How Time in the US and Race/Ethnicity Shape Parents Feeding Practices and Child Diet Quality

Abstract

Childhood obesity prevalence is high among children from immigrant/refugee households who live in high-income countries. Poor child dietary intake is a critical risk factor for elevated obesity prevalence and food parenting practices have been found to be associated with child dietary intake and eating behaviors. The main aim of this study was to examine the associations between migrants’/refugees’ food parenting practices, the length of residence time in the US, race/ethnicity, and child diet quality. The current study included 577 families from three racial/ethnic groups that include mostly foreign-born parents (Latino, Hmong, and Somali/Ethiopian), and a comparison group of 239 non-Hispanic White families. Results showed that for Latino and Hmong parents, some food parenting practices varied by how long they had lived in the US. For example, more recently moved parents engaged in more non-directive (e.g., avoid buying sweets) practices compared with US-born parents. In contrast, Somali/Ethiopian parents engaged in different food parenting practices than White parents, regardless of time in the US. Results also showed that diet quality among Hmong children was lower if their parents were US-born compared to foreign-born. Future researchers may want to consider studying why some food parenting practices change when parents move to the US and explore whether there is a combination of food parenting practices that are most useful in promoting a healthful child’s diet and weight among immigrant and refugee families.

Introduction

Prevalence and consequences of childhood obesity

Children from immigrant/refugee households in the United States (US) have a high prevalence of overweight and obesity (Akresh, 2007). In addition, the weight status of children of migrant/refugee parents is often higher than non-migrants’ children of the same race/ethnicity when living in high-income countries (Amirazodi, Turcotte, & Hunter, 2018), which may explain some of the disparities in childhood obesity by race/ethnicity. Childhood obesity often tracks into adulthood (Jennings, 2011; Cardel, Maga Rey, 2003; Freedman et al., 2001) and is associated with multiple short- and long-term physical and psychological health consequences such as sleep apnea, type 2 diabetes, asthma, non-alcoholic fatty liver cirrhosis, and some cancers (Bass, & Eneli, 2015; CDC, n.d.). Many of these diseases (e.g., diabetes) can change the quality of life for a child and potentially the whole family, in addition to being a financial burden. As a result, it is anticipated that the current American generation of children will have a shorter life span compared to the previous two generations of children due to high childhood obesity prevalence (Olshansky et al., 2005).

Factors related to obesity, child diet, and food parenting practices

One factor contributing to elevated obesity prevalence is poor child dietary intake (Nicklas et al., 2000). In a review by Ochoa and Berge (2017), food parenting practices (e.g., restriction, pressure-to-eat) were found to be a key influential factor on child dietary intake and eating behaviors. Food parenting practices have also been identified as a main factor within the home environment linked to childhood obesity (Birch et al., 2003; Cardel et al., 2015). Prior studies have shown that migrant/refugee parents have unique predictors of engaging in food parenting practices, such as culture, migrants’ previous experience with deprivation (Cheah & Hook, 2012), socioeconomic status (Lindsay et al. 2017), desire to keep their heritage (Roche, Goto, Zhao, Wolff 2015), and a different acculturation pace between parents and their kids. Immigrant parents have been found to be more likely to either control their child’s dietary intake in order to maintain their traditional eating habits (Wilson & Renzaho, 2015) or give children “American” foods to avoid conflict (Trofholz et al., 2020; Colón-Ramos et al. 2017).

Time in the US

Overweight and obesity rates are positively associated with length of time in the US for migrant adults (Singh et al., 2011). Likewise, the length of time living in high-income countries has been positively associated with childhood obesity in migrant/refugee children (Amirazodi, Turcotte, & Hunter, 2018; Murphy, Robertson & Oyebode 2017). Non-US-born parents may need time to become familiar with the language and food environments in the US. The “obesogenic environment” of the US where energy-dense, nutrient-poor foods are accessible and available is another well-reported risk factor that can influence food parenting practices (Swinburn, Egger, Raza, 1999; Osei-Assibey et al., 2012).

Theoretical model

This study was guided by the Assimilation Model (Brym et al., 2020; Brown & Bean, 2006), which describes the process by which immigrants and host people come to resemble each other. The model highlights the changes in lifestyles and behaviors of immigrants/refugees over the years in their host country. The process of change occurs over several generations (MPI, 2006). Analyses in the current study focus on the assimilation, or lack thereof, of food parenting practices over time to the dominant US race group – White families – and to families of the same race/ethnicity who migrated to the US in earlier cohorts.

