Shopping for food with children: A strategy for directing their choices toward novel foods containing vegetables

Accepted Manuscript Shopping for food with children: A strategy for directing their choices toward novel foods containing vegetables Xavier Allirot, Edurne Maiz, Elena Urdaneta PII:

S0195-6663(17)30782-1

DOI:

10.1016/j.appet.2017.09.008

Reference:

APPET 3610

To appear in:

Appetite

Received Date: 29 May 2017 Revised Date:

24 July 2017

Accepted Date: 9 September 2017

Please cite this article as: Allirot X., Maiz E. & Urdaneta E., Shopping for food with children: A strategy for directing their choices toward novel foods containing vegetables, Appetite (2017), doi: 10.1016/ j.appet.2017.09.008. 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.

1

ACCEPTED MANUSCRIPT 1

TITLE: Shopping for food with children: a strategy for directing their choices toward novel

2

foods containing vegetables.

3

5 6

AUTHORS: Xavier Allirota,*, Edurne Maiza, Elena Urdanetaa

7 a

Basque Culinary Center, Paseo Juan Avelino Barriola, 101 – 20009 San Sebastián, Spain.

9 CORRESPONDING AUTHOR: Xavier Allirot PhD

M AN U

10

SC

8

RI PT

4

Email: [email protected]

12

Phone: +34 943 574 533

AC C

EP

TE D

11

1

ACCEPTED MANUSCRIPT ABSTRACT

2

Involving children in the different steps of meal preparation has been suggested as a strategy

3

for enhancing dietary habits in childhood. It has previously been shown that involving children

4

in cooking can increase their willingness to taste novel foods and direct their food choices

5

towards foods containing vegetables. The objective of the present study was to assess the effect

6

of involving children in food purchasing on food choices, intake, liking and appetite. A

7

between-subject experiment was conducted with 86 children (from 8 to 10 years old). Forty-

8

three children (PURCHASE group) participated in a workshop dedicated to purchasing the

9

necessary ingredients online for the preparation of three unfamiliar foods containing

10

vegetables: apple and beetroot juice, zucchini tortilla sandwich and spinach cookies. Forty-

11

three children (CONTROL group) participated instead in a creativity workshop. Afterwards, all

12

the children were invited to choose, for an afternoon snack, between three familiar vs.

13

unfamiliar foods: orange vs. apple and beetroot juice, potatoes vs. zucchini tortilla sandwich

14

and chocolate vs. spinach cookie. The mean number of unfamiliar foods chosen per child was

15

higher in the PURCHASE (0.70 ± 0.14) vs. CONTROL (0.19 ± 0.07) group (P = 0.003). The

16

liking for 1 of the 3 unfamiliar foods was higher in the PURCHASE group (P < 0.05). We did

17

not find any difference between the two groups in food intake estimation and in the levels of

18

subjective appetite. This study demonstrates that involving children in purchasing food can

19

help in directing their food choices towards unfamiliar foods containing vegetables. It

20

highlights the importance of involving children in the different steps of meal preparation for

21

decreasing food neophobia.

AC C

EP

TE D

M AN U

SC

RI PT

1

22 23

KEYWORDS

24

Food purchasing

25

Food neophobia

2

ACCEPTED MANUSCRIPT Children

27

Food choices

28

Vegetables

29

Willingness to taste

AC C

EP

TE D

M AN U

SC

RI PT

26

3

ACCEPTED MANUSCRIPT INTRODUCTION

31

Great attention is currently being paid to children's eating behaviour and how to change it to be

32

more healthy (DeCosta, Moller, Frost, & Olsen, 2017). In particular, increasing vegetable

33

intakes in children is a key objective for both parents and health professionals (Jaime & Lock,

34

2009; Russell, Worsley, & Campbell, 2015). As low intakes of fruit and vegetables in

35

childhood track into adolescence and adulthood, many campaigns are trying to increase the

36

intake of these healthy foods in children (Bucher, Siegrist, & van der Horst, 2014). The

37

reluctance to eat novel foods, called food neophobia, is associated with a lower vegetable

38

liking and intake and with an overall less healthy diet (Cooke, Wardle, & Gibson, 2003;

39

Falciglia, Couch, Gribble, Pabst, & Frank, 2000; Russell & Worsley, 2013). As a result, trying

40

to decrease food neophobia is a frequently used strategy for enhancing diet quality in children.

41

Involving children in the different steps of meal preparation has been suggested as a strategy

42

for enhancing dietary habits in childhood (Nelson, Corbin, & Nickols-Richardson, 2013) and

43

particularly for reducing food neophobia (Allirot, da Quinta, Chokupermal, & Urdaneta, 2016).

