Using food as a reward: An examination of parental reward practices

Accepted Manuscript Using food as a reward: An examination of parental reward practices Lindsey Roberts, Jenna M. Marx, Dara R. Musher-Eizenman PII:

S0195-6663(17)30094-6

DOI:

10.1016/j.appet.2017.09.024

Reference:

APPET 3626

To appear in:

Appetite

Received Date: 19 January 2017 Revised Date:

1 September 2017

Accepted Date: 22 September 2017

Please cite this article as: Roberts L., Marx J.M. & Musher-Eizenman D.R., Using food as a reward: An examination of parental reward practices, Appetite (2017), doi: 10.1016/j.appet.2017.09.024. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

Using Food as a Reward: An Examination of Parental Reward Practices Lindsey Robertsab

Dara R. Musher-Eizenmanad

a

M AN U

SC

Bowling Green State University 822 East Merry Street Psychology Building Bowling Green OH 43403

RI PT

Jenna M. Marxac

Corresponding Author, [email protected]

c

[email protected]

d

[email protected]

AC C

EP

TE D

b

ACCEPTED MANUSCRIPT 1 1

Abstract Eating patterns and taste preferences are often established early in life. Many studies have

3

examined how parental feeding practices may affect children’s outcomes, including food intake

4

and preference. The current study focused on a common food parenting practice, using food as a

5

reward, and used Latent Profile Analysis (LPA) to examine whether mothers (n = 376) and

6

fathers (n = 117) of children ages 2.8 to 7.5 (M = 4.7; SD = 1.1) grouped into profiles (i.e.,

7

subgroups) based on how they use of food as a reward. The 4-class model was the best-fitting

8

LPA model, with resulting classes based on both the frequency and type of reward used. Classes

9

were: infrequent reward (33%), tangible reward (21%), food reward (27%), and frequent reward

M AN U

SC

RI PT

2

(19%). The current study also explored whether children’s eating styles (emotional overeating,

11

rood fussiness, food responsiveness, and satiety responsiveness) and parenting style

12

(Authoritative, Authoritarian, and Permissive) varied by reward profile. Analyses of Variance

13

(ANOVA) revealed that the four profiles differed significantly for all outcome variables except

14

satiety responsiveness. It appears that the use of tangible and food-based rewards have important

15

implications in food parenting. More research is needed to better understand how the different

16

rewarding practices affect additional child outcomes.

19 20 21 22 23 24

EP

18

AC C

17

TE D

10

Keywords: reward, eating, latent profile analysis, preschoolers

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

ACCEPTED MANUSCRIPT 2 Using Food as a Reward: An Examination of Parental Reward Practices

26

Parental feeding practices play a critical role in establishing young children’s food

27

preferences, eating patterns, and nutrition (e.g., Gerards & Kremers, 2015; Musher-Eizenman &

28

Holub, 2007). One such feeding practice—using rewards to encourage healthy eating—has

29

received some research attention, but results have been inconclusive on how using rewards

30

relates to children’s eating behaviors and health outcomes. The lack of clarity in the literature

31

may be due in part to researchers interchangeably studying parents’ use of foods as a reward for

32

eating (e.g., child gets a cookie if they eat their vegetables), parents’ use of non-foods as a

33

reward for eating (e.g., child gets extra play time if they eat their vegetables), and parents’ use of

34

foods as a reward for behavior (e.g., child gets candy if they receive good grades in school).

35

Each of these types of reward likely has a differential impact on children’s subsequent eating

36

behaviors; however, these distinctions have not been well-examined.

SC

M AN U

Food-based rewards for eating. The majority of the existing literature addressing how

TE D

37

RI PT

25

reward fits into food parenting has examined parents using preferred foods (treats) to reward

39

children for eating healthy foods. Studies have found that parents commonly (18-30% of parents)

40

encourage children to eat a non-preferred food by using a preferred food as a reward (e.g.,

41

allowing dessert if a child eats vegetables). Despite short-term effectiveness (i.e., the child eats

42

the non-preferred food , Orrell-Valente et al., 2007), longer-term effects of food-based rewards

43

for eating may be negative. For instance, some studies have suggested that introducing a reward

44

contingency might communicate to the child that the contingency food should be disliked

45

(Newman & Taylor, 1992). As Birch, Marlin, and Rotter noted (1984), even when using novel

46

foods that, by definition, are neither liked nor disliked, introducing a reward contingency where

47

the food is the instrumental component (i.e., not the reward or end goal) impacts children’s liking

AC C

EP

38

ACCEPTED MANUSCRIPT 3 of that food. For example, if a parent tells a child that they may eat dessert if they first eat their

49

brussels sprouts, the child then expects that brussels sprouts do not taste good, even without

50

tasting them. Similarly, other studies have found that when children are rewarded for eating

51

something that they already preferred, their preference for that food actually decreases (Birch,

52

Birch, Marlin, & Kramer, 1982). Overall, it appears that although reward contingencies increase

53

intake (e.g., of vegetables), they decrease preference (Birch et al., 1982; Cooke, Chambers,

54

Añez, Croker, et al., 2011), which may ultimately lead the child to eat less of the non-preferred

55

food in the longer term. Additionally, the child’s preference for the reward food (e.g., the

56

dessert) increases, adding yet another barrier to long-term health habits (Mikula, 1989).

M AN U

SC

RI PT

48

Non-food-based rewards for eating. In contrast, the literature suggests more

58

consistently positive outcomes when parents use non-food-based incentives (e.g., stickers or a

59

game) to encourage eating. Many studies have found that using non-food-based rewards

60

increases intake, with a few longitudinal studies suggesting that the increased intake is

61

maintained months later. For instance, Remington and colleagues (2012) found that children

62

rewarded with a sticker (i.e., a tangible, non-food incentive) for eating a neutral vegetable

63

demonstrated both increased intake and preference of the target vegetable at both post-test and at

64

3-month follow-up. Similarly, another randomized controlled trial found that tangible incentives

65

increased both liking and intake, with gains maintained at 3-month follow-up (Corsini, Slater,

66

Harrison, Cooke, & Cox, 2013). Horne and colleagues (2011) also found that rewarding children

67

with a tangible reward increased fruit and vegetable consumption. Further, these researchers

68

found that the increased intake generalized from the incentivized eating occasion to other meal

69

times where there were no rewards, and that the results were maintained at 6-month follow-up

70

(Horne et al., 2011). Finally, another longitudinal, experimental study found that effects on

AC C

EP

TE D

57

ACCEPTED MANUSCRIPT 4 consumption were larger in the tangible reward condition than in either the mere exposure or

72

control conditions; in this study, all three conditions increased liking of the food, but only

73

tangible reward affected actual intake (Cooke et al., 2011). In conclusion, although non-food

74

incentives for eating are far less studied than food-based incentives, recent research shows

75

stronger support for this practice and suggests that offering rewards such as stickers can

76

effectively increase children’s consumption of fruits and vegetables, without adversely affecting

77

children’s preference for the target foods.