The main aim of the current study was to examine the association between migrants’/refugees’ food parenting practices and child diet quality with respect to the length of residence time in the US, which is intended to capture acculturation due to longer exposure to US-born families, by race/ethnicity. We hypothesized that the length of time in the US among Latino, Hmong, and Somali/Ethiopian parents would be associated with differences in food parenting practices, which would then affect the diet quality of their children. Furthermore, we hypothesized that the acculturation of food parenting practices would differ by race/ethnicity; that is, for some groups, practices will converge to those of White parents (the dominant US “culture”), and for others, practices will converge to US-born parents of their same racial/ethnic group.

Methods

Data were taken from Family Matters, an incremental, mixed-methods study with two phases (Berge et al., 2017). Phase I was a cross-sectional, mixed-methods investigation of the home environment of low-income and racially/ethnically diverse families (n=150) with a 5-7-year-old child (Berge et al., 2017). Phase II, which the current study is built on, followed parents from low-income and racially/ethnically diverse households (n=1307) who had a 5-9-year-old child. Data collected in Phase II included both an online survey and ecological momentary assessment (EMA) with a sub-sample (n=627) (Berge et al., 2017). All study protocols were approved by the Institutional Review Board at the University of Minnesota (1107S02666).

Parent/child dyads were recruited for Phase II from primary care clinics in Minneapolis/St. Paul. The clinics identified 5-9-year-old children who had recently had a well-child visit, and therefore, a recent height and weight measurement. A letter was sent from the primary care physician inviting the family to participate in the Family Matters study; recruitment letters provided a unique code that parents could use to access the online survey. Parents/guardians consented online prior to beginning the survey.

Parents were eligible to participate in the study if they met the following eligibility criteria: (1) the child was 5-9 years old; (2) the person completing the survey was the primary guardian of the 5-9 year old child; (3) the child lived with the parent/guardian more than 50% of the time; and (4) the family self-identified with one of the following racial/ethnic groups: Latino, Hmong, Somali/Ethiopian, African-American, Native American, or non-Hispanic White (hereafter, White). These groups were chosen as they are prominent racial/ethnic groups in the Twin Cities, MN area (Darboe, 2003; Pfeifer, Sullivan, Yang, and Yang, 2012; Minnesota Department of Health, n.d.). This study focused on four of the six groups in the Phase II online survey sample (n=816) of Family Matters: Latino, Hmong, Somali/Ethiopian and White families.

Translation of Materials

All study materials (e.g., consent forms, survey questions) were translated into Hmong, Somali/Ethiopian, and Spanish. The translation was conducted by bi-lingual and bi-cultural staff members.

Survey Development

Development of the online survey followed best practices including using psychometric survey data and qualitative interview data from Phase I to inform survey items to be kept or cut, prioritizing validated measures and carrying out test-retest on all survey items. Test-retest reliability of the online survey was completed by 125 participants (~20 per race/ethnicity) by taking the survey two times within two weeks. Intraclass correlation coefficients (ICC) were estimated from mixed models with participant-level random intercepts to capture the degree of agreement between measurements. Agreement was high (ICC>0.8) for more static questions (e.g., receives SNAP, relationship status, height, household size), moderate (ICC<0.8 & ICC>0.6) for questions about most food-related attitudes and behaviors (e.g., meal frequency, food shopping and preparation attitudes and behaviors, child eating behaviors), and low (ICC<0.6) for questions that are expected to vary over time (e.g., stress, dietary intake, home food availability). A validity sub-study was also carried out in Phase I of Family Matters to test validity of the online dietary intake questions. Participants (n=75) completed the adapted Children’s Eating Habits questionnaire and three 24-hour dietary recalls (gold standard measure of child dietary intake). Correlations between parent-report of child dietary intake on the adapted Children’s Eating Habits questionnaire and the three 24-hour dietary recalls were similar to previously published food frequency questionnaires in children and adolescents (Lanfer et al., 2011; Rockett et al., 1995; Metcalf et al., 2003; Huybrechts et al., 2009; Buzzard et al., 2001).

Measures

Expanding on prior research (Loth et al., 2016), we categorized food parenting practices into four scales: non-directive practices (e.g., modeling healthy eating, and avoiding unhealthy foods), directive practices (e.g., pressuring to eat, restricting foods, bribing), controlling practices (e.g., limiting children’s choices around meals), and emotional practices (e.g., connecting food with comfort). Non-directive practices have been found to be health-promoting, while the other three practice scales are associated with less healthful eating (Ochoa and Berge, 2017; Birch et al., 2003; Cardel et al., 2012). See Table 1 for a description of the food parenting practice measures used in the current study and how they were operationalized. We also examined a composite child diet quality score modeled after the Healthy Eating Index, where a score of 0 represents the lowest possible diet quality and 100 represents the highest possible diet quality (Krebs-Smith et al., 2018). Table 2 describes how the diet quality score was created.