44

Involvement in meal preparation at home has been shown to be associated with overall better

45

diet quality (Chu, Storey, & Veugelers, 2014), more eating enjoyment and lower levels of

46

picky eating (van der Horst, 2012). Involving children in cooking can increase their willingness

47

to taste novel foods and direct their food choices towards foods containing vegetables (Allirot

48

et al., 2016). The importance of children’s involvement is also recognized by parents (Casey &

49

Rozin, 1989) and by food experts (Fordyce-Voorham, 2011).

50

Studies assessing the links between child involvement in meal preparation and diet quality has

51

mostly focused on one step of meal preparation: cooking (Chu et al., 2014; Leech et al., 2014;

52

van der Horst, Ferrage, & Rytz, 2014). Little information is available on involvement in other

53

steps. In particular, little is known about the links between children’s involvement in food

54

shopping and diet quality. One study (Nozue et al., 2016) showed that children (10-11y)

AC C

EP

TE D

M AN U

SC

RI PT

30

4

ACCEPTED MANUSCRIPT

involved in food-related activities (including cooking and shopping) showed more favourable

56

food intake and cooking skills than children who are not involved in food preparation

57

activities. On the contrary, in another study with female adolescents (11-18y), food shopping

58

frequency was related to greater consumption of fried foods (Larson, Story, Eisenberg, &

59

Neumark-Sztainer, 2006). Other studies investigating food shopping behaviours in childhood

60

have focused on the children’s influences on in-store purchases (O'Dougherty, Story, & Stang,

61

2006), without assessing the links with food purchasing and the quality of foods consumed by

62

children. Even if little evidence exists, shopping skills are considered by some authors as

63

essential food skills to be included in healthy eating programs (Fordyce-Voorham, 2009). To

64

the best of our knowledge, the effects of involving children in shopping activities as a strategy

65

for enhancing their eating behaviours have not been studied yet.

66

The objective of the present study was to assess, in an experimental setting, the effects of

67

involving children in food purchasing on their food choices, food intake, food liking and

68

appetite. Specifically, we examined the short-term effect of a one-hour food shopping

69

workshop on children’s willingness to select and to taste unfamiliar food items containing

70

vegetables for an afternoon snack. We hypothesized that involving children in purchasing

71

foods could help in directing their choices towards unfamiliar foods containing vegetables. As

72

exactly the same study design was used in the present study and in a previous one assessing the

73

effects of involving children in cooking (Allirot et al., 2016), we also compared the effects of

74

food purchasing involvement with cooking involvement on children’s willingness to select and

75

to taste unfamiliar food items.

76

METHODS

77

Participants

78

The study sample comprised 86 children aged between 8 and 10 years. They were recruited

79

from three conveniently selected schools in Gipuzkoa (northern Spain). Parents or legal

AC C

EP

TE D

M AN U

SC

RI PT

55

5

ACCEPTED MANUSCRIPT

guardians of children were sent a letter inviting their son/daughter to take part in the study.

81

School principals then sent the research team the contact details of all parents or legal

82

guardians who expressed an interest in being involved in the study. Afterwards, we called the

83

parents or legal guardians in order to explain the study and to request that they complete an

84

online questionnaire about food purchasing, cooking involvement, eating habits, and food

85

allergies of their children. Questions on food purchasing, children’s eating habits and food

86

allergies, were developed by the researchers. Other questions on cooking involvement, Food

87

Neophobia were extracted from other papers (Fernández-Ruiz, Claret, & Chaya, 2013). Based

88

on two items: “how often do you purchase food?” and “how often does your child go with you

89

for purchasing food?”, weekly food purchasing frequency of the child was calculated. The item

90

to assess cooking involvement (i.e. “how often does your child help you to prepare a meal?”)

91

was rated on a 4-point Likert scale with options ranging from once monthly or less (1) to more

92

than once a week (4) (van der Horst et al., 2014). Several items were also included in the

93

questionnaire in order to evaluate children’s eating habits: (a) the frequency of eating in the

94

canteen (5-point scale: from never to every school day), (b) the frequency of fruit intake (5-

95

point scale: from less than once a week to more than once a day), (c) the frequency of

96

vegetable intake (5-point scale: from less than once a week to more than once a day), and (d)

97

the liking of the vegetables that will be later used in the experiment was assessed with the

98

following questions: “does your child like beetroot / zucchini / spinach?” (6-point scale: 1 =

99

he/she hates it; 2 = he/she don’t like it; 3 = he/she likes it; 4 = he/she likes it very much; 5 =

AC C

EP

TE D

M AN U

SC

RI PT

80

100

he/she never tasted it; 6 = I don’t know).