SC

78

RI PT

71

Food-based rewards for behavior. Finally, use of food to reward behavior is pervasive, even though well-respected professional organizations, including the American Academy of

80

Pediatrics, American Academy of Family Physicians, and American Psychological Association

81

all advise against this practice as it may adversely affect physical health, learning, and behavior

82

(Fedewa & Davis, 2015). Although not focusing solely on reward practices, one study found that

83

maternal use of food as a reward for behavior was the only feeding practice significantly

84

associated with children’s eating in the absence of hunger (Remy, Issanchou, Chabanet, Boggio,

85

& Nicklaus, 2015). A separate study examining the reinforcing value of food found that children

86

with higher BMI and reward sensitivity responded to a task at a faster rate, and consumed more

87

calories in a snacking session when granted access to food (Rollins, Loken, Savage, & Birch,

88

2014a). These results are further supported by a study by Lu and colleagues (2015), who found

89

that children consumed more daily fat, carbohydrate, and total energy when parents used food to

90

reward behavior than did children whose parents did not use food as a reward. As in the earlier

91

study, these researchers found stronger effects for children with higher reward sensitivity.

92

Together, these findings suggest that use of food as a reward for behavior increases overall

93

caloric intake and may put children at risk for negative future outcomes.

AC C

EP

TE D

M AN U

79

ACCEPTED MANUSCRIPT 5 94

Parenting Style. A recent systematic review of the literature examined seven studies for connections between parenting style and child feeding behaviors (Collins, Duncanson, &

96

Burrows, 2014), and found weak to moderate associations. For example, regarding parental

97

monitoring of child food intake, the reviewers found that Authoritative and Permissive parenting

98

were both associated, but in opposite directions; Authoritarian parenting, however, was

99

associated with pressuring a child to eat and restricting food, suggesting that parenting style and

100

feeding are associated. A separate review found that children with Authoritative parents ate more

101

healthfully, were more active, and had lower BMI’s than other children (Sleddens, Gerards,

102

Thijs, Vries, & Kremers, 2011). Conversely, another study found that general parenting style did

103

not significantly affect children’s consumption (Vereecken, Legiest, De Bourdeaudhuij, & Maes,

104

2009). However, specific parenting behaviors (e.g., negotiation, pressure, giving-in, and

105

permissiveness) did have an impact on children’s eating (Vereecken et al., 2009). Although more

106

research in this area is needed, existing studies suggest that there may be a connection between

107

parenting style and feeding behaviors.

SC

M AN U

TE D

108

RI PT

95

Child Eating Behaviors. Research has also found modest links between parenting style and children’s eating behaviors. Generally, children’s eating behaviors are established at a young

110

age, are fairly stable over time, and are often linked to health outcomes (e.g., Ashcroft, Semmler,

111

Carnell, Jaarsveld, & Wardle, 2008). One systematic review found that there was substantial

112

evidence that parenting practices were related to children’s eating behaviors (Ventura & Birch,

113

2008), and concluded that there is likely a bidirectional relation. The bidirectional nature of the

114

relation has been further supported in recent studies (Harris, Fildes, Mallan, & Llewellyn, 2016;

115

Jansen et al., 2017), which have found that feeding practices often emerge in response to

116

individual child characteristics (such as pickiness and neophobia). In other words, children

AC C

EP

109

ACCEPTED MANUSCRIPT 6 respond to their parents’ feeding practices, and likewise, parents respond to their children’s

118

eating habits. Other research has found a relation between specific eating behaviors and general

119

parenting style. For example, Topham and colleagues (2011) found that children’s emotional

120

overeating was negatively predicted by authoritative parenting style, suggesting that overall

121

parenting style may be related to specific child outcomes. Similarly, another study found that

122

parental reward practices (specifically using food to reward behavior) were related to increased

123

food avoidance (i.e., fussiness) (Powell, Farrow, & Meyer, 2011). Other research has focused on

124

the relation between child outcomes and reward processes. For example, one study found a link

125

between overeating and reward sensitivity (Davis, Strachan, & Berkson, 2004), with a higher

126

sensitivity to reward positively correlated with emotional overeating. Collectively, this research

127

suggests that child eating outcomes are related to both parenting style and to reward practices.

SC

M AN U

128

RI PT

117

The Current Study. Although experimental trials have studied the effects of incentives on eating, few studies have addressed how parents actually use these practices in their homes.

130

The current study aims to address this gap by using a Latent Profile Analysis (LPA) to explore

131

whether parents of 3- to 7-year-old children fall into profiles (i.e., subgroups of parents) based on

132

their use of incentives, including items asking about food-based incentives for eating, food-based

133

incentives for behavior, and tangible incentives for eating. Additionally, this study explores

134

whether feeding profiles vary in relation to parental attempts to control their child’s weight.

135

Finally, this study explores how various reward strategies are related to both parenting style

136

(Authoritarian, Authoritative, and Permissive) and child eating outcomes (emotional overeating,

137

food fussiness, food responsiveness, and satiety responsiveness).

138

AC C

EP

TE D

129

Methods

ACCEPTED MANUSCRIPT 7 Participants. The sample included 495 parents of children between the ages of 2.8 and

140

7.5 years old (mean child age = 4.7 years, SD = 1.1). The sample comprised both mothers (n =

141

376) and fathers (n = 117), and two parents who did not specify gender; parents’ ages ranged

142

from 19 to 65 years old (M = 32.5 years; SD = 6.6). Seventy-nine percent of the sample

143

identified as Caucasian, 8% as African-American, 5% as Multiracial, 4% as Hispanic, and 2% as

144

Asian. Nearly half of respondents reported that they were employed full time (44%), with 14%

145

employed part-time, and 32% homemakers. The sample was largely middle class, with only

146

7.5% of the sample reporting a household income below $20,000 per year, 39% earning $20,000

147

to $50,000 per year, 34% earning between $50,000 and $80,000 per year, and 17% earning more

148

than $80,000 per year. Geographically, the sample comprised participants from 49 of the 50

149

United States, with no more than 8% from any one state (California).

SC

M AN U

150

RI PT

139

Procedure. As part of a larger study on parents’ feeding practices, parents were recruited using Amazon’s Mechanical Turk (MTurk), which has been shown to provide high quality,

152

generalizable data (Berinsky, Huber, & Lenz, 2012). Parents who were interested in participating

153

were provided with a link to an online survey hosted by Qualtrics. Data were screened based on

154

total completion time, patterned responses, and incorrect responses to multiple quality control

155

items that instructed parents to select a certain response. Fewer than 10% of those who

156

completed the survey were rejected. Participants who provided high quality data were

157

compensated $0.75 through MTurk.

