary pattern of a group of Mexican Americans in the border region of Arizona.

Conceptual Framework
The theoretical framework used to guide the present study is based on two ecosystem perspectives.It integrates the conceptual systems from the model of communication-acculturation developed by Kim (1991) [23], and the ecological perspective on nutrition proposed by Pelto (1981) [24], and Sims (1972) [25].
The conceptual framework, incorporates an attempt to examine how differentiation (defined as the extent of the presence of dietary risk factors) in dietary behavior and food consumption among Mexican American families is related to the level of acculturation, socioeconomic, and demographic characteristics.
Consequently, the dependent or outcome variables of the study were food consumption and dietary change, which are the main components of dietary behavior.Additional outcomes resulting from differentiated dietary behaviors and food consumption included an obesity indicator or Body Mass Index (BMI), and body fat distribution (from measures of waist-hip circumferences ratio).Some other variables that can affect the dependent variables were also explored such as smoking behavior, physical activity, alcohol consumption and stress.

Sample and Methods
Although the data analyzed for this study was collected during the summer of 1994 to spring of 1995, some arguments support its current validity.First, The Hispanic Paradox hypothesis documented from evidence that Hispanics living in the US have higher prevalence of several CV risk factors but lower mortality (Ayala et al., 2008; Cortes-Bergoderi et al., 2013) [8] [13]; this paradox has been related to differentiated nutritional, psychosocial and genetic factors.Second, studies on the effect of acculturation on dietary intake and health risk factors among Hispanics and Latinos in the U.S., have reported that conclusive evidence is still lacking (Pérez-Escamilla, 2011) [22].Third, the urgent need for more information regarding obesity and chronic degenerative disease risk factors, such as poor diet, across the U.S. and Mexico border (Healthy Border 2020: A Prevention & Health Promotion Initiative (2010) [5].

Study Design and Sampling
The type of research was cross-sectional and included a purposive site selection from which systematic random selection of units of study (women) were enrolled.Two populations were of interest for this study; migrant Mexican American families (FM) living in Yuma County, Arizona, and low-income nonmigrant Mexican families (NM) from Hermosillo, Sonora, Mexico.This last group was considered as a reference group for nutritional variables as dietary intake, obesity and body composition indicators.Sample size was calculated using general formulas for group comparison and association analysis (Bowner et al., 1988) [26], and resulted in 129 families per group; we interviewed 150 wom- en in each group.The selection of women was carried out at The WIC Clinic in Yuma, Arizona and at a Public Health Center in Hermosillo, México.As women were waiting for their turn to see a doctor or routine evaluation, we selected each 10 th woman in line that met the eligibility criteria (age ≥ 18 years, at least one year of residence in Hermosillo or Yuma, having preschool children, and willingness to participate in the study).Each participating women signed an informed consent.This study was carried out in accordance with the recommendations of the University Committee on Human Subjects from Cornell University in 1994.

First and Second Face-to-Face Interviews
The first interview was conducted at the clinic and lasted 45 to 105 minutes; it consisted of a semi-structured 20-page questionnaire that included a 24-hour recall, a semiquantitative food frequency questionnaire (data not-shown, except for alcohol intake), validated scales for acculturation and emotional stress, dietary change and food preparation, behavioral health risk factors (smoking behavior, physical activity, weight concern, and health care), migration, employment, and educational background (Appendix A).We also collected anthropometric data.The second interview involved a 24-hour recall conducted at least one month from the first interview at the participant's home, and lasted 20 to 30 mins (Conway et al., 2003) [27].Quantitative estimation of food was supported by the use of common food containers, spoons, glasses, and cups.Food components were calculated by using a food dictionary containing foods and beverages consumed by women at two sites.The dictionary included foods from the USDA food data bank (Geghardt and Matthews, 1988) [28], traditional Mexican foods from the INN (National Institute of Nutrition) Food Composi-tion Data for Mexican Populations (Hernández and Chávez, 1980) [29] and the Regional Food Composition Databank provided by C.I.A.D. in Hermosillo, Mexico (Grijalva, et al., 1995) [30].

