Maternal prompting types and child vegetable intake: Exploring the moderating role of picky eating

Appetite 146 (2020) 104518

Contents lists available at ScienceDirect

Appetite journal homepage: www.elsevier.com/locate/appet

Maternal prompting types and child vegetable intake: Exploring the moderating role of picky eating

T

Ariel A. Jordana, Danielle P. Appuglieseb, Alison L. Millerc, Julie C. Lumengd, Katherine L. Rosenblume, Megan H. Peschf,∗ a

Morehouse School of Medicine, 720 Westview Dr SW, Atlanta, GA, 30310, USA Appugliese Professional Advisors, P. O. Box 71, North Easton, MA, 02334, USA c Department of Health Behavior and Health Education, School of Public Health, University of Michigan, 1415 Washington Heights, Ann Arbor, MI, 48109, USA d Division of Developmental and Behavioral Pediatrics, Department of Pediatrics, University of Michigan, Center for Human Growth and Development, University of Michigan, Department of Nutritional Sciences, School of Public Health, University of Michigan, 300 N. Ingalls Street, 10th Floor, Ann Arbor, MI, 48109-5406, USA e Department of Psychiatry, Medical School, University of Michigan, Center for Human Growth and Development, University of Michigan. 4250 Plymouth Road, Rachel Upjohn Building, Ann Arbor, MI, 48109, USA f Division of Developmental and Behavioral Pediatrics, Department of Pediatrics, University of Michigan, Center for Human Growth and Development, University of Michigan, 300 N. Ingalls Street, 1111 SE, Ann Arbor, MI, 48109-5456, USA b

A R T I C LE I N FO

A B S T R A C T

Keywords: Picky eating Mothers Child Vegetables Prompts

Background: It is unknown if using different maternal prompting types is associated with vegetable intake in children perceived to be picky versus non-picky. Objectives: 1) To test the correlation of counts of maternal prompting types with child vegetable intake, and picky eating, 2) to examine the interaction of prompting types and picky eating status on vegetable intake. Design/methods: Low-income mother-child dyads (N = 199, mean child age 6.0 years) participated in a videotaped laboratory eating protocol with green beans, a familiar vegetable. A coding scheme was developed and reliably applied to categorize mothers’ prompting types. The prompting types were: Coercive Control (SubCategories: Reward and Pressure-to-Eat), Autonomy Promotion (Sub-Categories: Modeling, Reasoning, Praise, and Question), and Total Prompts (sum of all prompts). Mothers completed questionnaires. Bivariate analyses tested the association between counts of maternal prompting types with amount of green beans eaten, and picky eating. Regression analyses examined the interaction of picky eating status with counts of maternal prompting type on amount of green beans eaten. Results: Mothers used on average 1.66 prompts. Greater use of Coercive Control, Autonomy PromotionModeling, and Total Prompts were all inversely correlated with amount of green beans eaten. Greater use of Autonomy Promotion-Praise was directly correlated with amount of green beans eaten. In stratified models, greater use of Coercive Control prompts was negatively associated with amount of green beans eaten by the child in non-picky eaters, but not in picky eaters. There was no interaction between other prompting types and child picky eating status in predicting amount eaten. All p-values < 0.05. Conclusions: Mothers use different prompting types to encourage their children to eat vegetables depending on their picky eating status, most of which may be correlated with reduced intake.

1. Introduction Parental feeding practices, defined as specific behaviors parents use to influence what, when, and how much a child eats, can have varying effects on a child's mealtime behavior (Fries, Martin, & Van der Horst, 2017; Mallan et al., 2018) and whether a child is a picky eater (Jansen et al., 2017). Mealtimes can prove to be particularly frustrating for

parents, as picky eaters often end up crying and refusing to try the served food, leading to parental concern that their child is not meeting their recommended dietary needs (Fries et al., 2017; Reedy and KrebsSmith, 2010). Prior studies have looked at prompting as it relates to feeding particularly with mother-child dyads (Haycraft, Blissett, Maternal, & Paternal Controlling, 2008; Klesges et al., 1983) (J. Blissett, Bennett, Donohoe, Rogers, & Higgs, 2012), finding that there is a

∗

Corresponding author. E-mail addresses: [email protected] (A.A. Jordan), [email protected] (D.P. Appugliese), [email protected] (A.L. Miller), [email protected] (J.C. Lumeng), [email protected] (K.L. Rosenblum), [email protected] (M.H. Pesch). https://doi.org/10.1016/j.appet.2019.104518 Received 16 May 2019; Received in revised form 16 October 2019; Accepted 13 November 2019 Available online 14 November 2019 0195-6663/ © 2019 Elsevier Ltd. All rights reserved.

Appetite 146 (2020) 104518

A.A. Jordan, et al.