Table 1:

Description of Study Measures

Question Wording	Response Options	Variable Creation
Racial/Ethnic Group & Time in the US
Please select the race/ethnicity that best characterizes your household (e.g., the foods you cook for your family, the language you speak at home, the holidays you celebrate).	White/Caucasian; Latino; Hmong; Somali/Ethiopian	Eight indicators were created: US-born Latinos, Latinos living in the US 10+ years, Latinos living in the US <10 years, US-born Hmong, Hmong living in the US 10+ years, Hmong living in the US <10 years, Somali/Ethiopian living in US 10+ years, and Somali/Ethiopian living in US <10 years (no Somali/Ethiopian parents in the sample were US-born. In addition, we created three indicator variables only among the Latino, Hmong, and Somali/Ethiopian race/ethnic groups: Born in the US, 10+ years in the US, and <10 years in the US.
Were you born in the United States?	Yes; No
How long have you lived in the United States?	Less than 1 year; 1 to less than 5 years; 5 to less than 10 years; 10 or more year
Food Parenting Practices
Controlling
Allow [child] to choose which foods to have for meals (SD=1.1)	Disagree; Slightly Disagree; Neutral; Slightly Agree; Agree	Items are reverse coded and summed so that higher score indicates more controlling food parenting practices. Alpha=0.64
Let [child] decide when s/he would like to have his/her meal (SD=1.2)
Allow [child] to wander around/get up from the table during a meal (SD=1.1)
Emotional
Give [child] something to eat to make him/her feel better (e.g., when anxious, sad) (SD=1.2)	Never; Once in a while; About half of the time; Very often; Always	Items are summed. Alpha=0.91
Give [child] something to eat to make him/her feel better when s/he has been hurt (SD=1.2)
Non-Directive: Modeling & Covert
I model healthy eating for [child] by eating healthy foods myself (SD=1.1)	Disagree; Slightly Disagree; Neutral; Slightly Agree; Agree	Items are summed. Alpha=0.83
I try to eat healthy foods in front of [child] even if they are not my favorite (e.g., vegetables) (SD=1.2)
Avoid going to restaurants or fast food places which sell unhealthy foods with [child]? (SD=1.1)	Never; Once in a while; About half of the time; Very often; Always
Not buy foods that you like because you do not want [child] to have them? (SD=1.2)
Try not to eat unhealthy foods when [child] is around? (SD=1.3)
Avoid buying cookies, candy, and other treats to avoid bringing them into the house? (SD=1.2)
Avoid having snack foods such as candy and chips in the house? (SD=1.2)
Avoid having unhealthy foods in the house? (SD=1.2)
Directive: Instrumental, Pressure-to-Eat, Restriction
If [child] misbehaves, I withhold his/her favorite foods/sweets/desserts (SD=1.3)	Never; Once in a while; About half of the time; Very often; Always	Items are summed. Alpha=0.75
I use desserts as a bribe to get [child] to eat his/her main course (SD=1.1)
I have to be especially careful to make sure [child] eats enough (SD=1.2)	Rarely; Sometimes; Often; Always
[Child] should always eat all the food on his/her plate (SD=1.4)	Disagree; Slightly Disagree; Neutral; Slightly Agree; Agree
If [child] says “I’m not hungry,” I try to get him/her to eat anyway (SD=1.3)
If I did not guide or regulate how [child] eats, he/she would eat much less than he/she should (SD=1.5)
I have to make sure that [child] does not eat too many sweets (candy, ice cream, cake, or pastries) or his/her favorite foods (SD=1.3)
I intentionally keep some foods out of [child]’s reach (SD=1.5)
I offer sweets (candy, ice cream, cake, pastries) or favorite foods to [child] as a reward for good behavior (SD=1.4)
If I did not guide or regulate [child]’s eating, he/she would eat too many junk foods/favorite foods (SD=1.4)
Covariates
Does your family receive any of the following? (Check all that apply)	Food support/stamps (SNAP/EBT); WIC; Free or reduce cost school breakfast; Free or reduced cost school lunch; SSI; MFIP; Healthy start or daycare assistance; None	An indicator was created for those receiving any of the 7 listed types of public assistance
What was the total income of your household before taxes in the past year?	Less than $20,000; $20,000-$34,000; $35,000-$49,000; $50,000-$74,900; $75,000-$99,999; $100,00 or more	Four indicators were created for income $20,000-$34,999, $35,000-$49,999, $50,000-$74,999; and $75,000 and above (income less than $20,000 is the reference category).
What is the highest level of education that you have completed?	Middle school or junior high; Some high school; High school graduate or GED; Vocational, technical, trade, or other certification; Associate degree; Some college; Bachelor’s degree; Graduate or professional degree; Other	Two indicators were created for completing less than a high school degree or equivalent or completing a 4-year college degree or more (completing a high school degree or equivalent and/or some college is the reference category).
Child Weight Status	Child height and weight were taken from their most recent well-child visit (within 6 months).	Two indicators were created for having a percentile BMI between 85% and 95%, and a percentile BMI >95% (having a percentile BMI <85% is the reference category).
Food insecurity
In the LAST 12 MONTHS, did you (or other adults in your household) ever cut the size of your meals or skip meals because there wasn’t enough money for food?	Yes; No	Families who responded Yes and/or Often true/Sometimes true for more than 1 of the 6 questions were determined to be food insecure. This is the 6-item short form of the food security survey module developed by the USDA (Blumberg et al., 1999; USDA, n.d.).
How often did this happen?	Almost every month; Some months but not every month; Only 1 or 2 months
In the LAST 12 MONTHS, did you ever eat less than you felt you should because there wasn’t enough money to buy food?	Yes; No
In the LAST 12 MONTHS, were you ever hungry but didn’t eat because you couldn’t afford enough food?	Yes; No
In the LAST 12 MONTHS, the food that we bought just didn’t last, and we didn’t have money to get more.	Often true; Sometimes true; Never true
In the LAST 12 MONTHS, we couldn’t afford to eat balanced meals.	Often true; Sometimes true; Never true