Dietary Change
Concepts such as retention or addition of foods to common dietary practices among Mexican Americans were used to measure dietary change (Dewey et al., 1984;Romero et al., 1993) [31] [32].Common dietary changes during the year before to the interview and during the last five years were also recorded, as well as reasons for reported changes (Sanjur in 1995) [33].
From height and weight measurements we calculated the Body Mass Index (BMI).We used the WHO categories to define weight status (WHO, 1995) [36].

Acculturation
Acculturation was operationalized through a score obtained by using an acculturation scale developed from selected variables in Kim's (1988) [23] model of communication/acculturation (language preference and use, ethnic identity, and social networks, as well as perceived discrimination) and from the scale of acculturation for Mexican Americans developed by Cuellar (1980) [38] (language preference and use, ethnic identity and social networks).

Other Health Related Variables
In the present study, only current smokers were considered; operationalization of alcohol consumption was described from the data in the food frequency questionnaire.From the daily activities reported in the pilot study done in the summer of 1993 in Arizona, an index of heavy, moderate, and light physical activity (PAL) was developed and registered depending on the activities reported by women.A modified scale developed by Krause and and Goldenhar (1992) [39] was used to measure stress and acculturation in a sample of elderly Hispanics.This scale involved the domains of financial strain, social isolation, and effects of depression.From our previous 1993 pilot study in Yuma, items such as concern about children's behavior and communication with relatives in Mexico were added.A total score of stress was derived.

Ethnography
In addition to the cross-sectional research described before, an ethnographic study was conducted in both studied communities (Patton, 1990) [40].The purpose of this qualitative work was to characterize the communities and to explore food related issues among families in order to further the understanding and support the cross-sectional research with participant observation, and informal talks with key informants.

Statistical Analysis
Questions belonging to dietary practices and socioeconomic and demographic variables were coded and input using the SAS and SPSS statistical software [41] [42] were used on exploratory (multicollinearity, frequency distributions, plot analysis) and descriptive data analysis (means, standard deviations, skewness, kurtosis).The analysis of association and group comparisons were performed using multiple regression analysis and the Pearson product-moment correlation.
Chi-square analysis was also used to compare proportions.

Results
152 migrant women from Yuma County, and 157 non-migrant women from Hermosillo, Mexico were interviewed.Nineteen (12%) of the migrant women (Yuma) and 28 (18%) of the non-migrant women (Hermosillo) were pregnant at the time of the interview, and 9% and 13% (respectively) were lactating.

Socioeconomic and Demographic Characteristics
Table 1 contains data on the main socioeconomic and demographic characteristics of participant women.There were no significant differences (tested by one way ANOVA) between migrant women and non-migrant women concerning family size mother's and father's age, and level of schooling.However, there was a substantial difference in the family structure between the two groups.Migrant families had a higher percentage of female heads of household than did nonmigrant families.Since more than three quarters of migrant and non-migrant women were homemakers, they were not performing paid work during the pe- riod of the first interview.Migrant women were more likely to work in agriculture in Yuma County and as clerks before they migrated.

Acculturation of Migrant Women
The acculturation scale included 11 items which represented the proficiency of language use and preference of language (7 items), social networks (3 items) and identity (1 item).A test of reliability resulted in a Cronbach's alpha [43] of 0.77, indicating a moderately reliable scale.
Scale range varied from 15 points (which indicated that women preferred and used Spanish language at all times, had friends and neighbors only of Mexican descent, attended social gatherings only with Mexican individuals, and identified themselves always as Mexican) to 77 points (which indicated that women used and preferred to speak English, attended social gatherings with American individuals only, and identified themselves always as American).Some of the items in the scale were not relevant for women that reported that they did not listen to the radio, read books, or magazines, or go to parties or social gatherings.Items that were not applicable were coded as 0. Percentages of women who had one, two or three not applicable items were as follows: 30 (20%) for 1 item, 12 (8%) for 2 items and 1 (0.6%) for 3 items.