2. Materials and methods

significant correlation between maternal prompting types and child food intake. However, no studies to date have examined if prompting has a positive or negative association with child vegetable intake in picky versus non-picky eaters. Maternal prompting types, defined as practices mothers use to persuade their child to eat a healthy food (i.e. a vegetable) instead of an unhealthy food (i.e. a dessert), may influence the way the child perceives healthy foods and their willingness to eat healthy foods presented in the future (Pesch et al., 2016a,b). There are many different forms of maternal prompting types (Fries et al., 2017). One prompt type is Coercive Control, where a mother may compel her child to eat the healthy food by pressuring the child through threats or bribes (Vaughn et al., 2016). Another prompt type is Autonomy Promotion, where a mother provides limited structure, allowing the child to make age-appropriate decisions about trying the healthy food (Vaughn et al., 2016). Pressuring or prompting a child to eat healthy foods through coercion is thought to be an ineffective method, yet this method is often used if the child is perceived as a picky eater (Antoniou et al., 2015; Galloway, Fiorito, Francis, & Birch, 2006). Autonomy-Promoting prompts are thought to be more effective as they may teach the child decisionmaking skills and strengthen the mother-child relationship in a controlled, supportive environment (Steinsbekk, Bonneville-Roussy, Fildes, Llewellyn, & Wichstrom, 2017; Van der Kaap-Deeder et al., 2015). Few prior studies have examined the associations of these prompt types with child vegetable intake, especially differences among children with different eating temperaments (Edelson, Mokdad, & Martin, 2016; Fries et al., 2017; Lumeng & Burke, 2006). Children's eating temperaments, meaning the ways in which children are predisposed towards certain eating behaviors such as their food responsiveness, emotional eating or picky eating, have been shown to be associated with maternal feeding behaviors (Blissett, Bennett, Fogel, Harris, & Higgs, 2016; Faith & Hittner, 2016; Rodgers et al., 2013; Taylor, Wernimont, Northstone, & Emmett, 2015). Just as in general parenting, mothers may approach children with these differing eating-specific temperamental qualities differently in order to achieve their eating goals (Blissett et al., 2016). For instance, in order to persuade a child who is highly food responsive to eat their vegetables at dinnertime, a mother may offer the reward of dessert after the meal (Cooke, Chambers, Añez, & Wardle, 2011). This approach, however, may not be effective in a child with low food responsiveness. In the case of picky children, it is unknown if certain prompting types are more effective in promoting vegetable intake. For example, children who are picky may be more amenable to Autonomy Promotion prompts, but may be reluctant to comply when Coercive Control prompts are used because of a degree of anxiety or food dislike underlying their pickiness. If picky eating moderates the relationship between maternal prompting type and child vegetable intake, then identifying which maternal prompt types are associated with greater child vegetable intake for picky children can help create more effective interventions focused on appropriate execution of those prompting types during mealtimes. To our knowledge, the associations of different maternal prompting types with observed child vegetable intake in picky versus non-picky eaters are unknown. Therefore, the objectives of this study were 1) to identify the different maternal prompting types mothers give their children when presented with a familiar vegetable during a video recorded observational protocol, 2) to examine the association between maternal prompting types with child picky eating status and 3) to examine the moderating effect of picky eating status on the relationship between maternal prompt types and child vegetable intake. Based on prior work (Fries et al., 2017), we hypothesize that a greater number of maternal prompts in general will be associated with greater vegetable intake in non-picky eaters, but that there would be a lesser association in picky eaters. Furthermore, we hypothesize that a greater number of Autonomy Promotion prompts will be associated with greater vegetable intake in picky as compared to non-picky eaters.

2.1. Participants The participants selected for this study were 199 low-income childfemale primary caregiver dyads (mean child age 6.0 years) from Michigan who were originally enrolled in a larger longitudinal study from 2009 to 2011 examining contributors to children's risk of obesity. Participants in the original study (n = 380) were enrolled through their child's Head Start Program (a free, federally subsidized preschool program for low-income children) to participate in a study related to children's feeding behaviors. Exclusion criteria included the child having a gestational age of less than 35 weeks, significant perinatal or neonatal complications, serious medical problems or food allergies, disordered eating, foster care, not being the legal guardian of the child, or the mother not speaking English fluently and/or having a four-year college degree. Since all dyads were originally recruited from Head Start, the children were 3–4 years of age and living in low-income families at the time of recruitment. Of the sample in the original study, 95% of the female primary caregivers were biological mothers. The remaining 5% were grandmothers, adoptive mothers, and stepmothers; henceforth, we refer to the entire group as “mothers.” Approximately two years later, the dyads were invited to take part in a follow-up study, which was described as seeking to understand how mothers feed their children. A total of 296 dyads participated in this follow-up study, which included the Structured Eating Protocol (SEP), a laboratory-based eating interaction designed to capture children's and mothers' responses to different foods. Dyads were excluded (n = 50) from participating in this protocol if the mother had a food allergy or the child had a food allergy (which had developed since inclusion in the original study). An additional 3 dyads were not able to complete the SEP due to scheduling. Of the 243 who completed SEPs, 44 were excluded from this analysis for the following reasons: 1 for the child becoming ill during the protocol, 2 for the mother speaking a language other than English during the SEP, 8 for protocol violations, and 33 for inaccurate green bean weight (e.g. child throwing green beans in trash can) (See Fig. 1). This resulted in a final sample size of 199 dyads. The University of Michigan Institutional Review Board approved this study, mothers provided written informed consent, children provided assent, and dyads were compensated $60 for their participation. 2.2. Maternal prompts and child intake Maternal prompts of child food intake were measured in the SEP, which has been previously described in detail (Goulding et al., 2014; Pesch et al., 2016a,b; Radesky et al., 2015). The SEP was designed to

Fig. 1. Flow chart displaying participant inclusion and exclusion criteria. 2

Appetite 146 (2020) 104518

A.A. Jordan, et al.

Fig. 2. Hierarchical relationships in coding scheme categorizing maternal prompting types.

hypothesized that this familiar vegetable would elicit more prompts from the mothers than the other foods. Of the children, the majority (98.5%), were reported to be familiar with green beans, rating it on average 3.4 on a five point Likert scale, with higher ratings indicating more liking of the food. Portion sizes were identical for both mother and child. A coding scheme was developed to capture the prompts mothers gave to their children while eating a healthy food. The coding scheme was created using an iterative process over the course of several meetings between the first author and senior author. The development was informed by note-taking while watching 20 video segments from the SEP and review of the separate literatures around maternal prompting types (Fries et al., 2017; Vaughn et al., 2016). Prompts were defined as statements, utterances, questions, or phrases said by the mother to her child specifically about increasing food consumption. Physical prompts (e.g., the mother force feeding the child) were not included in this analysis as they were infrequent. Therefore, we focused only on the mothers’ verbal prompts. Coders first categorized each prompt into one of the two principal categories, Coercive Control or Autonomy Promotion. Coercive Control prompts were categorized as pressuring statements made by the mother to the child in an attempt to impose her will on the child to eat the specified healthy food (Vaughn et al., 2016). In contrast, Autonomy Promotion prompts were categorized as positive statements made by the mother to the child encouraging the child to eat the specified healthy food (Vaughn et al., 2016). Once a decision was made in selecting the appropriate principal category, the coder then assigned the prompt to the appropriate corresponding subcategory. The subcategories for Coercive Control were Pressure-to-Eat and Reward. Pressure-to-Eat included providing repeated commands to eat (e.g., “Eat your green beans!”). Reward included food-related and non-foodrelated rewards, which the child can earn if she/he complies with the mother's request to eat the specified food (e.g., “If you eat the green beans, you can watch an extra hour of TV today!”). The subcategories for Autonomy Promotion were Modeling, Praise, Reasoning, and Question Prompts. Modeling included the mother actively demonstrating and narrating healthy eating for the child (e.g., [Mother tries green beans]-“Try them with me, sweetie.”). Praise included the mother