101

Seven items from the Food Neophobia Scale (Fernández-Ruiz et al., 2013; Pliner & Hobden,

102

1992): were also used to measure children’s food neophobia: 1- my child is constantly

103

sampling new and different foods (reverse item); 2- my child does not trust new foods; 3- if my

104

child does not know what is in a food, he/she won’t try it; 4- at dinner parties, he/she will try a

6

ACCEPTED MANUSCRIPT

new food (reverse item); 5- he/she is afraid to eat things he/she has never had before; 6- he/she

106

is very particular about the foods he/she will eat; 7- he/she will eat almost anything (reverse

107

item). Based on the judgment of various authors (Cooke, Carnell, & Wardle, 2006), some items

108

were eliminated from the original version, due to lack of suitability for the child population,.

109

Parents answered the items depending on their degree of agreement with them on a 7-point

110

Likert scale, with options ranging from strongly disagree to strongly agree. Scores were

111

obtained by summing the values for each of the items. Scores could range from seven (low

112

child's trait neophobia score) to forty-nine (high child's trait neophobia score). A question

113

about children’s food allergies was also included in the questionnaire. Children presenting

114

allergies to any of the foods included in the experiment were excluded.

115

Based on a pilot study held in our laboratory, we calculated that a minimum sample size of

116

eighty four (84) subjects was necessary to observe a significant difference in willingness to

117

taste vegetables in two different conditions with a significance level of 0.05 and power of 80%.

118

The online questionnaire was sent to 99 parents of children who wanted to take part in the

119

study. From them, 93 completed the online questionnaire and three children were rejected due

120

to food allergy. Finally, 90 children aged between 8 and 10 years were included in the

121

experiment that was carried out between March and May 2016. In all cases, parents or legal

122

guardians explicitly authorized the child’s participation in the study by a written informed

123

consent. The study complied with the Second Declaration of Helsinki and it also obtained the

124

approval of the Ethical Commission of Basque Culinary Center- Mondragon Unibertsitatea

125

(005/2014).

126

Study design

127

Experimental sessions were conducted at Basque Culinary Center (BCC) in San Sebastian

128

(Spain). The study was conducted using a between-subject design. Sessions were organized in

129

group format (five children each; 18 groups altogether) and children came only once to the

AC C

EP

TE D

M AN U

SC

RI PT

105

7

ACCEPTED MANUSCRIPT

BCC to take part in the experiment. The 18 groups were randomly allocated to one of the two

131

possible conditions (i.e. PURCHASE or CONTROL). Due to absence on the day of session,

132

four children did not participate. Hence, 86 children (16 groups of 5 children and 2 groups of 4

133

children; one group per day) participated in the study. Those children showed up at the BCC

134

between 4:30pm and 5:00pm. Children were asked to refrain from eating until after the meal.

135

The experimental sessions included two different parts: (a) a one-hour purchasing food

136

workshop (in the PURCHASE condition) or a one-hour creative workshop (in the CONTROL

137

condition), and (b) the consumption of an afternoon snack (both conditions).

138

Purchasing workshop (PURCHASE condition)

139

Forty three children participated in the PURCHASE condition (purchasing workshop). The

140

purchasing workshop consisted of purchasing online the necessary ingredients for the

141

preparation of three unfamiliar food items containing vegetables: apple/beetroot juice, zucchini

142

tortilla sandwich and spinach cookies. Those ingredients are described in Table 1.

147 148 149 150 151 152 153 154

M AN U

EP

146

AC C

145

TE D

143 144

SC

RI PT

130

8

ACCEPTED MANUSCRIPT 155

Table 1: Necessary ingredients for the realisation of the purchasing workshop (PURCHASE

156

condition). Ingredients - Apples (2 kg) - Raw peeled beetroot (100 g)

Zucchini tortilla sandwich

- Unpeeled zucchini (275 g) - Eggs (5 units) - Olive Oil (2 tablespoons) - Salt (6 pinches) - Bread (1 slice of toast)

Spinach cookies

- Spinach (125 g) - Wheat flour (440 g) - Sugar (250 g) - Vanilla extract (2.5 tablespoons) - Baking powder (16 g) - Matcha green tea powder (2.5 tablespoons) - Chopped pistachios (125 g) - Softened butter (315 g) - Eggs (2.5 units)

M AN U

SC

RI PT

Name of the food item Apple/beetroot juice

TE D

157

The workshop was completed under the supervision of two researchers from the BCC in the