Table 2:

Description of Child Dietary Intake Measure

Question Wording	Response Options & Points Assigned	Healthy Eating Index (HEI) Equivalent Category
Overall Description: Child dietary intake was measured based on an adapted version of the Children’s Eating Habits Questionnaire that had been tested in a validity sub-study in Phase I of the Family Matters study (Lanfer et al., 2011; Trofholz, Tate, et al., 2017). This adapted measure is similar to a food frequency questionnaire. The parent is asked to report on the child’s dietary intake within the past four weeks. See specific categories below.
In the past month, how many times has [child] eaten…
Dark-Green vegetables (e.g. broccoli, collard greens, spinach, romaine lettuce)	Never or less than 1 time/week: 0 points Few times a week: 2 points Nearly everyday: 4 points 1+ time/day: 5 points	Greens & Beans Category: The scores for Dark-Green Vegetables and Legumes categories were summed, with 5 being the maximum Greens & Beans score.
Legumes (e.g., beans, split peas, lentils, hummus)	Never or less than 1 time/week: 0 points Few times a week: 2 points Nearly everyday: 4 points 1+ time/day: 5 points
Other Vegetables (e.g., tomatoes, carrots, cabbage, corn, potatoes)	Never or less than 1 time/week: 0 points Few times a week: 1 points Nearly everyday: 3 points 1 time/day: 4 points 2+ times/day: 5 points	Total Vegetable Category
Fruit (e.g., oranges, grapefruit, berries, apples, bananas)	Never or less than 1 time/week: 0 points Few times a week: 3 points Nearly everyday: 5 points 1 time/day: 7 point 2 times/day: 9 points 3 times/day: 10 points	Total Fruit Category: Fruit and 100% fruit juice scores are summed, with 10 being the maximum Total Fruit score.
100% fruit juice	Never or less than 1 time/week: 0 points Few times a week/Nearly everyday: 1 point 1+ times/day: 2 points
Whole grains (e.g., whole wheat bread, corn tortillas, brown rice, oatmeal)	Never or less than 1 time/week: 0 points Few times a week: 3 points Nearly everyday: 6 points 1 time/day: 8 point 2+ times/day: 10 points	Whole Grains category
1% or skim milk, white or flavored	Never or less than 1 time/week: 0 points Few times a week: 4 points Nearly everyday: 7 points 1+ times/day: 10 points	Dairy category: 1% or skim milk and Whole or 2% milk categories are summed, with 10 being the maximum points
Whole or 2% milk, white or flavored
Nuts, seeds, and nut butters	Never or less than 1 time/week: 0 points Few times a week: 2 points Nearly everyday: 3 points 1+ times/day: 5 points	Plant Proteins category
Refined grains (e.g., white bread, flour tortillas, white rice, white pasta)	Never or less than 1 time/week: 10 points Few times a week: 8 points Nearly everyday: 6 points 1 time/day: 3 point 2+ times/day: 0 points	Refined Grains category
Salty snacks (e.g., chips, pretzels, crackers, popcorn)	Never or less than 1 time/week: 10 points Few times a week: 8 points Nearly everyday: 6 points 1 time/day: 3 point 2+ times/day: 0 points	Sodium category
Fried vegetables (e.g., French fries, tater tots, onion rings)	Never or less than 1 time/week: 10 points Few times a week: 6 points Nearly everyday: 4 points 1 time/day: 2 point 2+ times/day: 0 points	Saturated Fats category: The scores for Fried vegetables and Baked goods are averaged to create the Saturated Fats score.
Sugar sweetened drinks (e.g., fruit drinks, Capri Sun, pop/soda, Kool Aid, Gatorade)	Never or less than 1 time/week: 10 points Few times a week: 8 points Nearly everyday: 6 points 1 time/day: 3 point 2+ times/day: 0 points	Added Sugars category: Scores for Sugar sweetened drinks, Baked goods, and Candy are averaged.
Baked goods (e.g., cookies, cakes, muffins, donuts)
Candy (e.g., chocolate, licorice, candy bars, fruit snacks)
Overall dietary intake quality score	To create an overall dietary intake score using the adapted Children’s Eating Habits questionnaire, staff dietitians created a score using the Healthy Eating Index-2015 (HEI) as a guide. The HEI was used as a guide because of its reliability and validity in assessing diet quality (Guenther et al., 2014). Similar to the HEI, the overall dietary intake score created using the Family Matters study gives participants points for consuming healthy food categories (e.g., Dark Green Vegetables) and for not consuming food categories that should be eaten in moderation (e.g., Sodium) (Trofholz, Tate, et al., 2017). The points for each category were then summed to create an overall dietary intake score. Higher score indicates a higher diet quality.