capture the mother's and child's responses to various foods in a controlled laboratory setting, thereby reducing the variability of home mealtime observations (e.g. sibling interference, mother attending to other family members, food preparation, etc.). The SEP has shown strong test-retest reliability across approximately 2.5 years, with correlations for the amount consumed by children for vegetables (r = 0.31) and total amount of food consumed (r = 0.43; p-values < .05 for all statistics reported) (Pesch, Miller, Appugliese, Rosenblum, & Lumeng, 2016). Validity of the SEP is supported by correlations between mothers' prompts of their children's intake during the SEP, as measured by the Bob and Tom's Method of Assessing Nutrition (Klesges et al., 1983), and mother's feeding practices and beliefs as measured by the Caregiver Feeding Style Questionnaire (CFSQ) (Hughes et al., 2005) and the Preschooler Feeding Questionnaire (PFQ) (Baughcum et al., 2001) (Pesch et al., 2016a,b). For example, higher rates of maternal Total Prompts during the SEP correlated with more maternal demandingness (r = 0.26) as measured on the CFSQ (Pesch et al., 2016a,b). The SEP has also been validated as a measure of observationally measured picky eating behavior against maternal report on the Children's Eating Behavior Questionnaire (Wardle, Guthrie, Sanderson, & Rapoport, 2001) and the Food Neophobia Scale (Pliner & Hobden, 1992) (Fernandez et al., 2018). Each mother-child dyad was asked to fast for two hours prior to the SEP. During the protocol, the dyad was seated at a table alone in a quiet room and videotaped throughout the procedure. Mothers were asked if their child had ever eaten each food before to assess familiarity with the food. A research assistant presented both the mother and child with individualized portions of four different foods that were offered one at a time, sequentially, and in randomized order, to the dyad concurrently. Each food was presented individually and removed after the allotted time period. The mother and child were invited to try the food if they wanted, and then were left alone for 4 min. Each food portion was weighed before and after the protocol to calculate the amount consumed by each participant. The foods varied based on familiarity and sweetness, and included halva (unfamiliar dessert), artichoke (unfamiliar vegetable), chocolate cupcakes (familiar dessert), and green beans (familiar vegetable). This analysis focused only on the fourminute segment in which the green beans were presented, as it was

3

Appetite 146 (2020) 104518

A.A. Jordan, et al.

measured using the six-item Food Fussiness subscale (CEBQ FF) (Wardle et al., 2001) (baseline Cronbach's α = 0.91). Items were answered on a Likert scale (range = 1–5) and averaged such that a higher score indicated greater picky eating. The mean CEBQ FF score for the entire study sample was calculated, which was 2.7, falling just below the moderately picky eating range based on prior work (Steinsbekk et al., 2017). Children whose CEBQ FF score was below the mean were considered as less picky, which we term “non-picky”, whereas children whose CEBQ FF score was equal to or greater than the mean were considered pickier, which we term “picky.”

Table 1 Characteristics of participant anthropometrics, demographics and maternal prompting types, (N = 199). Total sample (N = 199)

Participant characteristics

Non-Picky Eaters (FF ≤ mean, n = 100)

Picky Eaters (FF > mean, n = 99)

n (%) or mean (SD)

Child characteristics Age (years); mean, SD 6.0 (0.7) Child's sex; n (%) Male 98 (49.3) Female 101 (50.8) 0.9 (1.0) Body Mass Index (BMI) z-score; mean (SD) Categorical weight status; n (%) Underweight 3 (1.5) Normal weight 108 (54.6) Overweight 44 (22.2) Obese 43 (21.7) CEBQ FF 2.7 (0.76)

6.1 (0.6)

5.9 (0.7)

41 (41.0) 59 (59.0) 1.0 (1.1)

57 (57.6) 42 (42.4) 0.7 (0.9)

2 (2.0) 49 (49.0) 19 (19.0) 30 (30.0) 2.1 (0.5)

1 (1.0) 59 (60.2) 25 (25.5) 13 (13.3) 3.3 (0.4)

Mother characteristics Age (years); mean, SD 31.3 (7.2) 32.0 (7.2) Mother BMI; mean, SD 33.2 (9.6) 34.1 (9.7) Mother race/ethnicity; n (%) White non-Hispanic 143 (71.9) 69 (69.0) Not white, non56 (28.1) 31 (31.0) Hispanic or other Highest level of maternal education; n (%) ≤ high school 90 (45.2) 45 (45.0) diploma > high school 109 (54.8) 55 (55.0) diploma CES-D total score of 16 or above; mean, SD < 16 131 (65.8) 66 (66.0) ≥16 68 (34.2) 34 (34.0)

2.4. Mother and child characteristics Mothers reported child age and sex and their own race/ethnicity and education level. Mothers completed the Center for Epidemiologic Studies-Depression scale (20 items, Cronbach's α = 0.91), a valid and reliable measure that assesses mothers' symptoms of depression (Radloff, 1977). Clinically significant maternal depressive symptomatology was defined as a score greater than or equal to 16. Heights and weights of the mothers and children were measured according to standardized procedures (Shorr, 1986). Body mass index (BMI) was calculated as weight in kilograms divided by height in meters squared. For 13 (5.5%) mothers who were pregnant or had recently given birth, self-reported pre-pregnancy weight was used instead of measured weight. Children were categorized by weight status based on the United States Center for Disease Control and Prevention growth charts (Kuczmarski et al., 2000, pp. 1–27).