EP

AC C

158

TE D

151

Several items, including all of the items used in the LPA, were written for a study that

159

aimed to create a measure of food parenting that corresponds to the content map outlined by

160

Vaughn et al (2016). To create the first pool of items, items from three validated scales of food

161

parenting were drawn from the Comprehensive Feeding Practices Questionnaire (CFPQ, 49

ACCEPTED MANUSCRIPT 8 items; Musher-Eizenman & Holub, 2007), the Meals in our Household structure subscale (MioH,

163

10 items; Anderson, Must, Curtin, & Bandini, 2012), and the Feeding Strategies Questionnaire

164

(FSQ, 13 items; Berlin, Davies, Silverman, & Rudolph, 2011). Items were excluded from the

165

MioH and the FSQ if they duplicated items already included in the CFPQ or were not relevant to

166

parent feeding practices. Additional items were drawn from a review of published scales

167

(O’Connor et al., 2016), and the authors generated an additional 42 items to reflect concepts

168

described by Vaughn et al. (2016). Following item generation, an exploratory factor analysis

169

using Maximum Likelihood extraction and Promax rotation was used to test factor structure and

170

to create scales. Factor loadings for the items included in this study ranged from .39 to .94. To

171

help assess for content validity, open-ended responses were also asked of parents who reported

172

they engaged in a particular feeding practice either “often” or “always;” for example, “Earlier

173

you said that you often or always [e.g., offer your child his/her favorite foods in exchange for

174

good behavior]. Please briefly tell us more about the last time that you did this.” Responses were

175

coded to analyze whether items assessed the intended content. Finally, a subset of parents

176

answered items a second time, two weeks after the initial data collection. Both Cronbach’s

177

alphas at time one (ranging from .67 to .91) and test-retest reliabilities (r ranging from .66 to .71)

178

all demonstrated moderate to strong reliability

180

SC

M AN U

TE D

EP

Measures

AC C

179

RI PT

162

Reward. All items measuring rewarding practices are novel items written for the purposes

181

of validating a new measure (Musher-Eizenman et al., Unpublished Manuscript). A principal

182

components analysis with Promax rotation was used to identify subscales, resulting in three

183

subscales: Non-Food Incentives to Eat, Food-Based Incentives to Eat, and Food-Based

184

Incentives to Behave. Non-Food Incentives to Eat included 3 items that assess the frequency

ACCEPTED MANUSCRIPT 9 with which parents use a non-food reward (e.g., a sticker or a game) as a reward for trying foods;

186

internal consistency was high (Cronbach’s alpha = .90). Food-Based Incentives to Eat included 4

187

items that assess the frequency with which parents use a child’s preferred foods as a reward for

188

eating a less-preferred food; internal consistency was good (Cronbach’s alpha = .86). Food-

189

Based Incentives to Behave contained 5 items assessing the frequency with which parents use a

190

child’s preferred foods as a reward for certain behaviors (e.g., getting good grades); internal

191

consistency was acceptable (Cronbach’s alpha = .77). Participants responded on a 5-point Likert

192

scale, with answers ranging from Never (1) to Always (5), with higher scores indicating higher

193

frequency. See Appendix A for all reward items.

M AN U

SC

RI PT

185

Child Eating Behavior Questionnaire. Items from the Child Eating Behavior

195

Questionnaire (Wardle, Guthrie, Sanderson, & Rapoport, 2001) measured parents’ reports of

196

children’s emotional overeating, food fussiness, food responsiveness, and satiety responsiveness.

197

The current study included 3 items from the Emotional Overeating subscale (Cronbach’s alpha =

198

.91) that assessed children’s tendency to eat in response to emotions (such as anxiety or

199

frustration). The Food Fussiness subscale (Cronbach’s alpha = .90) comprised 6 items that

200

measured both children’s avoidance of novel foods (neophobia) and general pickiness around

201

food. The Food Responsiveness subscale (Cronbach’s alpha = .83) comprised 5 items that

202

assessed children’s responsiveness to food cues and desire to eat. The current study included 4 of

203

these original items. However, in our sample, these 4 items demonstrated unacceptable reliability

204

(Cronbach’s alpha = .55). Therefore, two items were removed from the scale, resulting in two

205

remaining items that were used to assess children’s satiety (“My child gets full before their meal

206

is finished” and “My child gets full up easily”; Pearson’s r = .50). Participants responded on a 5-

207

point Likert scale ranging from Never (1) to Always (5), with higher scores indicating more

AC C

EP

TE D

194

ACCEPTED MANUSCRIPT 10 208

emotional overeating, more fussiness or refusal when trying new foods, and a greater tendency to

209

eat when not hungry.

210

Parenting Style. General parenting style was measured using the Parenting Practices Questionnaire (Robinson, Mandleco, Olsen, & Hart, 1995), which follows Baumrind’s (1971)

212

models of Authoritative, Authoritarian, and Permissive parenting. Briefly, Authoritative

213

parenting is characterized by warmth, responsiveness and appropriate structure, while

214

Authoritarian parenting emphasizes restriction, and Permissive parenting lacks appropriate

215

structure. The Authoritative subscale comprised 15 items (Cronbach’s alpha = .89), the

216

Authoritarian subscale included 11 items (Cronbach’s alpha = .87), and the Permissive subscale

217

included 5 items (Cronbach’s alpha = .74). Participants responded on a 5-point Likert scale

218

ranging from Never (1) to Always (5), and scores were a mean on each of the three subscales for

219

all participants (i.e., participants received a continuous score on each parenting style).

221 222

SC

M AN U

Weight control. One additional item, “are you trying to control your child’s weight (e.g.,

TE D

220

RI PT

211

with a diet) for any reason?” was included as a covariate, with responses of “yes” (1) or “no” (2). Latent Profile Analysis. To assess different patterns of parental reward, a Latent Profile Analysis (LPA) was conducted using item-level indicators from three reward subscales. LPA is a

224

subset of Latent Class Analysis (LCA) that incorporates continuous, as opposed to categorical,

225

indicators. Broadly, LPA uses a person-centered (rather than an item-centered) statistical

226

approach to identify subgroups within a sample by estimating the most likely latent class

227

grouping based on patterns of responding. LPA allows us to investigate patterns of rewarding

228

behavior based on the frequency with which parents report using different reward practices.

AC C

EP

223

229

Beginning with a 2-class model, classes were added sequentially and each model was

230

assessed for fit. Multiple indicators of model fit were used to assess overall fit, including the

ACCEPTED MANUSCRIPT 11 Bayesian Information Criteria (BIC), Akaike Information Criteria (AIC), and Entropy (Henson,

232

Reise, & Kim, 2007; Nylund, Asparouhov, & Muthén, 2007). Additionally, a bootstrap

233

likelihood ration test (BLRT) and a Lo, Mendell, Rubin Test (LMRT; Lo, Mendell, & Rubin,

234

2001) were used to compare models by testing whether a model with k classes offered a

235

significantly improved fit over a model with k -1 classes (e.g., a 3-class model would be

236

compared to a 2-class model). Models were estimated using a maximum likelihood estimation

237

with robust standard errors (MLR), with respondents classified based on the likelihood of each

238

case belonging to each class. All models were tested using MPlus version 7.1 software (Muthén

239

& Muthén, 2012).

