|Home | About | Journals | Submit | Contact Us | Français|
This study examined father and child socio-demographic characteristics in relation to fathers’ feeding-related parenting strategies and whether his parenting strategies were associated with children’s-reported fruit and vegetable (FV), weekly fast food, and daily sugar-sweetened beverage intake among 81 Latino fathers-children pairs. Fathers’ employment status, acculturation, number of children in the home, and child’s age and weight status were associated with the use of different parenting strategies. Additionally, more control was associated with less FV intake, but more reinforcement was associated with more FV intake by children. Fathers play a role in their children’s diet and should be considered in future interventions.
Childhood obesity rates in the United States (US) have reached unprecedented levels with national surveillance estimates indicating that approximately 34% of children aged 6–11 years are overweight or obese (Ogden et al. 2014). Additionally, racial/ethnic disparities exist with Mexican-American boys at greatest risk of being overweight or obese compared to white and non-Hispanic black boys (Ogden et al. 2014). Because being overweight and obese puts children at risk for developing other chronic diseases (Park et al. 2012), additional research is needed to identify potential modifiable risk factors for childhood obesity. Understanding the roles that parents play in factors that affect their children’s weight such as diet can help inform interventions aimed at preventing and controlling childhood obesity.
Parents play a key role in determining their children’s diet by making specific foods available and accessible to their children in (Ding et al. 2012; G. X. Ayala et al. 2007) and outside (G. X. Ayala et al. 2008) the home, by modeling eating behaviors (Oliveria et al. 1992; Morgan et al. 2011; Freeman et al. 2011), and by their use of various parenting strategies (Arredondo et al. 2006). Research with mothers indicates that excessive use of feeding-related control and pressure to eat and use of unhealthy foods as a reward are associated with children’s preferences for unhealthy foods, energy intake, and weight (Faith et al. 2004). Research regarding the influence of fathers on children’s diet is limited (Faith et al. 2004; Khandpur et al. 2014) despite recent evidence that fathers influence children’s dietary intake through modeling and parenting behaviors (Khandpur, Blaine, Fisher, & Davison, 2014). Furthermore, fathers are increasingly responsible for caregiving and child feeding as mothers find employment outside the home (Bianchi 2000). Recognizing the scarcity of research on fathers’ feeding-related parenting strategies, Khandpur and colleagues (Khandpur et al. 2014) reviewed and summarized 20 studies. They determined that fathers consider child feeding well within their role as parents and are responsible for organizing meals at least half of the time and deciding what kinds of foods and how much of each to feed children (Khandpur et al. 2014; Vollmer, Adamsons, Gorin, et al. 2015). Paternal feeding-related parenting strategies most frequently used include pressuring to eat through the use of neutral prompts, physical prompts, and incentives. Compared to mothers, fathers report less frequent monitoring of child food intake, more frequent use of pressuring, and more frequent use of control strategies over the amount of food taken by the child and greater restriction of their child’s food intake (Haycraft en Blissett 2008; Blissett, Meyer, en Haycraft 2006; Hendy et al. 2009; Loth et al. 2013; Tschann et al. 2013; Vollmer, Adamsons, Foster, et al. 2015). Only one study strategically recruited fathers of Mexican origin, but did not examine child diet (Tschann et al. 2013). Understanding how Latino fathers influence their children’s diet is especially important given their central role in the family system (Parke et al. 2004) and the observations that Latino fathers spend more time in caregiving activities than fathers from other ethnic groups (Hofferth 2003).
To address these research gaps, this study described the feeding-related parenting strategies used by Latino fathers, as well as socio-demographic factors associated with their use of these strategies. The second aim of this study was to examine the associations between fathers’ feeding-related parenting strategies and children’s diet, including intake of fruits and vegetables, fast food, and sugar-sweetened beverages. We hypothesized that increased feeding-related control, limit setting, and discipline would be inversely associated with fruit and vegetable intake and positively associated with fast-food and sugar-sweetened beverage intake while monitoring and reinforcement would be associated with healthier eating behaviors. Lastly, given potential differences in fathers’ feeding-related parenting strategies with sons compared with daughters (Orrell-Valente et al. 2007; Khandpur et al. 2014), the third aim explored the moderating effects of child gender on these associations.