30.6 (7.2) 32.3 (9.3) 74 (74.8) 25 (25.3)

45 (45.5)

2.5. Statistical analysis

54 (54.6)

Univariate statistics were computed for all variables. Bivariate analyses examined counts of maternal prompting types and amount of green beans eaten by child picky eating status (picky vs. non-picky children). Correlations of each maternal prompting type with amount of green beans eaten by the child, and child picky eating (as reflected by the continuous CEBQ FF score) were calculated. Five separate linear regression models were constructed to test the interactive effects of the 5 maternal prompt types (Coercive Control-Total, Autonomy Promotion-Total, Total Prompts, Autonomy Promotion-Proportion, Coercive Control-Proportion) with picky eating status (CEBQ FF ≤ mean vs. not) on amount of green beans eaten by the child, controlling for child age and child sex. For those models in which the interaction of the prompt type and picky eating status was significant, we ran additional models stratified on picky eating status. A p-value of < .05 was considered statistically significant.

65 (65.7) 34 (34.3)

CEBQ FF stands for Children's Eating Behavior Food Fussiness Subscale, SD stands for standard deviation.

complementing the child for voluntarily trying the specified food or complying with a maternal prompt to try the food (e.g., “Thank you for trying the green beans!”). Reasoning included the mother using logic to convince the child why he/she should eat the specified food (e.g., “If you eat your green beans, you'll grow big and strong!”). Question Prompts included allowing the child to choose whether or not to try the food being offered, (e.g., “Would you like to try the green beans?”) (See Fig. 2). Reliability was calculated based on the counts of each code (Coercive Control-Pressure-to-Eat (Cohen's kappa = 0.84), Coercive Control-Reward (kappa = 0.87), Autonomy Promotion-Modeling (kappa = 0.88), Autonomy Promotion-Praise (kappa = 0.91), Autonomy Promotion-Reasoning (kappa = 0.87), and Autonomy Promotion-Question Prompts (kappa = 0.83) across each video segment. Counts of Coercive Control prompts and Autonomy Promotion prompts were summed for each dyad, creating the 6 variables that we call Coercive Control-Pressure-to-Eat, Coercive Control-Reward, Autonomy Promotion-Modeling, Autonomy Promotion-Praise, Autonomy Promotion-Reasoning, and Autonomy Promotion-Question Prompts (Fig. 2). The sum of each principal category and corresponding subcategories generated the two variables Coercive Control-Total and Autonomy Promotion-Total.

3. Results Characteristics of the participants are provided in Table 1. Just over half (50.8%) of the children were female and 21.7% were obese. Mean CEBQ FF score was 2.70 (SD 0.76) in the entire sample, and 3.3 in the group of picky children as compared to 2.1 in the non-picky children. Of the mothers, 71.9% were non-Hispanic white and over half (54.8%) had a high school education or less. The average maternal BMI was 33.2, which is in the obese weight status range. Mothers used on average 1.66 prompts during the 4-min segment, ranging from 0 to 16 prompts. Mothers most often used Coercive Control-Pressure-to-Eat (M = 0.94, SD 1.9) prompts, followed by Autonomy Promotion-Question prompts (M = 0.20, SD 0.50) and Autonomy Promotion-Modeling (M = 0.20, SD 0.6) (Table 2). Mothers less frequently used Autonomy Promotion-Reasoning (M = 0.17, SD 0.6), Autonomy Promotion-Praise (M = 0.15, SAD 0.5) and Coercive Control-Reward (M = 0.01, SD 0.01), (Table 2). With regard to the proportion of prompt types used, mothers used on average half Autonomy Promotion prompts (Autonomy Promotion–Proportion M = 0.50, SD 0.43), and half Coercive Control prompts (Coercive

2.3. Child picky eating status Mothers completed the 35-item Children's Eating Behavior Questionnaire (CEBQ). For this analysis, child picky eating was 4

Appetite 146 (2020) 104518

A.A. Jordan, et al.

Table 2 Univariate statistics and bivariate analyses of prompting types and amount of green beans eaten by child picky eating status, (N = 199). Prompting Types

Coercive Control-Reward Coercive Control-Pressure-to-Eat Coercive Control-Total Autonomy Promotion-Modeling Autonomy Promotion-Praise Autonomy Promotion-Reasoning Autonomy Promotion-Question Autonomy Promotion-Total Total Number of Prompts (Autonomy Promotion + Coercive Control) Coercive Control - Proportion Autonomy Promotion - Proportion Amount of Green Beans child ate (grams)a

Total Sample (n = 199)

Non-Picky Eaters (FF ≤ mean, n = 100)

Picky Eaters (FF > mean, n = 99)

Mean

SD

Mean

SD

Mean

SD

0.01 0.94 0.93 0.20 0.15 0.17 0.20 0.72 1.66

0.1 1.92 1.93 0.62 0.48 0.55 0.52 1.21 2.69

0.01 0.54 0.55 0.07 0.15 0.11 0.17 0.5 1.05

0.1 1.04 1.06 0.29 0.54 0.42 0.49 0.93 1.45

0.01 1.34 1.32 0.33 0.15 0.23 0.23 0.95 2.27

0.1 2.45 2.46 0.81 0.41 0.65 0.55 1.42 3.42

.99 < .0001 < .0001 < .0001 .98 .04 .33 .0002 < .0001

0.50 0.50 26.00

0.43 0.43 28.73

0.49 0.51 34.02

0.45 0.45 33.5

0.51 0.49 17.91

0.43 0.43 20.06

0.92 0.92 < .0001

p valuea

FF stands for Children's Eating Behavior Questionnaire Food Fussiness subscale score, SD stands for Standard Deviation. a Wilcoxon test used as the Amount of Green Beans eaten was not normally distributed.

The interactive effects of counts of maternal prompt types with picky eating status on amount of green beans eaten are shown in Table 4. The interaction between Coercive Control-Total and Picky Eating is significant (p = .01). For picky eaters, Coercive Control-Total was not associated with amount eaten (p = .13). For non-picky eaters, Coercive Control-Total was negatively associated with amount of green beans eaten (p = .01). The interaction between Autonomy PromotionTotal and Picky Eating was not significant (p = .96). The interaction between Total Prompts and Picky Eating was not significant (p = .06). The interactions between Coercive Control-Proportion with Picky Eating and Autonomy Promotion-Proportion with Picky Eating were also not significant (p = .09 and .09, respectively).