Statistical Analysis

Adjusted analyses were conducted using Ordinary Least Squares regressions where the outcome variable is either the child diet scale or one of the four food parenting practice scales, and the independent variables were indicators for both the length of time in the US and the child’s race/ethnicity, as well as the covariates described in Table 1. We ran one set of regressions on the full sample to capture differences relative to the White families (as the dominant US-born group in the US), and one set each for the Latino, Hmong, and Somali/Ethiopian families to capture associations with the length of time in the US within immigrant/refugee groups (relative to US-born families of the same race/ethnicity). Because there are very few Somali/Ethiopian parents who were born in the US (n=2), we pooled US-born Somali/Ethiopian parents with parents living in the US 10+ years.

Results

The study sample included 816 participants from four different racial/ethnic groups (White, Hmong, Latino, and Somali/Ethiopian). The 239 White families are compared to the 577 Latino, Hmong, and Somali/Ethiopian families. Table 3 provides descriptive statistics on all measures in this analysis in these two sub-samples.

Table 3:

Characteristics of Participants

	Latino/Hmong/Somali/Ethiopian Sample (n=577)	White Sample (n=239)
	N (percent)	N (percent)
Race/Ethnicity
Latino	215 (37.3%)	0 (0%)
Hmong	227 (39.3%)	0 (0%)
Somali/Ethiopian	135 (23.4%)	0 (0%)
White	0 (0%)	239 (100%)
Primary Caregiver Time in US
Born in US	152 (26.3%)	233 (97.5%)
10+ years in the US	322 (55.8%)	5 (2.1%)
<10 years in the US	103 (17.9%)	1 (0.4%)
Primary caregiver’s education
Less than HS degree	152 (26.3%)	8 (3.3%)
High School degree or some college	305 (52.9%)	37 (15.5%)
4-year College degree or more	120 (20.8%)	194 (81.2%)
Household receives public assistance	315 (54.6%)	32 (13.4%)
Household annual pre-tax income
Less than $20,000	153 (26.9%)	13 (5.5%)
$20,000 - $34,999	178 (31.3%)	13 (5.5%)
$35,000 - $49,999	122 (21.4%)	9 (3.8%)
$50,000 - $74,999	78 (13.7%)	33 (13.9%)
$75,000+	38 (6.7%)	170 (71.4%)
Missing	8 (1.4%)	1 (0.4%)
Marginal, low or very low food security	165 (28.6%)	18 (7.5%)
Child’s weight status
BMI <85th percentile	372 (64.4%)	195 (81.6%)
BMI 85th-95th percentile	99 (17.2%)	31 (13.0%)
BMI >95th percentile	106 (18.4%)	13 (5.4%)
Outcomes	Mean (SD)	Mean (SD)
Controlling Food Parenting Practices (range 3-15)	10.2 (2.7)	10.9 (2.1)
Emotional Food Parenting Practices Scale (range 2-10)	4.4 (2.4)	2.9 (1.2)
Directive Food Parenting Practices Scale (range 10-50)	30.2 (7.3)	25.5 (6.3)
Non-directive Food Parenting Practices Scale (range 8-40)	27.1 (6.7)	27.5 (5.4)
Child’s Diet Quality Scale (range 20-85)	57.4 (10.7)	62.0 (9.3)

In the Latino/Hmong/Somali/Ethiopian sample, 26% were born in the US, 56% have lived in the US for 10+ years, and 18% have lived in the US for less than ten years. There were also a small number of foreign-born individuals (2.5%) in the White sample. While there were noticeable differences in parental education, public assistance receipt, household income, food security, and child weight status across the sub-samples, food parenting practice scores and child diet scores were somewhat similar.