SC

M AN U

240

RI PT

231

ANOVA. After grouping participants according to most-likely class, one-way betweensubjects Analyses of Variance (ANOVA) were used to explore whether the child outcome

242

variables (i.e., emotional overeating, food fussiness, food responsiveness, and satiety

243

responsiveness) and the parenting styles (i.e., Authoritative, Authoritarian, and Permissive)

244

varied by assigned profile. To reduce family-wise error, a Bonferroni adjustment was applied to

245

Fisher’s Least Significance Difference (LSD) test in order to examine post-hoc pairwise

246

comparisons.

EP

248

Results

Descriptive Statistics. Descriptive statistics (means and standard deviations) for Reward

AC C

247

TE D

241

249

subscales are presented in Table 1. Means ranged from 1.69 to 2.93, indicating that, in general,

250

parents report that they infrequently to sometimes engage in these rewarding practices.

251 252

Table 1. Means and standard deviations for items assessing each rewarding practice. Item Mean SD Food-Based Incentives to Eat 1 1.88 0.99 2 1.89 1.00

ACCEPTED MANUSCRIPT

SC

254

Correlations. Correlations between the Reward subscales and other key variables are

M AN U

253

3 2.17 1.08 4 1.69 0.89 Food-Based Incentives to Behave 5 2.33 1.06 6 2.60 1.29 7 2.31 1.07 8 2.93 1.01 9 2.18 1.09 Non-food based incentives to eat 10 2.20 1.23 11 2.28 1.23 12 2.02 1.19 Notes. All items range from 1 (Never) to 5 (Always).

RI PT

12

presented in Table 2. Significant correlations with child outcomes were positive and ranged from

256

.10 (Food-Based Incentives to Behave and Satiety Responsiveness) to .47 (Food-Based

257

Incentives to Eat and Emotional Overeating). Regarding parenting style, Authoritative parenting

258

was modestly, negatively correlated with both Food-Based Incentives to Eat and Food-Based

259

Incentives to Behave. Conversely, Authoritarian and Permissive parenting styles were both

260

positively related to Reward subscales, with correlations ranging from .139 (Authoritarian and

261

Non-Food Incentives to Eat) to .422 (Authoritarian with Food-Based Incentives to Behave).

262 263 264

Table 2. Correlations between rewarding practices and child eating behaviors and parenting styles.

AC C

EP

TE D

255

Scale

Non-Food Incentives to Eat

Food-Based Incentives to Eat

Food-Based Incentives to Behave

.23*

.47*

.35*

.31*

.18*

Child Outcomes

Emotional Overeating Food Fussiness

-.04

Food Responsiveness

.17*

.32*

.38*

Satiety Responsiveness

.08

.11*

.10*

.030

-.26*

-.13*

Parenting Style Authoritative

ACCEPTED MANUSCRIPT 13

266

.14*

.42*

.42*

Permissive Notes: * p <.05

.16*

.32*

.27*

Model Selection. The loglikelihood ratio (LLR), BIC, and AIC continued to improve

RI PT

265

Authoritarian

through the 7-class model. As class sizes decreases when adding more classes, models

268

containing more than 7 classes were not tested in order to prevent classes from becoming

269

sample-specific. Ultimately, the 4-class model provided the best fit, with an entropy of 88%, a

270

LLR of -7720.34, an AIC of 15572.67, and a BIC of 15850.18. Although the LMRT indicated

271

that the 4-class solution did not offer a significantly better fit than the 3-class solution (-2ln =

272

410.770, p = .08), the BLRT supported a 4-class solution (-2ln = 415.499, p = .000). Despite the

273

non-significant LMRT, when compared to a 3-class model, the 4-class model offered an

274

improved BIC and a significant BLRT, both of which have been shown to be consistent

275

indicators of class enumeration (Nylund et al., 2007). Therefore, a 4-class model was selected.

276

Posterior probabilities for the 4-class model were good, ranging from .91 to .94. The probability

277

of being in a different class than assigned ranged from 0 to .05.

M AN U

TE D

Description of Classes. The means for each item within each class for the 4-class model

EP

278

SC

267

were compared to describe and name the classes. To compare means across classes, item-level

280

means within each class were standardized using the item-specific means and standard deviations

281

for the entire sample. Profile 1 consisted of 33% of the sample (n = 163) and can be described as

282

infrequent reward; this subgroup consists of parents who do not frequently reward their children

283

with either tangible or food rewards. Profile 2 consisted of 21% of the sample (n = 102), and can

284

be described as tangible reward; this class consists of parents who frequently use tangible

285

rewards to encourage their children to eat, but do not often use food as a reward for behavior or

286

for eating. Profile 3 consisted of 27% of the sample (n = 135), and can be described as food

AC C

279

ACCEPTED MANUSCRIPT 14 reward; this class consisted of parents who do not use tangible rewards, but who use food to

288

reward for both behaviors and for eating. Profile 4 consisted of 19% of the sample (n = 95), and

289

can be described as frequent reward; parents who frequently use both tangible items and food to

290

reward children for eating and for behaviors. Figure 1 shows the centered means for each item.

291 292 293

Figure 1. Item means (centered) for each of the parenting profiles in the best fitting (4-class) model.

300

Notes. “Eat” = Food-Based Incentives to Eat; “Behave” = Food-Based Incentives to Behave; “Tangible” = Non-Food Incentives to Eat; Profiles: 1 = Infrequent Reward, 2 = Tangible Reward, 3 = Food Reward, 4 = Frequent Reward.

AC C

294 295 296 297 298 299

EP

TE D

M AN U

SC

RI PT

287

Weight Control. To examine whether class membership varies when parents are trying

301

to control their child’s weight, a dichotomous indicator was included as a covariate. Analyses

302

revealed no significant differences between parents in each of the four profiles.

ACCEPTED MANUSCRIPT 15 303

Reward Practices. The ANOVAs confirmed that the four profiles differed significantly on using food-incentives for eating, food-incentives for behavior, and non-food incentives for

305

eating. See Table 3 for overall means and profile means, Table 4 for ANOVA results, and Tables

306

5 for post-hoc comparisons.

RI PT

304

Post-hoc comparison confirmed that parents in the infrequent reward profile had

308

significantly lower scores on all three reward practices than did parents in the other three

309

profiles. Additionally, parents in frequent reward profile had the highest mean scores for all

310

scales with one exception: they did not significantly differ from parents in the tangible reward

311

profile on the use of non-food based incentives to eat. Parents in the tangible reward profile had

312

significantly higher scores on the use of non-food incentives to eat than did parents in the

313

infrequent reward and food reward profiles. Lastly, parents in the food reward profile had higher

314

scores on both food-based incentives scales than did parents in the tangible or infrequent

315

profiles.