Control-Proportion, M = 0.50, SD = 0.43). Mothers in general used differing counts of the six prompting types with their picky vs. non-picky children (Table 2). Mothers of picky (vs. non-picky) children used more Coercive Control- Pressure-to-Eat prompts, more Coercive Control-Total Prompts, more Autonomy Promotion-Modeling, Autonomy Promotion-Reasoning, Autonomy Promotion-Total and more Total Number of Prompts (all p's < 0.05). There was no difference in the proportion of Autonomy Promotion Prompts or Coercive Control Prompts used between picky and nonpicky children (p = .92, and .92 respectively). Non-picky children ate double the amount of green beans than their picky counter parts (34 vs 17 g on average, p < .0001). Results of the analyses evaluating the correlations of prompting types with child green bean intake and CEBQ Food Fussiness (Picky Eating) are shown in Table 3. Amount of green beans eaten was negatively correlated with Coercive Control-Pressure-to-Eat (r = −0.21), negatively correlated with Coercive Control-Total (r = −0.20), and negatively correlated with Autonomy Promotion-Modeling (r = −0.17). Amount of green beans eaten was positively correlated with Autonomy Promotion-Praise (r = 0.24). Picky eating was positively correlated with Coercive Control-Pressure-to-Eat (r = 0.16), positively correlated with Autonomy Promotion-Modeling (r = 0.17), and positively correlated with Autonomy Promotion-Total (r = 0.16). The Coercive Control –Proportion was inversely correlated to the amount of green beans consumed by the child, whereas Autonomy Promotion – Proportion was positive correlated. Overall, total number of prompts was negatively correlated with amount of green beans eaten (r = −0.19) and positively correlated with picky eating (r = 0.18).

4. Discussion To our knowledge, this is the first report describing the associations between maternal prompting types and child familiar vegetable intake in picky versus non-picky eaters. We found that mothers engaged in prompting fairly often, and used an average of 1.66 prompts during the 4-min protocol segment. The most frequently used prompt type was Coercive Control-Pressure-to-Eat (e.g. “Eat your green beans!”). Mothers of picky vs. non-picky eaters used more Total Prompts, as well as more of almost all different prompting types, except Coercive Control-Reward, Autonomy Promotion-Praise and Autonomy Promotion-Question. The amount of vegetables eaten by the child was negatively correlated with Coercive Control-Pressure-to-Eat and Autonomy Promotion-Modeling. Autonomy Promotion-Praise was the only prompt type positively correlated with greater amount of

Table 3 Correlation coefficients and p-values for prompt types in relation to child green bean intake, and picky eating, (N = 199). Prompting Types

Amount of green beans child ate (g)

Picky eating (CEBQ FF continuous)

Coercive Control-Reward Coercive Control-Pressure-to-Eat Coercive Control-Total Autonomy Promotion-Modeling Autonomy Promotion-Praise Autonomy Promotion-Reasoning Autonomy Promotion-Question Autonomy Promotion-Total Total Prompts Coercive Control - Proportion Autonomy Promotion - Proportion

Correlation Coefficient -.03 -.21* -.21* -.17* .24** -.10 -.13 -.09 -.19* -.23* .23*

.02 .16* .15* .17* .03 .13 .01 .16* .18* .003 -.003

*p < .05 **p ≤ .01. 5

Appetite 146 (2020) 104518

A.A. Jordan, et al.

Table 4 Interactive effects in linear models of associations between counts of maternal prompts and child picky eating status on child vegetable intake. Prompting type

Sample (N)

Variable

Estimate

SE

Lower CI

Upper CI

Wald Chisquare

P-value

Coercive Control-Total

Whole sample (N = 199)

Coercive Control-Total CEBQ: Food Fussiness > mean (no vs yes) Child age (years) Child sex (Male vs. Female) Coercive Control-Total*CEBQ FF > mean Coercive Control-Total Child age (years) Child sex (Male vs. Female) Coercive Control-Total Child age (years) Child sex (Male vs. Female) Autonomy Promotion- Total CEBQ: Food Fussiness > mean (no vs yes) Child age (years) Child sex (Male vs. Female) Autonomy Promotion-Total*CEBQ FF > mean Total Prompts CEBQ: Food Fussiness > mean (no vs yes) Child age (years) Child sex (Male vs. Female) Total Prompts-Total*CEBQ FF > mean Autonomy Promotion-Proportion CEBQ: Food Fussiness > mean (no vs yes) Child age (years) Child sex (Male vs. Female) Autonomy Promotion-Proportion *CEBQ FF > mean Coercive Control-Proportion CEBQ: Food Fussiness > mean (no vs yes) Child age (years) Child sex (Male vs. Female) Coercive Control-Proportion *CEBQ FF > mean

−1.21 17.87 −0.06 −7.67 −6.93 −8.13 0.00 −10.38 −1.29 −0.09 −5.11 −1.29 14.39 0.07 −7.72 −0.19 −0.76 18.04 0.00 −7.53 −3.90 5.42 3.39 −0.20 −3.09 17.78

1.11 4.33 0.24 3.88 2.76 3.11 0.47 6.71 0.82 0.23 4.01 0.82 4.59 0.25 3.97 3.55 0.80 4.69 0.25 3.92 2.02 6.92 6.91 0.27 4.59 10.37

−3.38 9.38 −0.54 −15.28 −12.33 −14.22 −0.91 −23.53 −2.90 −0.55 −12.98 −2.90 5.40 −0.42 −15.50 −7.15 −2.34 8.85 −0.49 −15.20 −7.85 −8.14 −10.15 −0.72 −12.09 −2.54

0.95 26.35 0.42 −0.06 −1.52 −2.04 0.91 2.77 0.32 0.37 2.75 0.32 23.39 0.56 0.06 6.77 0.82 27.22 0.48 0.15 0.05 18.98 16.92 0.32 5.91 38.11

1.20 17.02 0.07 3.90 6.31 6.85 0.00 2.39 2.48 0.15 1.62 2.48 9.84 0.08 3.79 0.00 0.89 14.82 0.00 3.69 3.74 0.61 0.24 0.57 0.45 2.94

.27 < .0001 .80 .05 .01 .01 1.00 .12 .12 .70 .20 .12 .002 .78 .05 .96 .34 .001 .99 .05 .05 .43 .62 .45 .50 .09