Table 4 displays associations between race/ethnicity, time in the US, food parenting practices, and child diet quality, adjusted for parent’s education, receipt of public assistance, household income, food security status, and child’s weight status. The reference group in Table 4 is White families and their average outcome is shown in the top row for comparison. These results show to what extent migrant families acculturate to the dominant race/ethnic group in the US over time.

Table 4:

Adjusted Estimates of the relationship between child diet quality, food parenting practices, race/ethnicity, and time in the US

Dependent variable:	Controlling Food Parenting Practices (range 3-15)	Emotional Food Parenting Practices (range 2-10)	Directive Food Parenting Practices (range 10-50)	Non-directive Parent Food Parenting Practices (range 8-40)	Child’s Diet Quality Scale (range 20-85)
Full Sample (n=816)
Mean Outcome for Reference (SD)	10.9 (2.1)	2.9 (1.2)	25.5 (6.3)	27.5 (5.4)	62.0 (9.3)
Latino US-born (ref=White)	−0.271 (0.421)	0.321 (0.343)	1.331 (1.187)	−0.703 (1.054)	−2.585 (1.680)
Latino 10+ yrs	0.552 (0.358)	0.313 (0.292)	2.945** (1.011)	1.833* (0.897)	0.112 (1.430)
Latino <10 yrs	1.237** (0.464)	0.113 (0.379)	3.342* (1.310)	2.513* (1.164)	−2.427 (1.854)
Hmong US-born	−0.338 (0.368)	1.126*** (0.300)	3.609*** (1.038)	−2.211* (0.921)	−11.33*** (1.468)
Hmong 10+ yrs	−0.476 (0.331)	0.992*** (0.270)	3.209*** (0.933)	−1.660* (0.829)	−7.078*** (1.320)
Hmong <10 yrs	−1.757 (0.923)	2.676*** (0.753)	8.902*** (2.606)	4.068 (2.314)	−5.560 (3.687)
Somali/Ethiopian 10+ yrs	−1.413*** (0.381)	1.680*** (0.310)	3.128** (1.074)	1.717 (0.954)	−1.887 (1.519)
Somali/Ethiopian <10 yrs	−1.322** (0.454)	2.971*** (0.370)	4.226** (1.280)	2.104 (1.137)	−0.175 (1.811)

Coefficients from Ordinary Least Squares regressions presented with standard errors in parentheses.

^***p<0.001

^**p<0.01

^*p<0.05.

All models adjusted for primary caregiver’s education, whether the household receives any public assistance, household income, food security status, and child’s weight status.

Table 5 presents these same adjusted associations but examines each of the three racial/ethnic groups with the most migrants separately. Thus, the reference groups in Table 5 are those of the same race/ethnicity who have been in the US the longest: US-born parents for Latino and Hmong families, and parents who migrated to the US ten or more years ago for the Somali/Ethiopian families. As in Table 4, the mean outcome for the reference group is shown in the top row of each panel. These results show the extent to which migrant families acculturate to behaviors of families of the same race/ethnicity who have been in the US longer.

Table 5:

Adjusted Estimates of the relationship between child diet quality, parenting practices, and time in the US by race/ethnicity

Dependent variable:	Controlling Food Parenting Practices (range 3-15)	Emotional Food Parenting Practices (range 2-10)	Directive Food Parenting Practices (range 10-50)	Non-directive Food Parenting Practices (range 8-40)	Child’s Diet Quality Scale (range 20-85)
Latino Sample (n=215)
Mean Outcome for Reference (SD)	10.3 (2.1)	3.6 (2.0)	28.3 (7.4)	25.9 (5.9)	58.9 (9.0)
Latino 10+ yrs (ref=US born Latino)	0.903 (0.460)	−0.151 (0.408)	0.842 (1.249)	2.354* (1.115)	2.113 (1.755)
Latino <10 yrs	1.489* (0.592)	−0.311 (0.525)	0.871 (1.607)	2.706 (1.434)	−0.975 (2.258)
Hmong Sample (n=227)
Mean Outcome for Reference (SD)	10.3 (2.2)	4.5 (2.1)	30.8 (6.6)	24.2 (5.1)	50.3 (9.5)
Hmong 10+ yrs (ref=US born Hmong)	−0.029 (0.333)	−0.042 (0.274)	−0.252 (0.918)	0.950 (0.843)	5.291*** (1.413)
Hmong <10 yrs	−1.146 (0.928)	1.614* (0.763)	4.635 (2.558)	6.330** (2.349)	5.168 (3.940)
Somali/Ethiopian Sample (n=135)
Mean Outcome for Reference (SD)	9.0 (3.1)	5.1 (2.7)	30.1 (8.5)	28.7 (7.8)	60.0 (9.8)
Somali/Ethiopian <10 yrs (ref=Somali/Ethiopian 10+ yrs)	0.187 (0.577)	0.981 (0.533)	0.751 (1.667)	0.330 (1.476)	1.745 (1.836)