159

experimental kitchen of the BCC. The children from each group were positioned around a

160

table. The objective of the workshop was first explained. . Then, the co-creation process of the

161

shopping list between the children and the researchers began: (1) the children were first asked

162

to guess the ingredients necessary for each of the food items; then, (2) the correct list of

163

ingredients and the recipe for each food item were presented to the children; finally (3), the

164

shopping list, including all the ingredients of the three food items was displayed on a screen in

165

the room. After completing the shopping list, the children purchased all of the ingredients

166

online, using a HP 250 G2 (15.6”) (one per child) connected to the website of a Basque

167

supermarket chain. The children were asked to add to their virtual basket all the ingredients of

168

the shopping list. They were asked to choose the less expensive items and to select the correct

AC C

EP

158

9

ACCEPTED MANUSCRIPT

quantity. The researchers supervised this task and the children were free to ask the researchers

170

for help. Once all children had added the correct ingredients to their baskets, a real online

171

purchase was completed collectively (one per group). Finally, the real ingredients were brought

172

into the room by the researchers, who explained that they had received those bought by another

173

group of children the day before. The children were then presented each ingredient and were

174

asked to classify them per recipe: apple/beetroot juice, zucchini tortilla sandwich and spinach

175

cookies.

176

Creative workshop (CONTROL condition)

177

Forty-three children participated in the CONTROL condition (creative workshop). Two

178

researchers from the BCC were in charge of carrying out the workshop in a room furnished

179

with a table, chairs and a whiteboard. The first part of the workshop consisted of drawing a

180

dish including a vegetable or a fruit, as might be designed by a Chef, and in presenting orally

181

this drawing to the other children. The second part consisted in placing 19 photos of foods in a

182

pyramid in order to create a personal nutritional pyramid. The results were then discussed

183

between children and researchers. Finally, the last part of the workshop consisted of a game: in

184

turn, each child had to think of a food item and the others had to guess what it was, using

185

closed questions.

186

Afternoon snack (both conditions)

187

When both workshops (purchasing or creative) ended, children were invited to take a short

188

break of 15 minutes outdoors while researchers arranged the afternoon snack. Once the break

189

was finished, the children came into the room individually in order to choose their food for the

190

afternoon snack among six food items placed on a table (i.e. two types of juices, two types of

191

tortilla sandwiches and two types of cookies). Each child should pick one juice, one tortilla

192

sandwich and one cookie. Three familiar and unfamiliar food items were displayed so as to

193

evaluate the effect of the purchasing workshop on the willingness to choose an unfamiliar food

AC C

EP

TE D

M AN U

SC

RI PT

169

10

ACCEPTED MANUSCRIPT

composed of vegetables. Familiar food items were orange juice, potato tortilla sandwich and

195

chocolate cookie, while apple/beetroot juice, zucchini tortilla sandwich and spinach cookie

196

were considered as unfamiliar food items. The latter options were selected by the chefs from

197

the BCC and parents of 8-10 year old children. The description of these food items is available

198

in a previous study (Allirot et al., 2016). Nonetheless, the 86 children of this study were asked

199

if the presented food items were familiar or unfamiliar for them prior to the experimental

200

session. To assess familiarity, the percentage of affirmative responses to the question “Have

201

you ever tasted this food before today?” was considered. Results are shown in Table 2. For the

202

three a priori familiar foods the levels of familiarity were all higher than 94% with no

203

difference

204

between the PURCHASE and the CONTROL group, while the levels of familiarity for the

205

three a priori unfamiliar foods (i.e., apple/beetroot juice, zucchini tortilla sandwich and spinach

206

cookie) were 9.3%, 20.9%, and 4.6% respectively with no difference between the PURCHASE

207

and the CONTROL group.

TE D

208

Table 2: Level of familiarity with the food items proposed for the afternoon snack.

EP

209

M AN U

SC

RI PT

194

PURCHASE CONTROL group group (n = 69) (n = 68) 97.7 100 95.4 Orange juice 9.3 4.6 13.9 Apple/beetroot juice 98.8 100 97.7 Potato tortilla sandwich 20.9 18.6 23.2 Zucchini tortilla sandwich 94.1 93.0 95.3 Chocolate chip cookie 4.6 4.6 4.6 Spinach cookie 1 210 Familiarity is reported as the percentage of children answering “Yes” to

AC C

Food items

All children

Familiarity1 (%)

211

you ever tasted this food before today?”.