M AN U

TE D

Table 3. Mean scores for rewarding practices, child eating behaviors, and parenting style for each reward profile. Scale 1 2 3 4 Overall Reward Practices Food-Based Incentives to Eat 1.22 1.58 2.14 3.09 1.90 Food-Based Incentives to Behave 1.82 2.26 2.78 3.36 2.47 Non-Food Incentives to Eat 1.19 3.31 1.68 3.30 2.16 Child Eating Behaviors Emotional Overeating 1.33 1.53 1.71 2.30 1.66 Food Fussiness 2.60 2.41 3.02 3.04 2.76 Food Responsiveness 2.14 2.22 2.48 2.94 2.40 Satiety Responsiveness 3.00 3.06 3.03 3.18 3.06 Parenting Styles Authoritative 4.10 4.23 3.83 3.76 3.99 Authoritarian 1.65 1.63 1.90 2.35 1.85 Permissive 1.96 1.97 2.20 2.60 2.15

AC C

EP

316 317 318

SC

307

319

ACCEPTED MANUSCRIPT 16

325

Notes: Profiles: 1 = Infrequent reward, 2 = Tangible Reward, 3 = Food Reward, 4 = Frequent Reward. Table 4. Results of ANOVAs assessing differences between reward profiles on rewarding practices, child eating behaviors and parenting styles.

326

p

.665 .506 .727

316.81* 163.14* 424.85*

.000 .000 .000

.188 .081 .121 .007

36.98* 14.09* 21.95* 1.13

.000 .000 .000 .335

RI PT

F (3,478)

.089 .187 .101

15.54* 36.68* 17.96*

.000 .000 .000

TE D

Table 5. Post hoc comparisons (Fischer’s Least Square Difference with Bonferroni adjustment) assessing differences between reward profiles on rewarding practices. Food-Based Incentives Food-Based Incentives for Non-Food Incentives for for Eating Behavior Eating Profile - Profile Mean Diff p Mean Diff p Mean Diff p 1 2 -.357* .000 -.427* .000 -2.115* .000 3 -.910* .000 -.953* .000 -.493* .000 4 -1.847* .000 -1.517* .000 -2.105* .000 2 3 -.553* .000 -.526* .000 1.623* .000 4 -1.490* .000 -1.090* .000 .011 1.000 3 4 -.937* .000 -.564* .000 -1.612* .000 Note: * p <.05; Bonferroni corrections can result in p-values of greater than 1; in these cases, pvalues are reported as 1. Profiles: 1 = Infrequent Reward, 2 = Tangible Reward, 3 = Food Reward, 4 = Frequent Reward.

329 330 331 332 333 334

AC C

EP

327 328

R2

SC

Outcome Reward Practices Food-Based Incentives to Eat Food-Based Incentives to Behave Non-Food Incentives to Eat Child Outcomes Emotional Overeating Food Fussiness Food Responsiveness Satiety Responsiveness Parenting Style Authoritative Authoritarian Permissive Notes: * p <.05

M AN U

320 321 322 323 324

Child Outcomes. The ANOVAs revealed that the four profiles differed significantly on emotional overeating, food fussiness, and food responsiveness, but not on satiety responsiveness.

ACCEPTED MANUSCRIPT 17 335

See Table 3 for overall means and profile means, Table 4 for ANOVA results, and Table 6 for

336

post-hoc comparison on significant outcomes. Post-hoc comparisons for emotional overeating (i.e., children’s tendency to eat in

338

response to negative emotions) revealed that parents in the infrequent reward profile reported the

339

lowest scores. Conversely, parents in the frequent reward profile reported the highest scores.

340

There were no significant differences between parents in the tangible reward profile and those in

341

the food reward profile.

SC

RI PT

337

Regarding food fussiness (i.e., children’s refusal to try novel foods or general food

343

pickiness), post hoc comparisons revealed that parents who used infrequent reward and tangible

344

reward had lower scores than did parents who used food rewards and frequent reward.

345

M AN U

342

Finally, food responsiveness (i.e., children’s responsiveness to food cues and desire to eat) revealed similar patterns. Parents who used frequent reward reported the highest scores for

347

food responsiveness, with significantly higher means than parents in all other profiles. Parents

348

who use infrequent reward practices reported lower mean scores for food responsiveness than

349

did parents who used food rewards or frequent reward.

350 351 352

Table 6. Post hoc comparisons (Fischer’s Least Square Difference with Bonferroni adjustment) assessing significant differences between reward profiles on child eating behaviors.

EP

TE D

346

AC C

Emotional Overeating Food Fussiness Food Responsiveness Profile - Profile Mean Diff p Mean Diff p Mean Diff p 1 2 -.193 .169 .198 .435 -.085 1.000 3 -.360* .000 -.411* .000 -.334* .002 4 -.939* .000 -.425* .001 -.792* .000 2 3 -.167 .405 -.609* .000 -.249 .099 4 -.746* .000 -.623* .000 -.707* .000 3 4 -.579* .000 .118 1.000 -.458* .000 Note: * p <.05; Bonferroni corrections can result in p-values of greater than 1; in these cases, pvalues are reported as 1. Profiles: 1 = Infrequent Reward, 2 = Tangible Reward, 3 = Food Reward, 4 = Frequent Reward.

353 354 355 356

ACCEPTED MANUSCRIPT 18 357

Parenting Style. The ANOVA revealed that the four profiles differed significantly on all

358

three parenting styles; see Table 3 for group means, Table 4 for initial ANOVA results, and

359

Table 7 for post-hoc comparisons. Parents who used frequent reward had the lowest scores for Authoritative parenting,

RI PT

360

while parents who used tangible reward had the highest mean scores. Regarding Authoritarian

362

parenting practices, parents who used frequent reward had the highest scores for Authoritarian

363

parenting, while parents who used tangible reward had the lowest scores. Similar patterns were

364

seen for Permissive parenting practices: parents who used frequent reward had the highest mean

365

scores for Permissive parenting.

366 367 368 369

Table 7. Post hoc comparisons (Fischer’s Least Square Difference with Bonferroni adjustment) assessing differences between reward profiles on parenting styles.

M AN U

SC

361

Authoritative

3 4

Mean Diff .019

p 1.000

Mean Diff -.0116

p 1.000

.264*

.001

-.257*

.001

-.243*

.029

.324*

.000

-.700*

.000

-.648*

.000

.397*

.000

-.276*

.001

-.231

.097

4 .457* .000 -.719* .000 -.636* 3 4 .060 1.000 -.442* .000 -.405* Note: * p <.05; Bonferroni corrections can result in p-values of greater than 1; in these cases, pvalues are reported as 1. Profiles: 1 = Infrequent Reward, 2 = Tangible Reward, 3 = Food Reward, 4 = Frequent Reward.