Food Consumption/Meal Patterns
Migrant and non-migrant women's meal patterns were described using two criteria proposed by Sanjur (1995) [33]: the presence of each food item, and the usual combination of food items within their common cultural foodways.Typical meal patterns included foods consumed by at least 30% of the women in each group.Variations to the main meal pattern involved foods consumed by at least 10% of the women in each group.Comparison of typical meal patterns among migrants and non-migrants reflect the inclusion of very similar food items and meal patterns (breakfast, lunch, dinner, and snack).However, migrant women's lunch included beef as part of the usual meal pattern.In non-migrants diets beef is present for lunch, but just as a variation of the usual lunch pattern.In addition, beef is again included for dinner within the main meal pattern of migrant women and constitutes a more elaborate meal than among non-migrant women.
An explanation for this pattern could be that even when lunch is still the main meal of the day, dinner could become the main meal for those women that follow the American meal pattern.This is true especially when their families and they follow a work schedule in the U.S. Migrant and non-migrant women's fruits and vegetables intake was scarce.Tomatoes and onions were the only vegetables consumed by more than 50% of the sample; however, they were used just as condiments or in very small quantities when cooking foods.
There were no significant differences in mean consumption of total energy, energy from fat, saturated fat, and carbohydrates, as well as cholesterol, sodium, fiber, vitamin A, vitamin C, iron, and calcium intake between migrant and non-migrant women.Energy from total fat is over the recommendations in both studied groups, as well as saturated fat consumption in migrant women.On the other hand, fiber consumption seems to be high in both studied groups when compared to mean consumption in the total adult US population (≈17 g/day) (Storey and Anderson, 2014) [44].

Food Consumption/Dietary Change
Dietary change considered: a) perceived dietary change in the last year and in the last five years and the main reasons to change, and b) actual change in the consumption of 34 food items after migration for migrant women.Whether women's food consumption increased, decreased, did not change or they never tried three different sets of food items (eleven basic foods, twelve traditional foods, and eleven processed foods) was assessed as the percentage of women in every category.In addition, a total score of change was computed for each category.
Among migrants 17% and 23.2% of women responded that they have made some change in their food consumption or cooking methods during the last year and last five years, respectively.Among non-migrants, the percentages were quite similar, with 19% and 18% of respondents stating that they have made some change in the foods they eat or the ways of cooking it.Migrant women's main reasons for dietary change were driven by a change in legal status (i.e.marriage, 5%), more knowledge about cooking (3%), or the presence of an extended family member (i.e.mother-in-law, 2%).Only 8% of migrant women responded that they had made changes because of migration, 16% because of health reasons, and 2% because of food cost, which were originally expected to be the main reasons for dietary change.
Table 2 shows the percentage of migrant women that answered according to the different categories of increased, decreased, did not change, or never tried for each food item.As we can see for basic foods, the trends in change (shadow areas) show that more migrant women have increased their consumption of basic foods or that the consumption of basic foods has stayed the same after migration.In other words, the highest percentages of migrant women's responses to change are distributed in the categories of "increase" and "no change" for basic foods.Interestingly, almost half of migrant women (48%) stated that they had increased their consumption of vegetables and milk, as well as 45% and 44% for fruit and chicken.However, high percentages of migrant women have not changed their basic food consumption (especially for eggs, rice, pasta, and oil).
Regarding change in traditional food consumption, the trend is distributed mainly among "decrease" and "stayed the same" categories.Traditional foods such as Mexican sweet bread, nopal (cactus), chicharrón (pork rinds) and tamales were foods that more than 40% of migrant women reported diminishing their intake.Except for Mexican sweet bread, nopal, chicharrón, and tamales are traditional foods that are not part of the daily consumption of foods among Mexicans, but are eaten during weekends or special events within the Mexican cultural heritage.Traditional foods that constitute part of the basic cultural eat- ing patterns of Mexicans, such as beans, tortillas, and chili, were reported by high percentages of migrant women (72%, 63%, and 59%, respectively) as "stayed the same."This suggests that even after migration the majority of women (and, consequently, migrant families) continued to consume main traditional Mexican foods.Regarding processed foods intake, the trend was distributed mainly within the categories of "increased" and "stayed the same."Percentages of women that increased processed foods consumption were, however, higher across all food items (cold dry cereals, ice cream, instant soup, and canned fruit).