This cross-sectional study used baseline data from 81 Mexican-origin fathers and their children who participated in the Entre Familia: Reflejos de Salud study, a family-based, family-systems theory driven (Cox en Paley 1997) randomized controlled trial promoting healthy eating among Latino families in Imperial County, CA. In the Entre Familia study, a convenience sample of 361 Latina mothers and their children were recruited to participate. Although all family members were allowed to participate in the intervention, we recruited fathers from among a 25% of these households to participate in the evaluation cohort along with the mothers and children.
The Entre Familia study involved a partnership between the Institute for Behavioral and Community Health at San Diego State University Research Foundation and Clínicas de Salud del Pueblo, Inc., (CDSDP) a federally qualified health center providing comprehensive primary care services to residents throughout Imperial and Riverside Counties.
Imperial County is located in Southern California and borders Mexico to the south, Riverside County to the north, San Diego County to the west and the State of Arizona to the east. Population estimates for 2014 indicate that of the approximately 180,000 people living in Imperial County, approximately 82% are of Hispanic/Latino origin. Thirty-two percent are estimated to be foreign-born and 62% high school graduates. The median household income in 2009–2013 was estimated at $41,807 with 23% of persons living below the poverty level (U.S. Census Bureau 2015).
Trained bilingual and bicultural research assistants recruited and enrolled 361 families between June 2009 and January 2011. Participants were recruited several ways: engaging with project staff at local events and at community venues; responding to flyers; and, parents with children enrolled at CDSDP received a letter from the clinic informing them of the study. We attempted to recruit fathers from 25% of enrolled households; 81 (22%) fathers consented to participate in the study.
The Entre Familia study was a family-based randomized controlled trial with the primary outcomes based on the mother. Therefore, family eligibility was based on the mother and included: being married and living with her partner, being able to speak and read Spanish, and having a child between the ages of 7 and 13 years. For inclusion into the study, fathers also had to live with the child for at least four days per week. If more than one child was eligible for the study, the child with the month and day of birth closest to the day of the baseline assessment was recruited for the study. After family members provided consent, the research assistant administered an interview to each family member to obtain information on demographics, dietary behaviors and other outcomes of interest. The interviews were followed by measurement of the mother’s and child’s height and weight using NHANES protocols (NHANES 2007). All mothers completed the interview in Spanish and all fathers and children were given the option to complete it in either English or Spanish. A higher proportion of fathers (n=71, 88%) completed a Spanish interview, and similar proportions of children completed English (n=43, 53%) and Spanish (n=38, 47%) interviews. Parents were asked to think about the specific child enrolled in the study when responding to questions related to their parenting. All study protocols were approved by the Institutional Review Board of San Diego State University to ensure protection of human participants.
We used the fruit and vegetable 2-item cup instrument from the National Cancer Institute’s Food Attitudes and Behavior Survey (Yaroch et al. 2012) to assess children’s daily intake of fruits and vegetables. Children were asked “About how many cups of fruit, including 100% fruit juice, do you eat or drink each day?” and “About how many cups of vegetables, including 100% vegetable juice, do you eat or drink each day?” (Spearman’s ρ=0.32; p≤0.01). Children were also read the portions for one-cup equivalents and were asked to select from the following response options: “None”, “½ cup or less”, “½ to 1 cup”, “1–2 cups”, “2–3 cups”, “3–4 cups”, “4 or more cups”. We have used this assessment approach previously with children (G. X. Ayala et al. 2007). For analysis, the two items were transformed from categorical ranges to a continuous scale by taking the midpoint of the response ranges and then summed to obtain an estimate of daily cups of fruits and vegetables consumed.
Daily servings of sugar-sweetened beverages was assessed with one question: “In a typical day, how many cans or glasses of regular soda, Kool-aid, Tampico or punch do you drink?” (Horton et al. 2013).
Child height and weight were measured in triplicate using a stadiometer and Tanita weight scale, respectively. The average of the three height and three weight measurements was used to calculate body mass index (BMI), which was dichotomized as normal weight versus overweight/obese using CDC age and gender-specific cut-points (Centers for Disease Control and Prevention 2014).
Child demographics included age and gender (male or female).