AC C

370 371 372 373 374 375

3

p .409

Permissive

EP

2

Mean Diff -0.133

TE D

Profile - Profile 1 2

Authoritarian

Discussion Using a person-centered approach to exploring parental feeding practices revealed four

376

common patterns of incentive-based food parenting. These profiles revealed that two groups of

377

parents fall on opposite ends of a continuum, with some parents (frequent reward, profile 4)

378

offering both food and tangible rewards to their children for both eating and behavior, while

.000 .000

ACCEPTED MANUSCRIPT 19 other parents (infrequent reward, profile 1) offer few incentives of any type to encourage their

380

children to eat or behave in a certain way. Perhaps most interesting are the two profiles that fell

381

in the middle of the continuum—the group of parents who frequently use tangible rewards but

382

not food-based rewards (tangible reward, profile 2) and the group of parents who frequently use

383

food-based rewards but not tangible rewards (food rewards, profile 3). Both of these profiles are

384

underrepresented in the existing literature. However, as each of these profiles represent over 20%

385

of the sample, and together represent close to half of the sample (48%), these findings suggest

386

that an artificial dichotomy that simplifies whether or not parents use rewards in food parenting

387

does not accurately capture the feeding practices of a substantial portion of parents. By using an

388

LPA to explore feeding practices, we were able to take a more nuanced approach that accounts

389

for both differences in frequency and types of incentives. Analyses of Variance (ANOVA) and

390

post-hoc comparisons confirmed that these profiles differed significantly across reward practices

391

in the expected directions.

TE D

M AN U

SC

RI PT

379

Understanding how parents reward children for eating is an important first step in

393

understanding the effects of reward on outcome variables of interest. Most of the existing

394

literature focuses on either taste preference or intake (Cooke, Chambers, Añez, & Wardle, 2011),

395

with mixed findings. However, the recent literature that explores the use of tangible rewards has

396

shown positive results regarding fruit and vegetable consumption (Corsini et al., 2013; Horne et

397

al., 2011; Remington et al., 2012) and has generally not shown the negative results on taste

398

preference that accompany food-based rewards. Therefore, the two groups of parents who use

399

tangible rewards differently from food-based rewards (tangible reward, profile 2 and food

400

reward, profile 3) could be an important population to study, but have remained largely

401

unrepresented given the ways that the existing literature operationalizes rewards. Further, given

AC C

EP

392

ACCEPTED MANUSCRIPT 20 clear evidence linking the use of food to reward behavior with increased caloric intake (Fedewa

403

& Davis, 2015; Lu et al., 2015; Remy et al., 2015; Rollins, Loken, Savage, & Birch, 2014b), it

404

may be especially instructive to understand the portion of parents that do not report engaging in

405

this practice. Perhaps further research identifying individual difference variables could help to

406

facilitate more targeted health promotion programs for parents depending on their reward styles.

RI PT

402

The results of the current study found that across all outcomes measured, parents in the

408

tangible reward profile and parents in the infrequent reward profile did not significantly differ,

409

and these two groups had the highest mean scores on the scale measuring positive parenting

410

styles (i.e., Authoritative parenting practices) and the lowest scores on scales measuring negative

411

outcomes (i.e., emotional overeating, food fussiness, food responsiveness, Authoritarian

412

parenting practices, and Permissive parenting practices). Given that the primary similarity

413

between these two groups is their lack of food rewards, these results suggest that some of the

414

negative associations between reward practices and child eating outcomes (Birch et al., 1982,

415

1984; Newman & Taylor, 1992) may be due to the use of food as a reward, rather than to the use

416

of any reward. Parents in the infrequent reward and tangible reward profiles reported

417

significantly lower scores for their children on food fussiness than did parents in the food and

418

frequent reward profiles (where both groups use food as a reward); parents who reward with

419

food reported that their children tend to reject new foods. It should be noted that there were not

420

significant differences between parents using tangible rewards and those using infrequent

421

rewards, suggesting that using tangible rewards to encourage a child to try new foods may be a

422

suitable strategy for parents who struggle with neophobic or picky-eating children. It should also

423

be noted that overall, most parents report engaging in these practices somewhat infrequently;

AC C

EP

TE D

M AN U

SC

407

ACCEPTED MANUSCRIPT 21 424

even parents in the frequent reward profile reported mean scores that reflect that they

425

“sometimes” endorse these practices.

426

Similar patterns were seen in the parenting style outcomes. Parents in the infrequent reward and tangible reward profiles did not significantly differ on any of the three parenting

428

styles. However, these two groups did have significantly higher scores on Authoritative practices

429

than did parents in the food and frequent reward profiles. Conversely, parents in the food and

430

frequent reward profiles had significantly higher scores for Authoritarian practices. Although

431

Authoritative, Authoritarian, and Permissive styles of parenting are well established in the

432

literature (Baumrind, 1971), research in parent feeding has found mixed support for connections

433

between parenting style and child outcomes. Although at least some research suggests that

434

overall parenting style is related to outcomes (Collins et al., 2014; Sleddens et al., 2011), other

435

research supported the connection between specific parenting behaviors and outcomes

436

(Vereecken et al., 2009). The results of the current study indicate that while more research is

437

needed to better understand the nuanced associations between parenting style and eating

438

outcomes, overall parenting style is related to parental rewarding practices, which are tied to

439

child eating outcomes. Therefore, the use (or lack) of food as a reward might offer one link

440

between overall parenting style and child eating outcomes.

SC

M AN U

TE D

EP

As always, limitations of the research should be acknowledged. Primarily, the data are

AC C

441

RI PT

427

442

cross-sectional; we are only able to examine trends and are unable to draw causal conclusions.

443

Substantial research has demonstrated that feeding practices are bi-directional in nature—parents

444

adapt their feeding to suit their children’s needs, and children’s eating changes with their

445

parent’s feeding practices (Harris et al., 2016; Jansen et al., 2017). Individual characteristics (of

446

both the parent and the child) likely affect this relationship as well. Using cross-sectional data,

ACCEPTED MANUSCRIPT 22 we are unable to conclusively determine whether the use of food as a reward leads to undesirable

448

outcomes, such as food fussiness, or whether parents are offering their children food as a reward

449

because their children are already hesitant to try new foods. Other individual differences (of both

450

the parent and the child) may also contribute to the bi-directional nature of feeding practices.

451

For instance, a child who is not particularly responsive to food would likely not respond strongly

452

to the use of food as a reward, and the parent may then choose to reward with stickers instead.

453

The current study does not allow us to determine the directions of the findings. Additionally,

454

self-report data is inherently limited. However, by using self-reported cross-sectional data, we

455

were able to collect a much larger sample than we would have otherwise, which also allowed us

456

to include many more fathers, a historically underrepresented group in the parent feeding

457

literature. While the existing experimental literature offers many important contributions, self-

458

report data also allowed us to explore which practices parents engage in at home.

M AN U

SC

RI PT

447

The findings of the current study underscore the need to further explore the use of

460

rewards in food parenting and how it may be related to outcomes of interest, which might expand

461

beyond emotional overeating, food fussiness, food responsiveness, and satiety responsiveness.