Nutritional Status: Anthropometry
According to different categories of BMI, 37% of migrant women and 36% of non-migrant women were in the overweight range.However, more migrant women (32%) were categorized as obese, as compared with 19% among nonmigrant women.Migrant women exhibit an even higher degree in the category of extreme obesity (4%) as compared to non-migrant women (1%).Regarding waist/hip circumference ratios among migrant and non-migrant women, and using the cut-off point of 0.8%, 52% and 42% of migrant and non-migrant women were at some risk of cardiovascular disease according to their body fat distribution pattern.

Physical Activity
Migrant women showed physical activity level (PAL) ranging from light to moderate levels, while non-migrant women had moderate and heavy PAL levels.
Conversely, migrant women were more likely to be engaged in some kind of regular physical exercise (36%) than non-migrant women (13%).In general, however, high percentages of women in both migrant (64%) and non-migrant (87%) groups did not report engaging in any kind of regular physical exercise.

Smoking and Alcohol Consumption
Smoking did not appear as a prevalent health risk among the study sample, since only a small percentage of women reported being current smokers in both migrant and non-migrant groups (10 and 12% respectively).Among migrant women 16% (1% daily, 3% weekly, 4% monthly, and 6% yearly) consumed some kind of alcoholic beverage (beer, any kind of liquor, or wine).Among non-migrant women, 22% consumed alcoholic beverages (6% weekly, 7% monthly, and 8% yearly).

Stress (Only Migrant Women)
Among migrant women, the reliability test revealed a moderately reliable stress scale (Cronbach's alpha = 0.66) with a point range of 9 to 18. Mean score of stress for migrant women was 12.6 ± 2.2.

Multivariate Analysis
Comparisons of BMI, total energy, energy from total fat, saturated fat and car-bohydrates were done between migrant and non-migrant women, after adjusting for some socioeconomic and demographic variables.Variables included for adjustments were age, education, type of family, physical activity level, energy consumption (for BMI), lactation, pregnancy, and smoking.
Table 3 shows the estimated regression coefficients, significance of variables, R 2 , and model significance, which show significant higher BMI among migrant women.In the case of total energy, energy from total fat and saturated fat and carbohydrates (Table 4), results comparing migrant and non-migrant women's diets indicate that migrant women were consuming significantly less total energy compared to non-migrant women.Also, that there were no differences between migrant and non-migrant women regarding energy from total fat consumption.
Migrant women, however, consumed significantly more calories from saturated fat and fewer calories from carbohydrates than non-migrant women.

Determinants of Risk Factors among Migrant Women
Table 5 shows the predictor variables considered to explain the variation of BMI and waist/hip ratios among migrant women.It also shows the "best fitting" models for the same indicators.From the analysis of the models presented for BMI, it seems the exposure to the American culture (represented by age at migration) became an important variable in explaining differences in BMI in migrant women.The more exposure to the American culture the women have, the more likely they are to increase their BMI.This is explained by the negative direction of regression coefficient for age at migration in the model; when age at migration was not considered, acculturation was significantly and positively related to BMI.Other significant variables in the model were age, consumption of energy and physical activity level.In the case of fat distribution, the full model was not significant in explaining the differences in waist/hip ratios among migrant women.However in the "best" fitting model, acculturation was marginally significant.In addition to acculturation, education was significant in the same model, indicating that the variation in waist/hip ratio is also related to socioeconomic status.