−5.42 21.17 −0.20 −3.09 −17.78

6.92 6.93 0.27 4.59 10.37

−18.98 7.58 −0.72 −12.09 −38.11

8.14 34.75 0.32 5.91 2.54

0.61 9.33 0.57 0.45 2.94

.43 .002 .45 .50 .09

Non-picky (N = 100)

Picky (N = 99)

Autonomy Promotion-Total

Whole sample (N = 199)

Total Prompts

Whole sample (N = 199)

Autonomy PromotionProportion

Whole sample (N = 199)

Coercive Control-Proportion

Whole sample (N = 199)

SE stands for standard error, CEBQ stands for Children's Eating Behavior Questionnaire, CI stands for confidence interval.

be effective in either picky or non-picky children. Current guidelines, however, do not provide guidance for mothers on how to appropriately encourage their child to eat healthy foods (Mitchell, Farrow, Haycraft, & Meyer, 2013). Mothers need guidance on how to effectively encourage their child to eat healthy foods, especially children who are picky, rather than simply being told using coercive strategies is inappropriate and ineffective. Our study found that children who are picky did not have greater intake of familiar vegetables with more frequent Coercive Control prompting, and that non-picky children actually ate fewer vegetables with more Coercive Control prompts from their mothers. It may be that children who are picky are not impacted by their mother's prompts because of their strong “eating temperament” (Steinsbekk et al., 2017). If prompting is ineffective at promoting more vegetable intake in picky children, parents may be able to decrease their prompts, and thereby mealtime conflict, with the same intake results. There is still a need for further research into which types of prompt types are associated with greater child vegetable intake without the need to use threats and bribes, especially in picky children. The finding that greater Coercive Control prompts was associated with decreased green bean intake in non-picky children may suggest that children who are not picky may in fact “shut down” with Coercive Control prompts. In addition, we found that children whose mothers used a greater proportion of Coercive Control prompts have lower intake of vegetable. This finding is supported by prior studies (Fries et al., 2017; Galloway et al., 2006), hypothesizing that this more direct and sometimes negative approach to prompting can cause a paradoxical effect in vegetable intake. Due in part to these prior works, Autonomy Promotion prompting approaches have been recommended (Soenens et al., 2018; Van der Kaap-Deeder et al., 2015). Our study found that a greater proportion of Autonomy Promotion prompts was associated

vegetable intake. While there was no association between Coercive Control prompts and amount of vegetable consumed in picky eaters, non-picky eaters ate less with greater Coercive Control prompts. Of note, a greater proportion of Autonomy Promotion prompts was correlated with greater intake of vegetable in the sample as a whole; this relationship did not differ among picky vs. non-picky eaters. The inverse was true of Coercive Control – Proportion. The finding that mothers used both Coercive Control and Autonomy Promotion prompt types during the meal, and most commonly Coercive Control-Pressure-to-Eat, is consistent with prior work (Fries et al., 2017). Prior literature has shown that mothers most often use Coercive Control prompt types with little positive response from their child (Galloway et al., 2006; Gregory, Paxton, & Brozovic, 2010). During this 4-min segment with green beans, in both picky and non-picky eaters, mothers most often used Coercive Control-Pressure-to-Eat prompts, which was associated with lower intake of vegetables. These results may be interpreted in two different ways. First, it may be that Coercive Control-Pressure-to-Eat prompts are ineffective in encouraging children to eat more vegetables regardless of their picky eating status. Second, it may also be that mothers of children who do not prefer vegetables, such as green beans, may be more apt to prompt their children in a more direct and pressuring manner, as they may believe that gentler approaches are less effective. This, however, is not supported by prior work (Edelson et al., 2016) which has found that modeling is the most successful approach to getting a child to eat a target food. Prior studies have found that mothers are more apt to use pressuring strategies with children who are picky and those who are thinner (Jansen et al., 2017; Quah et al., 2019; Webber, Cooke, Hill, & Wardle, 2010). Mothers may also be motivated to engage in these behaviors in an attempt to promote a healthy lifestyle for their child by pressuring consumption of more healthful foods, although this study suggests that this approach may not 6

Appetite 146 (2020) 104518

A.A. Jordan, et al.

eating the green beans in fear of appearing to have limited control over their child's actions. Additional limitations include the study sample's make up of low-income mothers and early school-aged children from a single geographic location without a college degree, the majority of whom had white non-Hispanic race/ethnicity. Therefore, results may not be generalizable to other populations. Finally, the cross-sectional design limits any inferences about temporality or causation. Future longitudinal studies are needed to examine the association between maternal and paternal prompting types and child vegetable intake, taking into account socioeconomic and child picky eating status. Additional work should explore these associations in a variety of mealtime situations to more fully understand the dynamics of dyads during feeding interactions, as well as with unfamiliar foods.