462

Similarly, more research is needed to better understand which foods and/or behaviors parents are

463

targeting with their use of rewards. Further, future research could explore whether using food to

464

reward eating differs from using food to reward behaviors, and ways in which interventions

465

might target incentive practices. For instance, more research is needed to determine how to better

466

promote fruit and vegetable intake without the use of sugary food as a reward. Similarly, helping

467

parents understand how various profiles of rewarding practices are linked to children’s eating

468

behaviors could be an effective way to convey this information clearly and to encourage parents

469

to adopt the healthiest practices for their child.

470

AC C

EP

TE D

459

ACCEPTED MANUSCRIPT 23 471

References Anderson, S., Must, A., Curtin, C., & Bandini, L. (2012). Meals in our Household: Reliability

473

and initial validation of a questionnaire to assess child mealtime behaviors and family

474

mealtime environments. Journal of the Academy of Nutrition and Dietetics, 112, 276–

475

284. https://doi.org/10.1016/j.jada.2011.08.035

RI PT

472

Ashcroft, J., Semmler, C., Carnell, S., Jaarsveld, C. H. M. van, & Wardle, J. (2008). Continuity

477

and stability of eating behaviour traits in children. European Journal of Clinical

478

Nutrition, 62, 985–990. https://doi.org/10.1038/sj.ejcn.1602855

480 481

Baumrind, D. (1971). Current patterns of parental authority. Developmental Psychology, 4, 1–

M AN U

479

SC

476

103. https://doi.org/10.1037/h0030372

Berinsky, A., Huber, G., & Lenz, G. (2012). Evaluating online labor markets for experimental research: Amazon.com’s mechanical turk. Political Analysis, 20, 351–368.

483

https://doi.org/10.1093/pan/mpr057

484

TE D

482

Berlin, K., Davies, W., Silverman, A., & Rudolph, C. D. (2011). Assessing family-based feeding strategies, strengths, and mealtime structure with the Feeding Strategies Questionnaire.

486

Journal of Pediatric Psychology, 36, 586–595. https://doi.org/10.1093/jpepsy/jsp107

488

Birch, L., Birch, D., Marlin, D., & Kramer, L. (1982). Effects of instrumental consumption on children’s food preference. Appetite, 3, 125–134.

AC C

487

EP

485

489

Birch, L., Marlin, D., & Rotter, J. (1984). Eating as the “means” activity in a contingency:

490

Effects on young children’s food preference. Child Development, 55, 431.

491

https://doi.org/10.2307/1129954

ACCEPTED MANUSCRIPT 24 492

Collins, C., Duncanson, K., & Burrows, T. (2014). A systematic review investigating

493

associations between parenting style and child feeding behaviours. Journal of Human

494

Nutrition & Dietetics, 27, 557–568. Cooke, L., Chambers, L., Añez, E., Croker, H., Boniface, D., Yeomans, M., & Wardle, J. (2011).

RI PT

495

Eating for pleasure or profit: The effect of incentives on children’s enjoyment of

497

vegetables. Psychological Science, 22, 190–196.

498

https://doi.org/10.1177/0956797610394662

499

SC

496

Cooke, L., Chambers, L., Añez, E., & Wardle, J. (2011). Facilitating or undermining? The effect of reward on food acceptance. A narrative review. Appetite, 57, 493–497.

501

https://doi.org/10.1016/j.appet.2011.06.016

502

M AN U

500

Corsini, N., Slater, A., Harrison, A., Cooke, L., & Cox, D. (2013). Rewards can be used effectively with repeated exposure to increase liking of vegetables in 4–6-year-old

504

children. Public Health Nutrition, 16, 942–951.

505

https://doi.org/10.1017/S1368980011002035

506

TE D

503

Davis, C., Strachan, S., & Berkson, M. (2004). Sensitivity to reward: Implications for overeating and overweight. Appetite, 42, 131–138. https://doi.org/10.1016/j.appet.2003.07.004

508

Fedewa, A., & Davis, M. (2015). How food as a reward is detrimental to children’s health,

510 511 512

learning, and behavior. Journal of School Health, 85, 648–658.

AC C

509

EP

507

Gerards, S. M. P. L., & Kremers, S. P. J. (2015). The role of food parenting skills and the home food environment in children’s weight gain and obesity. Current Obesity Reports, 4, 30– 36. https://doi.org/10.1007/s13679-015-0139-x

ACCEPTED MANUSCRIPT 25 513

Harris, H., Fildes, A., Mallan, K., & Llewellyn, C. (2016). Maternal feeding practices and fussy

514

eating in toddlerhood: A discordant twin analysis. International Journal of Behavioral

515

Nutrition and Physical Activity, 13, 81. https://doi.org/10.1186/s12966-016-0408-4 Henson, J., Reise, S., & Kim, K. (2007). Detecting mixtures from structural model differences

RI PT

516

using latent variable mixture modeling: A comparison of relative model fit statistics.

518

Structural Equation Modeling: A Multidisciplinary Journal, 14, 202–226.

519

https://doi.org/10.1080/10705510709336744

520

SC

517

Horne, P., Greenhalgh, J., Erjavec, M., Lowe, C., Viktor, S., & Whitaker, C. (2011). Increasing pre-school children’s consumption of fruit and vegetables. A modelling and rewards

522

intervention. Appetite, 56, 375–385. https://doi.org/10.1016/j.appet.2010.11.146

523

M AN U

521

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

525

influences whom? Physiology & Behavior, 176, 101–106.

526

https://doi.org/10.1016/j.physbeh.2017.02.015

528

Lo, Y., Mendell, N., & Rubin, D. (2001). Testing the number of components in a normal mixture. Biometrika, 88, 767–778.

EP

527

TE D

524

Lu, J., Xiong, S., Arora, N., & Dubé, L. (2015). Using food as reinforcer to shape children’s non-

530

food behavior: The adverse nutritional effect doubly moderated by reward sensitivity and

531 532 533 534

AC C

529

gender. Eating Behaviors, 19, 94–97. https://doi.org/10.1016/j.eatbeh.2015.07.003

Mikula, G. (1989). Influencing food preferences of children by “if-then” type instructions. European Journal of Social Psychology, 19, 225–241.

https://doi.org/10.1002/ejsp.2420190304

ACCEPTED MANUSCRIPT 26 535

Musher-Eizenman, D., & Holub, S. (2007). Comprehensive Feeding Practices Questionnaire:

536

Validation of a new measure of parental feeding practices. Journal of Pediatric

537

Psychology, 32, 960–972. https://doi.org/10.1093/jpepsy/jsm037 Musher-Eizenman, D., Marx, J., Goodman, L., Roberts, L., Taylor, M., & Hoffman, D.

539

(Unpublished Manuscript). F-SCAPE: Feeding – Structure, Control, Autonomy

540

Promotion Evaluation. A theory-driven measure of food parenting practices. Bowling

541

Green State University.