Dietary Risk Factors and Acculturation
One of the main dietary risk factors for chronic diseases is, as referred earlier, high levels of total energy and fat intake, and especially the proportion of total energy from fat.In this study, intake of fat and calories provided by fat and saturated fat are thought to be influenced by dietary change, and in particular by that change related to the process of migration and exposure to a new culture.In this context, the variation in consumption of total energy, energy from total and saturated fat, and carbohydrates was explored mainly as a function of acculturation, controlling for socioeconomic, demographic, and health related variables.
The three sub-scales related to dietary change were also included in the model as was proposed initially in the conceptual model.
The full model presented in Table 6 was not significant as to explain energy consumption.The "best" fitting model, however, seems to explain, though in a Upon further exploring the association of the socioeconomic, demographic and health-related variables and calories provided by total fat and saturated fat consumption, the full model did not explain significantly the variation in consumption of these nutrients.However, there is a trend that shows that saturated fat intake increases as the level of physical activity level rises.The R 2 for this model indicates, however, that the magnitude of variance explained is very small.In summary, it seems that the more acculturated migrant women are, the more energy they consume.Having a higher BMI seems to be associated with the exposure to the new culture, as well as to lower levels of physical activity.
Although the data shows that energy consumption is negatively associated to BMI, when one separates the group of migrant women into women with BMI ≥ 30 and women with BMI < 30, the association of mean energy consumption and BMI behaves differently.Women with BMI < 30 showed a BMI that is negatively and significantly associated with energy consumption (p = 0.001).On the other hand, for women with BMI ≥ 30 the association is positive and significant (p = 0.041).If we look at the association of BMI with energy consumption, and adjusting for physical activity, women with BMI < 30 continue to show a significant and negative association; meanwhile in women with a BMI ≥ 30 association is no longer significant (p = 0.07).These results could mean interplay of effects of physical activity on the association of energy consumption and BMI.In addition, it seems that none of the socioeconomic, demographic, or cultural variables were associated with fat or saturated fat consumption.Fat consumption, however, could be associated to some other health-related variables or lifestyle characteristics not explored in our models.

Predictors of Acculturation
Table 7 presents the estimated regression coefficients and significance of predictor variables for acculturation.The full model contains socioeconomic, demographic, family context, and a self-identity related variable (perceived discrimination).The full model explained 41% of the variation in acculturation and was highly significant.However, a careful interpretation of the R 2 value should be done, since the model contained three variables with condition indexes higher than 30.The same variables that were significant in the full model became even more significant in the "best" fitting model.The proportion of variance explained by this model was still of good magnitude (38%) and significant.
Based on present findings, it seems that the variance in acculturation is ex-  Also, if a mother works, education increases, the family has income, and they are likely to be more acculturated.Surely, these variables could have an impact on the magnitude of social interactions and consequently on the extent of exposure to the U.S. culture.