with greater child intake of vegetable, however this relationship was not including the absolute number of Autonomy Promotion prompts. It may be that the proportion of a mother's prompts that are autonomy promoting has a greater impact on the child's intake than the absolute number of these prompts. A correlation between number of Autonomy Promotion-Praise prompts and amount of green beans consumed was found, however as with all prompting types, this work is unable to determine the timing of the Autonomy Promotion-Praise prompts with child intake. It may be that children who ate more green beans received more praise from their mothers, rather than the Autonomy PromotionPraise prompts encouraging more intake. It may also be that both of these pathways are somewhat true and that the relationship between child vegetable intake and Autonomy Promotion-Praise prompts is bidirectional. Future work examining the timing of maternal prompting types in relation to child bites of vegetables is necessary to disentangle these relationships. Coercive Control prompt types and Autonomy Promotion prompt types are strikingly different and may reflect more general parenting styles that could have varying impacts on emotional and cognitive development (Crouch et al., 2017; Soenens et al., 2018). Mothers may first use an Autonomy Promotion prompt type, like Question Prompts, when the child is first exposed to the meal but then resort to Coercive Control prompt types when she perceives that her child is refusing to be compliant. It is possible the child is deciding to eat very little or not at all in response to this harsher prompting type (Jani, Mallan, & Daniels, 2015; Jansen et al., 2017). Prior literature (Sanders, Patel, Le Grice, & Shepherd, 1993) has shown that mothers' aversive behaviors are correlated with the degree of a picky eaters' noncompliance. Perhaps this could be tested as an empirical question. Are picky eater children choosing to be less compliant because their mothers are using more threatening prompt types? Or will picky eaters choose to be noncompliant regardless of their mother's prompt types? A better understanding of the associations of coercive prompt types with children's compliance and emotional development is an important direction of future experimental studies for effective interventions. It should also be considered that the associations of maternal prompts with greater child vegetable intake might be related to their timing in feeding situations. For example, a child may be more likely to respond to a prompt at the beginning of the meal when he or she is considering whether or not to try the food. It could be that the timing of the prompt, and not the actual prompt type itself, may result in greater compliance. Future studies should examine the association of timing of maternal prompts with greater child vegetable intake, in addition to the prompt type in picky versus non-picky eaters. The order in which the green beans were presented during the SEP protocol could also have impacted the amount eaten. During the SEP the green beans were presented in a random order among three other foods. Prior work has shown that individuals consume more of a food presented first in isolation, whether it was a dessert or vegetable (Redden et al., 2015). Future work should examine the effects of the order of food presentation on child intake in picky and non-picky eaters in both laboratory and ecologically valid settings. Strengths of this study include a large sample size and the observational nature of data collection. Our mean cut off used for classification of picky eating was slightly lower than in prior work (Steinsbekk, Sveen, Fildes, Llewellyn, & Wichstrøm, 2017), which may impact results. While the laboratory setting of this study provided a carefully controlled environment, it also was a somewhat atypical eating environment, thereby reducing ecological validity. An example of this is the fact that Coercive Control-Reward was rarely seen in this study, possibly because no immediate rewards were present (e.g. the reward of offering dessert after finishing the green beans). Others have found that parents use food and non-food related rewards in order to encourage their child to eat a desired healthy food (Roberts, Marx, & Musher-Eizenman, 2018). However, due to the setting, mothers may have felt uncomfortable offering a reward to their child in exchange for

5. Conclusions Use of Coercive Control-Pressure-to-Eat prompts was most common in this study, but was associated with lower vegetable intake in nonpicky eaters. A higher proportion of Autonomy Promotion prompts was correlated with greater vegetable intake in children in general, regardless of picky eating status. Children who are picky may not be influenced by mothers’ prompts to eat more vegetables. Future studies should consider the associations of maternal prompts with vegetable intake in children considered picky and non-picky eaters, particularly investigating child factors associated with an effective command to eat more and independent decisions to eat healthy foods. Funding This work was funded by the Michigan Institute for Clinical & Health Research (UL1TR002240), an American Heart Association Fellow-to-Faculty Transition award (17FTF33630183) to MP, and a National Institute of Child Health and Human Development at the National Institutes of Health grant (R01 HD061356) to JL. AJ was supported by 2UL1TR000433. Appendix A. Supplementary data Supplementary data to this article can be found online at https:// doi.org/10.1016/j.appet.2019.104518. References Antoniou, E., Roefs, A., Kremers, S., Jansen, A., Gubbels, J., Sleddens, E., et al. (2015). Picky eating and child weight status development: A longitudinal study. Journal of Human Nutrition and Dietetics, 29, 298–307. Baughcum, A., Powers, S., Johnson, S., Chamberlin, L., Deeks, C., Jain, A., et al. (2001). Maternal feeding practices and beliefs and their relationships to overweight in early childhood. Journal of Developmental and Behavioral Pediatrics, 22, 391–408. Blissett, J., Bennett, C., Donohoe, J., Rogers, S., & Higgs, S. (2012). Predicting successful introduction of novel fruit to preschool children. Journal of the Academy of Nutrition and Dietetics, 112(12), 1959–1967. Blissett, Bennett, C., Fogel, A., Harris, G., & Higgs, S. (2016). Parental modelling and prompting effects on acceptance of a novel fruit in 2–4-year-old children are dependent on children's food responsiveness. British Journal of Nutrition, 115(3), 554–564. Cooke, L. J., Chambers, L. C., Añez, E. V., & Wardle, J. (2011). Facilitating or undermining? The effect of reward on food acceptance. A narrative review. Appetite, 57(2), 493–497. Crouch, J., Irwin, L., Milner, J., Skrowonski, J., Rutledge, E., & Davila, A. (2017). Do hostile attributions and negative affect explain the association between authoritarian beliefs and harsh parenting? Child Abuse & Neglect, 67, 13–21. Edelson, L. R., Mokdad, C., & Martin, N. (2016). Prompts to eat novel and familiar fruits and vegetables in families with 1–3 year-old children: Relationships with food acceptance and intake. Appetite, 99, 138–148. Faith, M. S., & Hittner, J. B. (2016). Shadows of temperament in child eating patterns: Implications for family and parenting research. Oxford University Press. Fernandez, C., DeJesus, J. M., Miller, A. L., Appugliese, D. P., Rosenblum, K. L., Lumeng, J. C., et al. (2018). Selective eating behaviors in children: An observational validation of parental report measures. Appetite, 127, 163–170. Fries, L., Martin, N., & Van der Horst, K. (2017). Parent-child mealtime interactions associated with toddlers' refusals of novel and familiar foods. Physiology & Behavior,

7

Appetite 146 (2020) 104518

A.A. Jordan, et al.

