Childhood obesity is a major public health issue( Reference Ogden, Carroll and Kit 1 ). Children who are overweight suffer from both psychological and physical ailments, including lowered self-esteem, social exclusion, skeletal pain, sleep apnoea, heart problems and diabetes( Reference Dietz 2 – Reference Benjamin 6 ). The causes of childhood obesity are multifaceted( Reference Harrison, Bost and McBride 7 ). A great deal of research has looked at both specific feeding practices and general parenting style and their relationship with children’s eating behaviours and weight status( Reference Pearson, Biddle and Gorely 8 – Reference Birch, Fisher and Davison 11 ). Investigations have examined these two aspects of the family environment together( Reference Hughes, Power and Fisher 12 – Reference Hubbs-Tait, Kennedy and Page 14 ), but none have looked at parenting style in relation to a wide range of parental feeding practices.
Parenting style encompasses the values, beliefs and practices of parents that are communicated to the child( Reference Darling and Steinberg 15 – Reference Baumrind 17 ). Although parenting style is not specific to the domain of feeding, it relates to children’s eating and weight. For example, children with authoritative parents have healthier eating behaviours and attitudes( Reference Pearson, Biddle and Gorely 8 , Reference Tung and Yeh 18 – Reference Topham, Hubbs-Tait and Rutledge 21 ), are more physically active( Reference Sleddens, Gerards and Thijs 22 ) and have lower BMI( Reference Topham, Hubbs-Tait and Rutledge 21 , Reference Berge, Wall and Loth 23 , Reference Gable and Lutz 24 ), whereas children with authoritarian parents have unhealthier eating behaviours( Reference Arredondo, Elder and Ayala 25 ) and higher BMI( Reference Rhee, Lumeng and Appugliese 10 , Reference Berge, Wall and Loth 23 ). Parents also influence children’s eating behaviours and weight outcomes through specific feeding practices. For example, controlling or instrumental feeding practices (e.g. pressure to eat, restriction of certain foods, using food as a reward) promote overeating and poor dietary habits( Reference Birch, Fisher and Davison 11 , Reference Fisher and Birch 26 – Reference Vereecken, Legiest and De Bourdeaudhuij 30 ) and interfere with children’s ability to self-regulate energy intake( Reference Johnson and Birch 31 ).
Few studies have examined how parenting style relates to specific feeding practices and most have studied only restriction, pressure to eat and monitoring( Reference Collins, Duncanson and Burrows 32 ). Hughes and colleagues measured these three feeding practices and parenting style, and found that authoritarian parents placed more pressure on children to eat than less demanding parents and were more likely to use food restriction( Reference Hughes, Power and Fisher 12 ). Also, authoritative parents were more likely to use monitoring, a feeding practice generally associated with positive outcomes( Reference Hughes, Power and Fisher 12 ). Hubbs-Tait and colleagues measured parenting style and six feeding practices with findings also in favour of authoritative style( Reference Hubbs-Tait, Kennedy and Page 14 ). These studies focused on children of preschool age; how feeding practices relate to parenting styles for older children is under-studied( Reference Taylor, Wilson and Slater 33 ).
Although findings suggest that parenting style relates to specific feeding practices, the practices investigated have been limited, focusing on restriction, pressure to eat and monitoring. Thus, the current study sought to expand on existing findings by examining the relationship of parenting style to parental feeding practices as measured by the Comprehensive Feeding Practices Questionnaire (CFPQ)( Reference Musher-Eizenman and Holub 34 ). Based on previous research, we hypothesized that more flexible parenting styles would relate to healthier practices and that more laissez-faire and rigid parenting styles would relate to less healthy practices.