Coefficients from Ordinary Least Squares regressions presented with standard errors in parentheses.

^***p<0.001

^**p<0.01

^*p<0.05.

All models adjusted for primary caregiver’s education, whether the household receives any public assistance, household income, food security status, and child’s weight status.

Latino Families

US-born Latino families were not significantly different from White families in terms of all four food parenting practices (Table 4). Latino parents who migrated to the US more than 10 years ago used more directive (p<0.01) and non-directive (p<0.05) food parenting practices than White parents. Latino families who have lived in the US for less than 10 years engaged in more controlling (p<0.01), directive (p<0.05) and non-directive (p<0.05) food parenting practices than White families. Compared to US-born Latino parents (Table 5), Latino parents who migrated to the US 10 or more years ago used more non-directive food parenting practices (p<0.05), and those who migrated less than 10 years ago engaged in more controlling food parenting practices (p<0.05). Thus, Latino parents appear to acculturate to US food parenting practices with respect to controlling and non-directive practices, but generally have higher directive food parenting practices than White parents (i.e., 28.3 (Table 5) vs. 25.5 (Table 4) and within-group differences were not significant (Table 5)).

The average diet quality of Latino children in this sample was not significantly different from White children or Latino children whose parents were born in the US regardless of time in the US.

Hmong Families

In contrast to Latino parents, US-born Hmong parents had significantly different food parenting scores than White parents (Table 4). Specifically, US-born Hmong parents engaged in significantly more emotional food parenting practices (p<0.001), more directive food parenting practices (p<0.001), and less non-directive food parenting practices (p<0.05) than White parents. Hmong parents who migrated to the US more than 10 years ago also used emotional food parenting practices (p<0.001), more directive food parenting practices (p<0.001), and less non-directive food parenting practices (p<0.05). Hmong families who have lived in the US for less than 10 years engaged in more emotional (p<0.001) and directive (p<0.001) food parenting practices than White families. US born Hmong parents (Table 5) and Hmong parents who migrated to the US 10 or more years ago were not significantly different in terms of all four parenting practices. Hmong parents who migrated less than 10 years ago engaged in more emotional (p<0.05) and non-directive food parenting practices (p<0.01) than US-born Hmong parents. Thus, Hmong parents appear to acculturate to US food parenting practices only with respect to non-directive practices, but generally have higher emotional (i.e., 4.5 (Table 5) vs. 2.9 (Table 4)) and directive food parenting practices (i.e., 30.8 (Table 5) vs. 25.5 (Table 4)) than White parents.

The average diet quality of Hmong children with parents who migrated to the US less than 10 years ago was lower (but not significantly so) than those of White children (Table 4). The difference in child diet quality when comparing Hmong children with parents who migrated ten or more years ago to White children was large and significant (p<0.001). When comparing Hmong children with US born parents to those with parents who migrated (Table 5), the diet quality is better among children in migrant families (significantly so for those in the US 10+ years). These results combined indicate that the average diet quality of recent migrant Hmong children is the same as White children, but becomes worse as the families live in the US longer such that Hmong children with US-born parents have an average diet quality score of 50.3 (Table 5), compared to 62.0 (Table 4) for White children.

Somali/Ethiopian Families

Somali/Ethiopian parents had significantly different food parenting scores than White parents (Table 4). Both recent migrants and those living in the US for 10 or more years engaged in less controlling practices (p<0.01), more emotional food parenting (p<0.001), and more directive food parenting practices (p<0.01) than White parents. There were no differences in food parenting practices among Somali/Ethiopian parents by time in the US (Table 5) suggesting that Somali/Ethiopian parents are not changing their parenting practices over time. Instead, Somali/Ethiopian parents just have different parenting behaviors than White parents.

The average diet quality of Somali/Ethiopian children in this sample was not significantly different from White children and did not vary by the length of the time that their parents lived in the US.