544 545 546

SC

543

Muthén, L., & Muthén, B. O. (2012). MPlus user’s guide: Statistical analysis with latent variables. Muthén & Muthén.

M AN U

542

RI PT

538

Newman, J., & Taylor, A. (1992). Effect of a means-end contingency of young children’s food preferences. Journal of Experimental Child Psychology, 64, 200–216. Nylund, K. L., Asparouhov, T., & Muthén, B. O. (2007). Deciding on the number of classes in latent class analysis and growth mixture modeling: A Monte Carlo simulation study.

548

Structural Equation Modeling, 14, 535–569.

TE D

547

O’Connor, T., Pham, T., Watts, A., Tu, A., Hughes, S., Beauchamp, M., … Mâsse, L. (2016).

550

Development of an item bank for food parenting practices based on published

551

instruments and reports from Canadian and US parents. Appetite, 103, 386–395.

552

https://doi.org/10.1016/j.appet.2016.04.033

554 555

AC C

553

EP

549

Orrell-Valente, J. K., Hill, L., Brechwald, W., Dodge, K., Pettit, G., & Bates, J. (2007). “Just three more bites”: An observational analysis of parents’ socialization of children’s eating

at mealtime. Appetite, 48, 37–45. https://doi.org/10.1016/j.appet.2006.06.006

ACCEPTED MANUSCRIPT 27 556

Powell, F., Farrow, C., & Meyer, C. (2011). Food avoidance in children. The influence of

557

maternal feeding practices and behaviours. Appetite, 57, 683–692.

558

https://doi.org/10.1016/j.appet.2011.08.011 Remington, A., Añez, E., Croker, H., Wardle, J., & Cooke, L. (2012). Increasing food

RI PT

559

acceptance in the home setting: a randomized controlled trial of parent-administered taste

561

exposure with incentives. American Journal of Clinical Nutrition, 95, 72–77.

562

https://doi.org/10.3945/ajcn.111.024596

563

SC

560

Remy, E., Issanchou, S., Chabanet, C., Boggio, V., & Nicklaus, S. (2015). Impact of adiposity, age, sex and maternal feeding practices on eating in the absence of hunger and caloric

565

compensation in preschool children. International Journal of Obesity, 39, 925–930.

566

https://doi.org/10.1038/ijo.2015.30

567

M AN U

564

Robinson, C. C., Mandleco, B., Olsen, S. F., & Hart, C. H. (1995). Authoritative, authoritarian, and permissive parenting practices: Development of a new measure. Psychological

569

Reports, 77, 819–830.

570

TE D

568

Rollins, B., Loken, E., Savage, J., & Birch, L. (2014a). Effects of restriction on children’s intake differ by child temperament, food reinforcement, and parent’s chronic use of restriction.

572

Appetite, 73, 31–39. https://doi.org/10.1016/j.appet.2013.10.005

574 575 576

Rollins, B., Loken, E., Savage, J., & Birch, L. (2014b). Measurement of food reinforcement in

AC C

573

EP

571

preschool children. Associations with food intake, BMI, and reward sensitivity. Appetite,

72, 21–27. https://doi.org/10.1016/j.appet.2013.09.018

Sleddens, E., Gerards, S. M. P. L., Thijs, C., Vries, N., & Kremers, S. P. J. (2011). General

577

parenting, childhood overweight and obesity-inducing behaviors: A review. International

578

Journal of Pediatric Obesity, 6, e12–e27. https://doi.org/10.3109/17477166.2011.566339

ACCEPTED MANUSCRIPT 28 579

Topham, G., Hubbs-Tait, L., Rutledge, J., Page, M., Kennedy, T., Shriver, L., & Harrist, A. (2011). Parenting styles, parental response to child emotion, and family emotional

581

responsiveness are related to child emotional eating. Appetite, 56, 261–264.

582

https://doi.org/10.1016/j.appet.2011.01.007

583

RI PT

580

Vaughn, A., Ward, D., Fisher, J., Faith, M., Hughes, S., Kremers, S., … Power, P. (2016). Fundamental constructs in food parenting practices: a content map to guide future

585

research., Fundamental constructs in food parenting practices: a content map to guide

586

future research. Nutrition Reviews, Nutrition Reviews, 74, 74(2, 2), 98, 98–117.

587

https://doi.org/10.1093/nutrit/nuv061, 10.1093/nutrit/nuv061

M AN U

588

SC

584

Ventura, A., & Birch, L. (2008). Does parenting affect children’s eating and weight status?

589

International Journal of Behavioral Nutrition and Physical Activity, 5, 15.

590

https://doi.org/10.1186/1479-5868-5-15

Vereecken, C., Legiest, E., De Bourdeaudhuij, I., & Maes, L. (2009). Associations between

TE D

591 592

general parenting styles and specific food-related parenting practices and children’s food

593

consumption. American Journal of Health Promotion, 23, 233–240. Wardle, J., Guthrie, C. A., Sanderson, S., & Rapoport, L. (2001). Development of the Children’s

EP

594

Eating Behaviour Questionnaire. Journal of Child Psychology & Psychiatry & Allied

596

Disciplines, 42, 963.

597 598 599 600

AC C

595

ACCEPTED MANUSCRIPT 29 601

RI PT

Variable names correspond to items listed in Figure 1. All items except weight control were answered on the following 5-point Likert scale. 1. Never 2. Rarely 3. Sometimes 4. Often 5. Always

M AN U

SC

Food-Based Incentives to Eat. 1. I offer my child sweets (candy, ice cream, cake, pastries) as a reward for eating fruits or vegetables. 2. I offer my child sweets (candy, ice cream, cake, pastries) as a reward for trying new foods. 3. I promise to give my child a food they like in exchange for eating a disliked food (e.g., if they eats soup for lunch, they will get pizza for dinner). 4. I give my child an unhealthy food or drink if they promise to eat a healthy food later.

TE D

Food-Based Incentives to Behave. 5. I offer sweets (candy, ice cream, cake, pastries) to my child as a reward for good behavior. 6. I withhold sweets/dessert from my child in response to bad behavior. 7. I offer my child his/her favorite foods in exchange for good behavior. 8. I take my child out for ice cream or give them other food treats to celebrate (e.g., after winning a sports game, after completing a big school project). 9. I give my child food treats (e.g., M & Ms) to encourage them to complete tasks (e.g., practicing an instrument, doing chores).

EP

Non-Food Incentives to Eat. 10. I offer my child a sticker, activity, or other non-food reward for eating fruits or vegetables. 11. I offer my child a sticker, activity, or other non-food reward for trying new foods. 12. I offer my child a sticker, activity, or other non-food reward in exchange for eating their whole meal.

AC C

602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637

Appendix A

Weight control. 1. Are you trying to control your child’s weight (e.g., with a diet) for any reason? 1. Yes 2. No