Methods
Participants
One hundred and ninety-three mothers (mean age 34·3 (sd 6·81) years) were recruited online via Amazon’s Mechanical Turk (MTurk). Research finds that MTurk users more closely match US demographics with regard to sex, race, age and education than other Internet and university samples, and the data obtained through this recruitment method are at least as reliable as those from more traditional recruitment methods( Reference Buhrmester, Kwang and Gosling 35 , Reference Paolacci, Chandler and Ipeirotis 36 ). The majority of participants were Caucasian (76 %) and had an undergraduate degree (37 %) or some college (35 %). Forty-three per cent were employed full- or part-time, and 36 % were homemakers or unemployed. Approximately 3 % of the mothers were underweight (BMI <18·5 kg/m2), 48 % were normal weight (BMI between 18·5 and 24·9 kg/m2) and 49 % were overweight or obese (BMI>25·0 kg/m2). Approximately 8 % of the children were underweight, 46 % were normal weight and 42 % were overweight or obese. Child ages were 7 years (n 51), 8 years (n 44), 9 years (n 42), 10 years (n 32) and 11 years (n 24). Inclusion criteria included: (i) at least one child (aged 7–11 years); (ii) a two-parent household (to control for feeding differences in single-parent households); (iii) English fluency; (iv) US citizenship; and (v) an MTurk rating of successful previous survey completion of 95 % or higher. Participants received a total of $US 0·75 for their participation in the study.
Procedure
A post on MTurk described the purpose of the study, eligibility requirements and the reward for participation. Individuals who wished to participate were directed to an external website to complete the survey. Participants agreed to an informed consent and completed a screening questionnaire about the inclusion criteria. The study was conducted according to the guidelines laid down in the Declaration of Helsinki and all procedures involving human subjects were approved by the Institutional Review Board of Bowling Green State University.
Measures
Parenting styles
The Parental Authority Questionnaire (PAQ)( Reference Buri 37 ) assessed the three parental authority types proposed by Baumrind: permissive, authoritarian and authoritative( Reference Baumrind 17 ). It contains thirty items. The items were reworded to measure mothers’ reports of their own parenting style( Reference Smetana 38 ). Mothers responded to each item using a five-point scale (1=‘strongly disagree’, 2=‘disagree’, 3=‘neutral’, 4=‘agree’, 5=‘strongly agree’). The PAQ has adequate discriminant and criterion-related validity( Reference Kremers, Brug and de Vries 19 , Reference Buri, Louiselle and Misukanis 39 ). In the current study the measures also had adequate reliability (Authoritative α=0·82, Permissive α=0·83, Authoritarian α=0·85).
Parental feeding practices
The CFPQ( Reference Buri 37 ) assessed twelve aspects of parental feeding behaviour (see Table 1). Mothers indicated the frequency with which they engaged in each behaviour using a five-point scale (‘never’ to ‘always’ and ‘disagree’ to ‘agree’). The CFPQ has adequate validity and reliability( Reference Doaei, Kalantari and Gholamalizadeh 40 ).
The Parental Authority Questionnaire (PAQ)( Reference Buri 37 ) was used to assess three parenting style types: Authoritative, n 59; Authoritarian, n 63; Permissive, n 63. The Comprehensive Feeding Practices Questionnaire (CFPQ)( Reference Buri 37 ) and assessed twelve aspects of parental feeding behaviour: (i) child control over feeding (five items); (ii) using food to regulate the child’s emotions (three items); (iii) encouraging balance and variety (four items); (iv) making healthy foods available in the home (four items); (v) using food as a reward (three items); (vi) involving the child in meal planning and preparation (three items); (vii) monitoring the child’s food intake (four items); (viii) modelling healthy eating (four items); (ix) pressuring the child to consume more food at meals (four items); (x) restricting a child’s intake for health-related reasons (four items); (xi) restricting the child’s food intake for weight reasons (eight items); and (xii) teaching about nutrition (three items). All variables are parent report.
a,b,cMean values within a row with unlike superscript letters were significantly different (P<0·05).
Demographic questionnaire
Participants completed questions about their age, race/ethnicity, level of education, employment status, income, height and weight, as well as their child’s age, height and weight. All variables were self- or parent report.
Statistical analyses
ANOVA and post hoc analyses were run to examine if and how key study variables differed between demographic groups. Consistent differences in key study variables between demographic groups were controlled for in subsequent analyses. To investigate the main study questions, participants were grouped by parenting style according to standardized scores. For example, participants whose authoritarian style Z-score was highest were categorized as authoritarian. A MANOVA was then used to examine how these parenting style groups differed on the twelve feeding practices, with post hoc analyses to identify specific differences. Data were analysed using the statistical software package IBM SPSS Statistics Version 22, with P<0·05 taken as the criterion for significance.