Father’s feeding-related parenting strategies associated with his child’s diet were assessed using the Parenting Strategies for Eating and Activity Scale (PEAS) (Larios et al. 2009), which was modified for the Entre Familia study based on our formative research and given our exclusive focus on diet (Guadalupe X Ayala et al. 2011). The 25-item scale was composed of the following six subscales (Table S1): control–five items measuring the father’s use of controlling strategies such as telling his child to eat everything on his/her plate (α=0.74); limit setting-three items measuring the frequency with which the father set limits on the amount of food his child ate including high-fat snacks and sugar-sweetened beverages (α=0.46); monitoring–five items measuring the frequency with which the father monitored his child’s eating by keeping track of the sweet snacks, fruits and vegetables, and sugar-sweetened beverages his child ate (α=0.87); reinforcement–four items measuring the father’s use of praise when his child ate healthy foods (α=0.74); permission–four items measuring the frequency with which the child had to ask for permission before getting a second helping, eating a snack, or drinking sugar-sweetened beverages (α=0.80); and discipline–four items measuring the frequency with which the father disciplined his child for getting a second helping, eating a snack, or drinking sugar-sweetened beverages without his permission (α=0.72). To improve the reliability of the limit setting subscale, one item (frequency of limiting the number of fruits and vegetables the child ate) was excluded which resulted in a Spearman’s ρ correlation of 0.55 (p≤0.001) for the remaining two items. Responses were made on a 5-point scale ranging from “Never” to “Very often.” A mean score was computed for each subscale with higher scores indicating more frequent use of paternal feeding-related control, limit setting, monitoring, reinforcement, permission, and discipline.
Acculturation was assessed using the 24-item Bidimensional Acculturation Scale (Marin en Gamba 1996), which provides an acculturation score for two major cultural dimensions (Hispanic and non-Hispanic domains) across three language-related domains (language use, linguistic proficiency, and electronic media use). For example, participants were asked rate how often they speak English with friends, how well they understand television programs in English, and how well they understand radio programs in Spanish. Responses to the 12 Hispanic (α=0.80) and 12 non-Hispanic (α=0.94) domain items were made on a 4-point Likert scale with two sets of response options: “almost never” to “almost always” and “very poorly” to “very well.” Mean scores across the two domains were used to classify participants as traditional – individuals who have retained a strong attachment to their Latino culture and have not assimilated to the United States way of life (Hispanic domain score > 2.5 and a non-Hispanic domain score ≤ 2.5), or bicultural – individuals who have adhered to their Latino culture but who have also adopted many practices of the dominant culture in the United States (Hispanic and non-Hispanic domain scores > 2.5).
Father demographic characteristics included age, dichotomized education (less than high school versus high school graduate or greater), employment status (unemployed/employed part time; employed 40 hours/week; and employed >40 hours/week), and a combined indicator of immigration status using country of origin and years in the US (US-born; foreign-born and living in the US less than 15 years; and foreign-born and living in the US more than 15 years (Reyes-Ortiz et al. 2009)). Monthly household income (four income brackets ranging from <$999 to $3,000+) and number of children under the age of 18 living in the home were obtained from the mother’s interview.
We examined father and child characteristics as correlates of fathers’ feeding-related parenting strategies using Multiple Linear Regression (MLR) via PROC REG on SAS. We used an alpha level of 0.05 and a Bonferroni corrected p-value threshold of 0.008 (α = 0.05/6). Regression diagnostics included normal probability plots for examining non-normality, studentized residuals for detecting extreme observations, and tolerance levels to assess collinearity between independent variables; no violations were observed. Variables that were associated with any of the feeding-related parenting strategies in the MLR analyses at alpha = 0.05 and child gender were retained for further analyses. We conducted three separate MLR analyses with each of the child dietary outcomes as a dependent variable. Unadjusted models included the six parenting strategies and adjusted models included all potential confounders previously identified (father’s employment status, acculturation, number of children in the home, and child’s age, weight status, and gender). We also examined a series of main effects with corresponding two-way interactions between child gender and paternal parenting strategies to examine the moderating effects of child gender.