Discussion
Regarding dietary risk factors, this study suggests that they were higher in the population of migrant women when compared to non-migrant women.Findings also suggest that careful attention should be paid to a potential decrease in the consumption of complex carbohydrates among migrant women, which could potentially imply a "substitution effect" of an increase in consumption of fat, and perhaps saturated fat.According to the overall data from the multiple regression analysis, consumption of calories from saturated fat and BMI seem to be significantly higher in migrant women, compared to non-migrant women, even after adjusting for other weight and diet related variables such as smoking, lactation and pregnancy.There are other variables that in addition to energy are asso-ciated with BMI in migrant women, for example physical activity level.This may indicate that in addition to diet, lifestyle characteristics of migrant women contribute to the differences found in BMI.This difference could well be related to a better economic situation of migrant families and consequently more access to often expensive foods such as beef.
Differences in dietary patterns among migrant and non-migrant women seem to be related to social and cultural ways within the host country.Even when lunch is still the main meal of the day, dinner could become the main meal for those women that follow the American meal pattern.This is true especially when they and their families follow a work schedule in the U.S.This cultural adaptation could have strong consequences for dietary change since women could end up consuming two elaborate meals instead of one.
These changes in meal patterns that affect nutrient consumption have been reported by Sanjur (1995) [33] among Hispanics in the U.S. and specifically in Mexican Americans.This author discusses that "although indigenous food habits of any population group are deeply rooted in local environment as well as local culture, food habits constantly change and are influenced by many factors".
In the context of dietary risk and protective factors, it is worth noting that dietary change carries the presence of both components during the process of change.Increase in the consumption of basic foods, such as vegetables and fruit, could mean a positive change and a contribution to the presence of protective food components as vitamins, minerals, and fiber.On the other hand, increased consumption of foods with a high content of fat, sugar, and sodium, increases health risk factors.The presence of these components can be particularly noted in the higher consumption of whole milk, ice cream, instant soup, and canned fruit, whose consumption is reported as being increased by migrant women.
It is also important to highlight the continued consumption of beans and tortillas, whose supply of complex carbohydrates and fiber to Mexicans is well recognized.These findings are consistent with the apparent similarities of meal patterns among migrants and non-migrants, the largest percentage of calories coming from carbohydrates, and with the high consumption of fiber among both groups.From our own participant observation study we learned that even after migration, Mexican American families continue to consume Mexican food at home.They find the food ingredients from grocery stores in Yuma County, or they bring the indigenous ingredients from the closest Mexican border town (San Luis, Mexico).
According to Pelto (1981) [24], family, with its particular composition and structure, is the central unit in which lifestyle is shaped.Different lifestyles come from different contexts of household structure and composition, and Pelto maintains that families with similar lifestyles will have similar dietary behaviors.
In a similar way, Sims (1972) [25] proposed that, "food intake is a result of transactional patterns between resources which may enter the family from the distal environment, and the family's use of such resources, will vary as a result of that family's decision-making and valuing process."Information that the mother gets from the media, social institutions in the community, interpersonal relationships, and from the members of her family, in addition to physical resources, can be utilized in different ways and can be transformed into specific behavior according to each type of family.
Overall, acculturation seems to be associated mostly with access to food rather than with change in consumption of dietary risk components.This is supported by the data relative to dietary change and from ethnographic fieldwork in the migrant community.Although there is an increase in consumption of processed foods and basic foods, traditional foods such as beans and tortillas are kept in the diets of migrant women.Informants reported having an increase in total consumption of foods after migration, while keeping their Mexican cooking customs.Finding ingredients was not a constraint to their dietary practices given the closeness to their home country and the frequent interaction with relatives in Mexico.Furthermore, acculturation was positively associated with socioeconomic status, indicating that interplay of socioeconomic and cultural variables is associated with food behavior in this group of Mexican immigrants.A study by Gregory-Mercado et al. (2006)  [45], reported a negative association of fruit and vegetables (FV) intake and acculturation among older low socioeconomic status Mexican American women in Arizona; they also reported that persistent low FV intake increased BMI risk in this sample of Mexican-American women.These authors discussed, that socioeconomic status, as well as education, among other social and cultural variables, could modify the relationship of acculturation and dietary change among the Mexican American population; and stressed the need for studies that include dietary patterns, food availability, social structure, health concern, purchasing power, and food preparation responsibilities to make adequate inferences in this population.
On the other hand, several studies have reported contrasting results on the relationship of acculturation and BMI.Khan et al. (1997) [46] found that an increased use of English language (as a measure of acculturation) was associated with a decrease in BMI among Mexican American women.Their analysis, however, included first, second and third generation Mexican Americans from the HHANES (1982-84) national study.Pérez-Escamilla (2011) [22] in a review of studies on acculturation, nutrition and health disparities of Latino populations in the U.S., reported that most of the reviewed studies found a positive relationship of acculturation and obesity; however, this author suggests that to reach appropriate conclusions on the cause-effect of this association, a comparison group of Latino-born population must be included, "since obesity rates change as a function of age and have increased across time in immigrant and nativeborn minority populations".Pérez-Escamilla concludes that to better understand the effect of acculturation on health risks factors, "a cohort study should include a parallel group of the immigrant's home Country".
In this study, the relationship of acculturation with BMI was positive, but women in the sample were by design first generation Mexican immigrants exclusively; they represent a group that is located in the first part of the curve of acculturation and BMI.Moreover, we found that immigrant women had higher BMI than non-migrant women living in Mexico.
For this study sample, there is better access to food than there was in Mexico.
It is, still, however, a low income population whose closeness to its country of origin, together with its primarily Mexican cultural heritage, allows it to keep, to a large extent, traditional food consumption patterns, while having better access to food in general.Their socioeconomic status, while low to U.S. standards, may still signify an improvement over what they previously had in Mexico.