Longitudinal analysis between maternal feeding practices and body mass index (BMI): A study in Asian Singaporean preschoolers. Frontiers in Nutrition, 6, 32. Radesky, J., Miller, A. L., Rosenblum, K. L., Appugliese, D., Kaciroti, N., & Lumeng, J. C. (2015). Maternal mobile device use during a structured parent–child interaction task. Academic Pediatrics, 15, 238–244. Radloff, L. (1977). The CES-D scale a self-report depression scale for research in the general population. Applied Psychological Measurement, 1, 385–401. Redden, J. P., Mann, T., Vickers, Z., Mykerezi, E., Reicks, M., & Elsbernd, S. (2015). Serving first in isolation increases vegetable intake among elementary schoolchildren. PLoS One, 10(4) e0121283. Reedy, J., & Krebs-Smith, S. (2010). Dietary sources of energy, solid fats, and added sugars among children and adolescents in the United States. Journal of the American Dietetic Association, 110, 1477–1484. Roberts, L., Marx, J., & Musher-Eizenman, D. (2018). Using food as a reward: An examination of parental reward practices. Appetite, 120, 318–326. Rodgers, R. F., Paxton, S. J., Massey, R., Campbell, K. J., Wertheim, E. H., Skouteris, H., et al. (2013). Maternal feeding practices predict weight gain and obesogenic eating behaviors in young children: A prospective study. International Journal of Behavioral Nutrition and Physical Activity, 10(1), 24. Sanders, M., Patel, R., Le Grice, B., & Shepherd, R. (1993). Children with persistent feeding difficulties: An observational analysis of the feeding interactions of problem and non-problem eaters. Health Psychology, 12, 64–73. Shorr, I. (1986). How to weight and measure children. New York: United Nations. Soenens, B., Park, S., Mabbe, E., Vansteenkiste, M., Chen, B., Petegem, S., et al. (2018). The moderating role of vertical collectivism in South-Korean adolescents' perceptions of and responses to autonomy-supportive and controlling parenting. Frontiers in Pscychology, 9, 1080. Steinsbekk, S., Bonneville-Roussy, A., Fildes, A., Llewellyn, C., & Wichstrom, L. (2017). Child and parent predictors of picky eating from preschool to school age. Journal of Behavioral Nutrition and Physical Activity, 14, 1–8. Taylor, C. M., Wernimont, S. M., Northstone, K., & Emmett, P. M. (2015). Picky/fussy eating in children: Review of definitions, assessment, prevalence and dietary intakes. Appetite, 95, 349–359. Van der Kaap-Deeder, J., Vansteenkiste, M., Soenens, B., Loeys, T., Mabbe, E., Gargurevich, R., et al. (2015). Autonomy-supportive parenting and autonomy-supportive sibling interactions. Personality and Social Psychology Bulletin, 41, 1590–1604. Vaughn, A., Ward, D., Fisher, J., Faith, M., Hughes, S., Kremers, S., et al. (2016). Fundamental constructs in food parenting practices: A content map to guide future research. Nutrition Reviews, 74, 98–117. Wardle, J., Guthrie, C. A., Sanderson, S., & Rapoport, L. (2001). Development of the children's eating behaviour questionnaire. Journal of Child Psychology and Psychiatry, 42(07), 963–970. Webber, L., Cooke, L., Hill, C., & Wardle, J. (2010). Associations between children's appetitive traits and maternal feeding practices. Journal of the American Dietetic Association, 110(11), 1718–1722.

176, 93–100. Galloway, A., Fiorito, L., Francis, L., & Birch, L. (2006). ‘Finish your soup’: Counterproductive effects of pressuring children to eat on intake and affect. Appetite, 46, 318–323. Goulding, A., Rosenblum, K., Miller, A., Peterson, K., Chen, Y., Kaciroti, N., et al. (2014). Associations between maternal depressive symptoms and child feeding practices in a cross-sectional study of low-income mothers and their young children. International Journal of Behavioral Nutrition and Physical Activity, 11, 75. Gregory, J., Paxton, S., & Brozovic, A. (2010). Pressure to eat and restriction are associated with child eating behaviours and maternal concern about child weight, but not child body mass index, in 2- to 4-year-old children. Appetite, 54, 550–556. Haycraft, E., & Blissett, J. (2008). Maternal and paternal controlling feeding practices: Reliability and relationships with BMI. Obesity, 16, 1552. Hughes, S., Power, T., Orlet Fisher, J., Mueller, S., & Nicklas, T. (2005). Revisiting a neglected construct: Parenting styles in a child-feeding context. Appetite, 44, 83–92. Jani, R., Mallan, K., & Daniels, L. (2015). Association between Australian-Indian mothers' controlling feeding practices and children's appetite traits. Appetite, 84, 188–195. Jansen, P., de Barse, L., Jaddoe, V., Verhulst, F., Franco, O., & Tiemeier, H. (2017). Bidirectional associations between child fussy eating and parents' pressure to eat: Who influences whom? Physiology & Behavior, 176, 101–106. Klesges, R., Coates, T., Brown, G., Sturgeon-Tillisch, J., Moldenhauer-Klesges, L., Holzer, B., et al. (1983). Parental influences on children's eating behavior and relative weight. Journal of Applied Behavior Analysis, 16, 371–378. Kuczmarski, R. J., Ogden, C. L., Grummer-Strawn, L. M., Flegal, K. M., Guo, S. S., Wei, R., et al. (2000). CDC growth charts: United States. Advance Data. Lumeng, J., & Burke, L. (2006). Maternal prompts to eat, child compliance, and mother and child weight status. The Journal of Pediatrics, 149, 330–335. Mallan, K., Jansen, E., Harris, H., Llewellyn, C., Fildes, A., & Daniels, L. (2018). Feeding a fussy eater: Examining longitudinal bidirectional relationships between child fussy eating and maternal feeding practices. Journal of Pediatric Psychology, 1–8. Mitchell, G., Farrow, C., Haycraft, E., & Meyer, C. (2013). Parental influences on children's eating behaviour and characteristics of successful parent-focused interventions. Appetite, 60, 85–94. Pesch, M., Appugliese, D., Kaciroti, N., Rosenblum, K., Miller, A., & Lumeng, J. (2016). Maternal encouragement and discouragement: Differences by food type and child weight status. Appetite, 101, 15–22. Pesch, M., Miller, A., Appugliese, D., Kaciroti, N., Rosenblum, K., & Lumeng, J. (2016). Low-income mothers' feeding goals predict observed home mealtime and child feeding practices. Child: Care, Health and Development, 42, 934–940. Pesch, M. H., Miller, A. L., Appugliese, D. P., Rosenblum, K. L., & Lumeng, J. C. (2016). Affective tone of mothers' statements to restrict their children's eating. Appetite, 103, 165–170. Pliner, P., & Hobden, K. (1992). Development of a scale to measure the trait of food neophobia in humans. Appetite, 19(2), 105–120. Quah, P. L., Chun, J., Fries, L. R., Chan, M. J., Aris, I. M., Lee, Y. S., et al. (2019).

8