Demographic characteristics of the 81 fathers and their children are presented in Table 1. Fathers were on average 41 years old (SD=9), most were employed 40 hrs/week or more (59%), all self-identified as Latino, Hispanic, Mexican/Mexican-American or of Spanish origin (100%) and most were born in Mexico (81%). The majority (63%) were also considered traditional based on the Hispanic (mean=3.4, SD=0.4) and non-Hispanic (mean=2.2, SD=0.8) domains of the acculturation scale (Marin en Gamba 1996). Children were on average 10 years old (SD=2) and 54% were categorized as overweight or obese. Children reported consuming on average 2.9 cups (SD=1.7) of fruits and vegetables per day, fast food 1.4 days per week (SD=1.3), and 1.9 servings of sugar-sweetened beverages per day (SD=1.2). Regarding their use of the parenting strategies, on average fathers reported sometimes-to-often using feeding-related reinforcement (mean=3.36, SD=0.92), limit setting (mean=3.13, SD=1.16), and control (mean=3.10, SD=0.88), and almost never-to-sometimes using permission (mean=2.63, SD=1.16), monitoring (mean=2.55, SD=1.06), and discipline (mean=2.07, SD=0.90) feeding-related parenting strategies.
As indicated in Table 2, socio-demographic characteristics associated with some feeding-related parenting strategies included father’s employment status, father’s acculturation, number of children in the home, and child’s age and child’s weight status (p≤0.05). Compared to fathers who were unemployed/employed part time, being employed over 40 hours per week was associated with more frequent use of limit setting and reinforcement. Being bicultural, rather than traditional, was associated with more frequent use of control and discipline. An increase in the number of children living in the home was associated with more frequent use of control, monitoring, reinforcement, and discipline. A one-year increase in the child’s age was associated with a decrease in the use of limit setting, reinforcement, and permission parenting strategies. Additionally, having a child who was overweight or obese was associated with more less frequent use of reinforcement of healthy eating behaviors by fathers. After Bonferroni adjustment, however, only employment status, number of children in the home, and child’s age were significantly associated with use of feeding-related reinforcement (p=0.006), discipline (p=0.008), and permission (p=0.007) parenting strategies, respectively. Father’s age, education, country of origin/years in the US, and child’s gender were not associated with the father’s use of parenting strategies (n.s.).
In the MLR analysis, more frequent use of feeding-related control by fathers was associated with less fruit and vegetable intake (β=−0.49, SE=0.25, p=0.05). This result persisted after adjusting for father’s employment status, acculturation, number of children in the home and child’s age, gender, and weight status (β=−0.71, SE=0.29, p=0.02). On the other hand, more frequent use of feeding-related reinforcement of healthy eating was associated with more fruit and vegetable intake after adjustment (β=0.69, SE=0.34, p=0.05). None of the parenting strategies were associated with fast food or sugar-sweetened beverage intake (n.s.).
There was a significant interaction between child gender and paternal use of feeding-related reinforcement (βInteraction=1.21; p=0.02). Depicted in Figure 1 is the moderating effect of child gender on the association between paternal use of low (2.0) versus high (4.0) reinforcement and child’s weekly fast food intake with continuous covariates entered at their means and categorical covariates entered at the referent. Low reinforcement was associated with more frequent fast food intake among girls and less frequent fast food intake among boys and high reinforcement was associated with similar fast food intake across both genders. There were no other significant interactions between child gender, other feeding-related parenting strategies, and fruit and vegetable intake and sugar-sweetened beverage intake (interaction n.s.).