Discussion

Overall, the findings support the main hypothesis and theoretical model that time living in the US and race/ethnicity is associated with food parenting practices and child diet quality among migrant/refugee families, and that there are differences in patterns by race/ethnicity. Specifically, recently arrived Latino and Hmong parents engaged in more non-directive (e.g., avoid buying sweets) food parenting practices than US-born counterparts, suggesting some acculturation of this parenting practice over time. Non-directive practices, like modeling healthy eating, have been found to promote a healthy diet (Loth et al., 2017).

Recently arrived Latino parents engaged in more controlling food parenting practices than US-born Latino parents. This is consistent with prior research indicating that non-US-born parents tend to be more controlling in terms of food parenting practices than White parents (Loth et al., 2017). In contrast, we found that Hmong parents engaged in controlling food parenting practices at the same rate as White parents while Somali/Ethiopian parents, regardless of time in the US, are significantly less controlling than White parents. This finding highlights differences across race/ethnicities among immigrant/refugee groups to the US.

We also found that immigrant parents had different food parenting practices than White parents that appear to persist over time. Regardless of the time length or birthplace (<10 years, > 10 years or US-born), Somali/Ethiopian, Hmong, and Latino parents engage in higher levels of directive food parenting practices compared to White parents, which aligns with prior literature (Berge et al. 2017; Loth et al., 2017). Directive food parenting practices (e.g., restriction and pressure-to-eat) can negatively influence kids’ dietary health (Birch et al., 2003; Ochoa and Berge, 2017), and thus may explain why the child dietary quality was found to be lower in these groups than for the White group in our analysis.

Recently moved Hmong and Somali/Ethiopian parents engaged in high levels of emotional food parenting practices. Certain previous experiences may potentially drive the increased use of emotional food parenting practices. This finding may warrant further investigation.

Finally, results showed that diet quality among Latino and Somali/Ethiopian children was not significantly different than White children regardless of time in the US, but that the average diet quality of Hmong children was lower if their parent had lived in the US longer. Thus, the migrants’ paradox, where migrants have better health outcomes than US counterparts of the same socioeconomic status (Stanek et al., 2020), is not observed in any of the three migrant samples examined. Potential hypotheses for lower Hmong child’s diet quality include: (1) the influence of less non-directive food parenting practices by time in the US, and/or (2) the potential role of cultural foods in Hmong children’s diet. For example, Berge et al. (2018) found that parents from Hmong backgrounds were more likely to pressure their kids to eat refined grains (i.e., white rice) because of the importance of this food to the Hmong culture.

Primary care providers, pediatricians, nurses, dietitians, and social workers may want to discuss food parenting practices with migrant/refugee families as soon as they arrive in the US to promote healthy food parenting practices and behaviors. Furthermore, while professionals can be part of the efforts working to pinpoint the root causes, at the same time they can help make parents aware of their food parenting practices and which practices are associated with better child diet.

This study is not without its limitations. First, while the full sample is large (n=816), when the sample is disaggregated by race/ethnicity and time in the US, the cell sizes become very small, reducing our power to precisely estimate differences across subgroups. Second, the measure of child diet quality is based on parent responses to survey questions instead of three 24-hour dietary recalls and thus may be subject to some measurement error. While all materials were translated into Spanish, Somali, and Hmong, we cannot be sure that participants were interpreting questions the same way, thus survey responses may have been influenced by the participants’ level of health literacy. Relatedly, because the main focus of this paper is on food parenting practices and overall diet quality, we did not examine components of the diet separately; future research may want to investigate specific types of food. Third, while it was beyond the scope of this study to examine the influence of food insecurity on our study outcomes, it would be important for future research to include food insecurity in studies to further understand the complicated relationships between food insecurity, food parenting practices, time in the US, acculturation, and child eating behaviors and diet quality. Fourth, this is a cross-sectional study, and thus we cannot ascertain whether the differences observed by time in the US are due to changes in parents’ behavior over time or due to changes in the cohorts arriving in the US over time. Finally, the sample of migrant/refugee families observed all live within the Twin Cities area and are not representative of US migrants/refugees generally.

Conclusion

Our study showed that time living in the US and race/ethnicity were associated with food parenting practices and child diet quality among migrant/refugee families. Some of the food parenting practices used by migrant/refugee families have been shown to be associated with poor child diet (e.g., emotional and directive) and some health-promoting food parenting practices used by migrant parents (e.g., non-directive) decline with time in the US. Thus, future research using longitudinal data is needed to determine whether and how parents’ practices change over time in the US and explore whether there is a combination of food parenting practices that is most useful in promoting a healthful child’s diet and weight.