Practical Implications
Some implications of the findings of this study for future sensitive and culturally appropriate nutritional interventions follow.There are subgroups of population living in the US region that based on their cultural heritage and socioeconomic characteristics should be thought of more as a native Mexican population than as a bicultural or acculturated population.The overall behavioral health and nutrition promotion programs should continue encouraging to reduce smoking and alcohol consumption, as well as the consumption of fiber-rich traditional foods as part of good health practices.Physical exercise should be part of the same programs since the data from this study supports the association between physical activity level and BMI.Physical exercise should also be promoted since both groups of women expressed a desire to lose weight.

Study Limitations and Future Research
Some study limitations should be noted.Collection of data was performed in the mid-late nineties; however, published research data on the study matter is limited in such border region, as well as studies that include comparison data from migrant's country of birth.Sampling locations in both study sites differ; in Yuma County the WIC clinic offered nutritional counseling and food aid to immigrant women, but in Hermosillo, Mexico, women attended a doctor's appointment which did not include any nutritional guidance.Finally, it is recommended that in the same areas of the United States, the intergenerational food-related behavior should be studied.For research purposes, the collection and analysis of information within these groups living close to Mexico can improve understanding of food-related behavior and acculturation of future generations.For community-based studies, understanding dietary change and examining points for interventions that differ from those needed by first generation Mexican Americans will greatly improve the effectiveness of strategies for nutritional interventions.
plained by a set of variables representing different domains: family context variables, socioeconomic, and demographic variables.The model seems to indicate that having a father born in the U.S. affects the process of acculturation negatively, while having a grandmother born in the U.S. influences acculturation positively.The more acculturated the women are, they depend less on food stamps.

Table 1 .
Selected sociodemographic characteristics among migrant and non-migrant families.

Table 2 .
Dietary changes among migrant women: increase, decrease, or no change in basic, traditional, and processed foods after migration.

Table 3 .
Estimated regression coefficients and significance of predictor variables for BMIs (log): comparing migrant and non-migrant women.

Table 4 .
Comparing migrant and non-migrant women: estimated regression coefficients and significance of predictor variables for total energy, energy from total and saturated fat, and carbohydrates consumption (squared root).

Table 5 .
Predictor variables for BMIs and waist/hip ratio among migrant women: (full and "best fitting" model).

Table 6 .
Predictor variables, estimated regression coefficients and significance of variables for total energy, energy from total and saturated fat and carbohydrates among migrant women.magnitude, some of the variance in energy consumption.Acculturation was positively and significantly related to energy consumption.Physical activity level as well was significantly, but negatively, related to energy consumption. smaller

Table 7 .
Estimated regression coefficients and significance of predictor variables for Acculturation (log): Full and "best fitting" model.
What other jobs have you had, and where?__________________________ 44.What is your highest grade you completed at school?Have you smoked at least 100 cigarettes in your entire life?Yes___ No___ 47.If yes in number 46: about how old were you when you first started smoking cigarettes fairly regularly?_____________________48.On the average of the entire time you smoked, how many cigarettes did you smoke per day?__________________ 49.Do you smoke cigarettes now?Yes___ No___ 50.If not: How old were you when you stopped smoking?_______yrs.51.If yes: On the average, about how many cigarettes a day do you smoke In the past few weeks, have you felt depressed and very unhappy?Weekly mean Expenses when eating out _____________________ dlls.