Self-reported and observed feeding practices of Rhode Island Head Start teachers: Knowing what not to do

Abstract

Purpose

Through their feeding practices, adult caregivers play an important role in shaping children's eating behaviors. However, the feeding practices of child care teachers have received little attention. The purpose of this study was to compare child care teachers' self-reported feeding practices and observed feeding practices during a preschool meal.

Methods

Rhode Island Head Start teachers (n = 85) were observed during breakfast and lunch where feeding practices were coded using a tool adapted from the Environmental Policy Assessment and Observation (EPAO) tool. Teachers completed a questionnaire adapted from the EPAO Self-Report to capture self-reported feeding practices. Agreement between reported and observed was compared by percent agreement.

Results

Teachers were predominantly White (89%) and female (98%). There was a higher level of agreement among self-reported and observed controlling feeding practices (78.8–97.6% agreement) compared to healthful feeding practices (11.8–20.0% agreement).

Conclusions

Although self-report measures are typically used to capture feeding practices, there are inconsistencies between self-report and observation measures. The inconsistencies found among healthful self-reported and observed feeding practices have implications for future research protocols, measurement refinement, and training of child care teachers.

Introduction

Childhood obesity is a serious public health issue. Early prevention efforts targeting preschool-age children is of particular importance given that taste preference and eating behaviors develop early in life (Birch and Fisher, 1998, Davison and Birch, 2001, Grimm et al., 2014, Skinner et al., 2002), and often persist into adulthood (Kelder, Perry, Klepp, & Lytle, 1994). Mothers have been considered the primary caregivers during this critical period (McBride and Mills, 1993, Nicklas et al., 2001), and their feeding practices during meals have been shown to influence a child's dietary intake (Hoerr et al., 2009, Papaioannou et al., 2013, Park et al., 2015, Vereecken et al., 2010) and weight status (Cardel et al., 2012, Hughes et al., 2008, Lumeng et al., 2012). However, an increase in the annual enrollment of full day child care programs from 34 percent in 1990 to 49 percent in 2014 (Natonal Center for Education Statistics, 2016) suggests a need to examine the feeding practices of child care teachers (Dev et al., 2013, Larson et al., 2011). In order to do this, appropriate measurement tools that capture feeding practices of teachers are needed.

Measuring feeding practices of teachers is a nascent area and of the few studies that have done this, most have used self-report measures, which can be easy to use and have less participant burden (Dev et al., 2013, Dev et al., 2014, Hendy and Raudenbush, 2000, Hendy, 2002, Larson et al., 2011). Major limitations of using self-report measures (e.g., surveys, questionnaires, focus groups, interviews), however, include response bias, such as social desirability, as well as problems with memory, judgments and generalizations (Hughes et al., 2013, Lanigan, 2012, Lumeng et al., 2008, Mita et al., 2013, Sharma et al., 2013). Rather than capturing actual feeding practices that occur, self-report measures may instead be tapping into a teacher's perception of what feeding practices they think they are using during meals (Hughes et al., 2007). Few studies have used observation to capture teacher feeding practices, which may provide valuable information beyond self-reported data.

Given that most of the feeding practice literature has been conducted with parents, this research helps inform the measurement of teacher feeding practices. For example, several studies with mothers and their children are novel in that they have directly observed the caregiver-child feeding relationship in naturalistic environments (Bergmeier et al., 2015, Gardner, 2000, Hughes et al., 2013, Edelson et al., 2016, Fries et al., 2017, Johnson et al., 2017, Pesch et al., 2016). In particular, observations are effective in providing insight to the verbal and nonverbal contexts that occur during feeding (Hughes et al., 2013). Of the studies completed with parents, most focus only on measuring the feeding environment in one setting at one given time, rather than examining whether feeding practices are stable across environments with other influential caregivers such as child care teachers (Hughes et al., 2013).

Combined, mixed-method approaches utilizing both self-report and observation provide the opportunity to draw on the strengths of each methodology, enabling a more rigorous study to draw stronger inferences than either method alone (Bergmeier et al., 2015, Hughes et al., 2013, Zoellner and Harris, 2017). For example, mixed-method approaches have been utilized to evaluate mother-child mealtime behaviors (Haycraft and Blissett, 2008, Sacco et al., 2007). However, a systematic review by Bergmeier, Skouteris, and Hetherington (2015) found no significant relationships between self-reported and observed maternal feeding practices (Bergmeier, Skouteris, & Hetherington, 2015). This review also found the most widely used measure of self-reported parent feeding practices, the Child Feeding Questionnaire, was not always significantly associated with observational measures of parent feeding practices (Bergmeier et al., 2015, Lewis and Worobey, 2011). Although one would not expect the two measures to be identical, one would expect that the observational coding system would capture similar constructs as the traditional and validated self-report instrument chosen for their study (Hughes et al., 2013, Bergmeier et al., 2015, Vaughn et al., 2013). In contrast to the home environment, teacher feeding practices in a child care setting have shown moderate congruency between observed and self-reported assessments of feeding practices (Hughes et al., 2007). Given the more structured feeding environment of a child care center compared to a home, other external factors that influence feeding behaviors may come into play. For example, child care environments that enforce nutrition policies and programs may influence the feeding practices of teachers during meal times (Dev et al., 2014, Hughes et al., 2007).

Given the mixed results of the caregiver-child feeding practice literature and the limited research with child care teachers in particular, it is important to further evaluate the agreement between self-reported and observed measures with child care teachers. Thus, the purpose of this study was to compare self-reported and observed feeding practices in a group of Rhode Island (RI) Head Start teachers. Given the more structured environment of Head Start centers, we hypothesized that self-reported and observed feeding practices among child care teachers would be higher in agreement, compared to a home environment.