In this study, correlates of fathers’ feeding-related parenting strategies related to children’s dietary intake included father’s employment status, acculturation status, the number of children living in the home, and child’s age and weight status. Additionally, more frequent use of feeding-related control was associated with less fruit and vegetable intake, but more frequent use of reinforcement was associated with more fruit and vegetable intake in children. Level of acculturation has been shown to influence mothers’ parenting strategies; more acculturated mothers may be more likely to view bribes, threats, and punishment as effective strategies for eliciting specific eating behaviors from their children (Kaiser et al. 2001). In the present study, being a bicultural father, rather than traditional, was associated with more frequent use of controlling and disciplining strategies, which are negatively associated with children’s healthy eating (Khandpur et al. 2014). This is consistent with previous research with Latino fathers in which traditional Mexican beliefs among non-acculturated fathers, such as greater endorsement of positive machismo attitudes including beliefs of familial responsibility, dignity, honor, and respect for parents and elders, is associated with greater child-report of positive father involvement (Cruz et al. 2011). While we found no association between country of origin and use of the parenting strategies, in particular feeding-related permission, one small study of mostly US-born Latino fathers found that both mothers and fathers had a predominantly permissive parenting style in which children chose their food with little to no direction about unhealthy foods (Turner et al. 2014). In addition, in our study, feeding-related permissive parenting was not related to any of the child dietary behaviors. In the present study, more children in the household was associated with more frequent use of feeding-related control, monitoring, reinforcement, and discipline parenting strategies, which may reflect more involvement in childrearing by fathers of larger families or potentially, fewer resources available and thus greater use of these feeding-related strategies. We found that child’s age was inversely associated with limit setting and permission required by fathers. Developmental literature suggests that parents seek to foster autonomy and independence, which, in turn, can result in less family communication, and by extension, less use of these strategies (Racz en McMahon 2011). Our observation that child weight status was inversely associated with feeding-related reinforcement is consistent with other studies that have found that higher child adiposity is associated with other parenting strategies such as control and low monitoring of food intake (Loth et al. 2013; Brann en Skinner 2005). Lastly, child gender was not associated with any of any of the feeding-related parenting strategies, which is consistent with several studies which have found no difference in paternal feeding practices for boys and girls (Khandpur et al. 2014). Future studies should continue to explore these associations as they provide possible targets for father-focused interventions.
Consistent with prior research (Khandpur et al. 2014; Faith et al. 2004) and with research with mothers (Arredondo et al. 2006), results of this study indicate that more frequent use of control, such as telling children to eat all the vegetables on their plate, is inversely associated with healthy eating and this association was similar across sons and daughters. Additionally, the use of more feeding-related reinforcement was positively associated with consumption of healthy foods. However, the association between reinforcement of healthy eating and fast food intake was modified by child gender; among fathers with low use of feeding-related reinforcement, daughters consumed fast food more frequently than sons. This finding is consistent with other studies that have found that mothers’ use of reinforcement is inversely associated with intake of nutrient-dense foods (Lloyd et al. 2014) and is potentially explained by the hypothesis that children may interpret the unwarranted reinforcement of healthy foods as coercion (Davison en Campbell 2005). While fathers are likely to employ different parenting strategies with sons and daughters (Orrell-Valente et al. 2007), additional research is needed to understand how low feeding-related reinforcement may lead to differences in dietary behaviors between sons and daughters.
This study used cross-sectional data and so causality and directionality cannot be determined. Child diet was based on self-report and was limited to three single-item questions. While children as young as six years old have been shown to be capable of reliably reporting dietary intake, at least for one main meal at school (Hunsberger et al. 2013), future studies should consider more extensive validated measures of child diet such as the use of food frequency questionnaires (Garcia-Dominic et al. 2011) or 24-hour dietary recall (Thompson en Subar 2001) given the potential for social desirability bias. Also, feeding-related parenting strategies were based on father self-report which are susceptible to social desirability bias (Kristjansdottir et al. 2009). Lastly, the small sample size available for analyses could have resulted in Type II error and subsequently in failure to observe associations in the outcomes examined. The small convenience sample also limits the generalizability of study findings. This may be especially true for fathers who do not live with their children and, therefore, may have different interactions with them. Despite these limitations, the study includes data from both the father and the child, and as such, contributes to our understanding of the intricate relationships between fathers and their children in a population for which obesity is a serious public health challenge.
These findings emphasize the importance of incorporating fathers into behavioral and family-based research. Interventions aimed at improving diet and health among children may be able to effect change by intervening with fathers. An important next step is to determine methods for engaging fathers given challenges associated with recruiting men into health promotion research (Mitchell et al. 2007). This study further adds to the literature on feeding-related parenting strategies and shows that fathers’ use of behaviors such as control and reinforcement influences their children’s behaviors. Moreover, these behaviors may be operating on sons and daughters differently. Future research should continue to explore these observations so that effective interventions for Latino families can be developed. As obesity in both children and adults becomes increasingly prevalent, involving fathers in research is essential.
This research was funded by the American Cancer Society (Grant G00007105) awarded to Dr. Guadalupe X. Ayala. American Cancer Society had no role in the design, analysis or writing of this article.