Results
Demographic differences
Feeding practices differed (P<0·05) based on maternal race/ethnicity, weight status, employment groups and child weight status. Post hoc analyses with pairwise comparisons showed that mothers of underweight children were less likely to restrict food for weight control than all other mothers. Moreover, mothers of normal-weight children were less likely to restrict food for weight control than mothers of obese children. Mothers of children who were obese were more likely than mothers of underweight children to offer food as a reward. Mothers who were Caucasian were more likely to encourage balance and variety and to make healthy foods available in the home than non-Caucasian participants. Non-Caucasian participants were more likely to offer food as a reward and to teach their children about nutrition than Caucasian participants. Mothers who were unemployed or homemakers were less likely to use food as a reward than mothers who were full- or part-time employed or students. Authoritarian mothers’ mean BMI (29·08 kg/m2) was higher than that of permissive (26·02 kg/m2) and authoritative mothers (25·83 kg/m2; F(2, 190)=4·17, P<0·05). Thus, maternal race/ethnicity (Caucasian v. non-Caucasian), maternal weight status, maternal employment, (unemployed/homemaker v. full-/part-time employed or student) and child weight status were controlled for in subsequent analyses. No feeding practices were related to child age after controlling for demographic covariates.
Relationships among parenting styles and feeding practices
A MANOVA was conducted to examine how feeding practices differed between parenting style groups. After controlling for relevant covariates, the multivariate effect of parenting style was significant (F(24, 336)=3·64, P<0·001, partial η 2=0·21). In univariate tests, parental reports of eleven of the twelve feeding practices differed based on parenting style (see Table 1). Authoritative parents reported using more healthy practices and fewer unhealthy practices than both authoritarian and permissive parents. Permissive and authoritarian parents differed on the unhealthy practices they employed. Pressure to eat was the only practice that did not differ by parenting style.
Discussion
The current study investigated the relationship between parenting styles and parental feeding practices. Overall, parents who had authoritative parenting styles reported more healthy feeding practices than parents with an authoritarian or permissive style. Differences in feeding practices were seen by parenting style in eleven out of the measured twelve feeding practice variables, demonstrating the far-reaching connection between parenting style and feeding practices.
In general, authoritative parents establish rules and standards, but are flexible and receptive to their child’s needs( Reference Rhee, Lumeng and Appugliese 10 ). In contrast, authoritarian parents are inflexible and can be insensitive to their child’s needs, whereas permissive parents do not enforce many rules( Reference Rhee, Lumeng and Appugliese 10 , Reference Buri 37 , Reference Baumrind 41 ). Consistent with this, the feeding practices used by authoritarian and permissive parents can lead to children having difficulty regulating their own food intake, including eating in the absence of hunger( Reference Birch, Fisher and Davison 11 , Reference Arredondo, Elder and Ayala 25 , Reference Francis and Birch 29 , Reference Vereecken, Legiest and De Bourdeaudhuij 30 ).
The future health of children is based, in part, on the behaviours that parents engage in. Parental education on healthy feeding behaviours and parenting styles may contribute to the health of future generations. However, there are several limitations worth noting. While feeding behaviours directly relate to food consumption, parenting style itself does not always directly relate to eating( Reference De Bourdeaudhuij, te Velde and Maes 42 ). The present sample was predominantly Caucasian and well-educated. Also, only mothers in two-parent households with children aged 7–11 years were recruited for the study. Given research indicating that racial/ethnic background and socio-economic status are correlated with both parenting style( Reference McLoyd 43 ) and feeding practices( Reference Skala, Chuang and Evans 44 ), this is especially important. The sample size for the current study was relatively small for the number of tests; therefore the results should be interpreted with caution.
In summary, given the critical role that the parents have in the development and maintenance of children’s eating behaviours and attitudes( Reference Lindsay, Sussner and Kim 45 , Reference Scaglioni, Salvioni and Galimberti 46 ), parenting style should be a key focus in future intervention and prevention programmes.
Acknowledgements
Financial support: This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors. Conflict of interest: None. Authorship: A.K.-B. contributed to the study design, research question, heavy data analysis, heavy manuscript writing. D.H. contributed to the study design, research question, light data analysis, heavy manuscript writing. S.Z. contributed to the study design, heavy data analysis, manuscript writing. D.M.-E. contributed to the study design, data analysis, manuscript editing. Ethics of human subject participation: This study was conducted according to the guidelines laid down in the Declaration of Helsinki and all procedures involving human subjects were approved by the Institutional Review Board of Bowling Green State University. Participants gave their informed consent.