Parental Guidance and Children’s Healthy Food Consumption: Integrating the Theory of Planned Behavior with Interpersonal Communication Antecedents

Abstract

The study of parenting practices on child food consumption has often been characterized as having great utility but lacking in theory. In contrast, the theory of planned behavior (TPB) has often been suggested to be limited in its utility. To address these gaps, interpersonal constructs – the concepts of active parental guidance (e.g., nutrition education) and restrictive parental guidance (e.g., rule-setting) – were integrated as antecedents to the original TPB variables in predicting child fruits and vegetables (FV) consumption. We surveyed 210 child/adolescent participants, aged between 10 and 16 in Singapore. Results from structural equation modeling showed that the integrative TPB model displayed an acceptable fit. Parental active guidance was associated with all three proximal predictors of behavior. Our results suggest that there are promising contributions to the theory of planned behavior in predicting child food consumption behavior by considering interpersonal antecedents.

Notes

¹ Having been conceptualized for the purposes of understanding the role parents play in managing and regulating children’s television usage, the concept of parental mediation was developed to measure the frequency of different interpersonal communication strategies parents utilize to mitigate the effects of television on their children. The core assumption is that social interactions regarding television between parents and children are an important socialization process that can have protective effects on children (Nathanson, 1999; Clark, 2011). While cultivation theory posits that television viewing can potentially have negative effects on people (Gerbner & Gross, 1976), parental mediation suggests that certain interpersonal communication strategies parents use can mitigate and moderate these effects (Nathanson, 1999). Parental mediation have been explicated to comprise of three different strategies with regards to child television viewing: active mediation, which refers to the frequency of discussion between parents and children on television content; restrictive mediation, which refers to the frequency of rule-making about television viewing; and co-viewing, which simply refers to the frequency with which parents watch television together with their children (Nathanson, 1999; Valkenburg, Krcmar, Peeters, & Marseille, 1999).

² Vaughn et al. (2016) conceptualized rules and limits as distinct from restriction that is coercive in nature. Rules and limits reflect the “what, when, where, and how much” (p.105) foods can be eaten by the child. In contrast, restriction refers to “parent-centered, authoritarian-type limits on a child’s access to foods or opportunities to consume those foods” (p.100). It was further argued that restriction reflected coercive practices that involve no reasoning, and can manifest in threats and punishments. To be clear, this study examines restrictive guidance as the verbal setting of rules and limits, and does not conflate the stylistic imposition of those rules.

³ Originally, Ajzen (1991) defined normative beliefs as “the likelihood that important referent individuals or groups approve or disapprove of performing a given behavior (p.195)”. This means that Ajzen (1991) conceived of subjective norms as reflecting a person’s belief about the social approval of a behavior. In addition to research on the theory of planned behavior, research conducted within the social norms approach (SNA) has identified two distinct types of normative beliefs (Perkins & Berkowitz, 1986). First, injunctive norms are beliefs about what others think one ought to do (in the context of a specific behavior). In this sense, injunctive norms are similar to the original conceptualization of subjective norms in the theory of planned behavior (Lapinski & Rimal, 2005; Rivis & Sheeran, 2003). Second, descriptive norms are beliefs about what others actually do. In other words, descriptive norms are the perceived popularity of a behavior. In a paper that sought to provide guidance on how to construct a questionnaire that tapped on the theory of planned behavior variables, Ajzen (2006) suggested that the operationalization of subjective norms in the theory of planned behavior should encompass both descriptive and injunctive components. As what ought (injunctive norms) is sometimes confused with what is (descriptive norms) (Cialdini, Reno, & Kallgren, 1990), the two different types of perceived norms can sometimes be difficult to untangle (Rimal & Real, 2005). This might be especially true for young children and adolescents, who might perceive social approval and popularity as similar. For example, when children are seated in the canteen and see all their friend group having vegetables on their plates, they might think that they ought to eat vegetables too.

⁴ Regarding attitudinal beliefs, the evaluation of an object or behavior is based on the expectancy-value model, which posits that individuals hold beliefs about the expected value outcome of a behavior (Ajzen, 2001). These beliefs about the expected value of the behavior summate into an overall attitude towards the behavior (Fishbein, 1963). Although there can be many differing, and even conflicting, beliefs about the value of a behavior in question, only beliefs that are accessible in memory can contribute to an individual’s prevailing attitude about the behavior (Ajzen, 2001), highlighting the central role communication plays in attitude formation in the area of food consumption. Parents who communicate frequently and emphatically about why eating certain foods are important, can increase the accessibility of beliefs that lead to a child’s attitude towards certain foods, shaping the likelihood of them eating or choosing to avoid certain foods. On the other hand, rules can also shape attitudes, through the implicit message to children that some foods are explicitly not good for them.

⁵ First, a list of public schools in Singapore was obtained from Singapore’s Ministry of Education. We then divided the list of schools into five regions (North, South, East, West, Central) based on their official administrative regions as demarcated by the Singapore Urban Redevelopment Authority. We randomly selected schools from within each of these administrative regions, and commenced the data collection process at the residential block with the smallest property number immediately adjacent to the school premises. We started recruitment from the top floor of each block and began with the apartment with the smallest unit number. If the attempt was successful, the researcher moved on to the adjacent unit. Otherwise, the researcher skipped one unit. After every floor, the team moved down one level. Upon completion of the residential block, the researchers moved to the block with the next smallest property number and repeated the process until every block adjacent to the school premises had been attempted.

⁶ In addition to univariate normality illustrated in Table 3, Mardia’s multivariate kurtosis was calculated using DeCarlo’s (1997) SPSS macro. The multivariate kurtosis calculated was 77.35, with a standardized measure of 9.26 (p < .001), indicating non-normal data.

⁷ The estimator “MLM” was defined in order to correct for biased estimates among the fit indices, resulting in a mean-adjusted chi-square test statistic that is robust to non-normal data (Satorra & Bentler, 1994).

⁸ MeMoBootR, similar to the SPSS PROCESS macro, utilizes a product-of-coefficients strategy to examine the significance of indirect effects, argued to be more powerful compared to the Baron and Kenny approach and the Sobel Test (Preacher & Hayes, 2008). Bootstrapped confidence intervals were estimated to avoid statistical power problems caused by asymmetric and non-normal distributions.

Additional information

Funding

This work was supported by the National Medical Research Council [NMRC/HSRG/0046/2013].