212

2

213

CONTROL groups.

chi-square value2

P2

χ2(1) = 2.04 0.15 χ2(1) = 2.20 0.14 2 χ (1) = 1.01 0.31 χ2(1) = 2.81 0.59 2 χ (1) = 0.21 0.65 χ2(1) = 0.00 1.00 the question: “Have

P and chi-square values are results for the Chi-square test comparing PURCHASE and

11

ACCEPTED MANUSCRIPT 214

Regarding the procedure for food selection, several considerations were made: (a) all food

216

items were placed on a table and clearly labelled, (b) white plastic plates and transparent plastic

217

cups were used to aid the recognition of food items, (c) the sequence of presenting both

218

familiar and unfamiliar food items was randomly decided, and (d) the researcher always used

219

the same questions (i.e.“For your afternoon snack, which juice do you want? You can choose

220

between an orange juice and an apple and beetroot juice. Now, which tortilla sandwich do you

221

want? You can choose between a potato tortilla sandwich and a zucchini tortilla sandwich.

222

Finally, which cookie do you want? You can choose between a chocolate cookie and a spinach

223

cookie”). No additional information concerning the food items was provided during this

224

process. All foods were placed in a tray after making the choice and children were taken to the

225

kitchen in order to eat the afternoon snack. Once all members of the group had completed their

226

selection, children were told they could eat whatever they wanted. The only rule for the

227

afternoon snack was that they could not swap food items among them or having someone

228

else’s foods. However, all children were sat together at a table and it was permitted to talk

229

among themselves or with the researchers while having the afternoon snack. Once all the

230

members of the group indicated they ended eating, they were given the option to taste the food

231

items that were not initially selected.

232

During the experimental session drinking water was permitted anytime. The session finished

233

when children declared having finished eating, usually between 6:30pm and 7:00pm. Before

234

leaving the BCC, all children were rewarded with an apron to express our gratitude for their

235

participation.

236

Measurements

237

The following measurements were taken into consideration in the PURCHASE and in the

238

CONTROL conditions.

AC C

EP

TE D

M AN U

SC

RI PT

215

12

ACCEPTED MANUSCRIPT Subjective hunger and satiety

240

During the experimental session subjective hunger and satiety was measured three times: prior

241

to the beginning of the workshop (T0), once the workshop had finished (T1) and when all the

242

members of the group asserted having ended their afternoon snack (T2). Bennet and Blisset’s

243

“Teddy the Bear” hunger and satiety scale was employed, which is adapted for primary school

244

children (Bennett & Blissett, 2014). “Teddy the Bear” is a scale where five black and white

245

cartoon bears silhouettes and their stomach area represent different levels of satiety. Black

246

circles with different sizes in the bear’s stomach area symbolize varying amounts of “food”.

247

The bigger the circle, the more food that has been ingested and consequently, the satiety level

248

of the bear. Together with the silhouettes, a writing description of the bear’s level of hunger

249

and satiety was placed on a 5-point Likert scale with options ranging from not satiated at

250

all/very hungry to not hungry at all/very full. The scale was translated to Spanish and the

251

children’s comprehension of the scale has been tested in a previous study (Allirot et al., 2016).

252

The scales were presented on a laptop HP 250 G2 (15.6”) and FIZZ Sensory Software 2.47B

253

by Biosystèmes (Couternon, France) was used for automatic data capture. The following

254

explanation was given by the researchers before the children answered the scale: “I was

255

wondering if you could tell me about how hungry you are feeling right now. If you think about

256

your own tummy and how empty or full it is right now, which Teddy would you say shows me

257

how hungry or how full you are feeling. There is no right or wrong answer; this is just about

258

how you feel.” (Bennett & Blissett, 2014).

259

Willingness to choose and taste unfamiliar foods containing vegetables

260

After the food selection task, a photograph of each child’s tray was taken in order to know

261

which food items they had chosen. This allowed us to count individually the quantity of

262

unfamiliar food items chosen spontaneously in the first selection process each child made

263

alone. Because after eating the first selected food items children were given the opportunity to

AC C

EP

TE D

M AN U

SC

RI PT

239

13

ACCEPTED MANUSCRIPT

taste the food items that were not selected at the beginning, we could determine the total

265

number of unfamiliar food items chosen and tasted throughout the experimental session.

266

Food intake estimation

267

In order to estimate children’s food intake, two photos of the trays with the selected food items

268

(lateral views) were taken exactly after the choice task and just after children had finished

269

eating their afternoon snack. Afterwards, two evaluators rated from the images the percentage

270

of intake for each food item (i.e. 10%, 25%, 50%, 75% or 100%) and consumption of juice,

271

tortilla sandwich and cookie in addition to the entire afternoon snack. Likewise, we ascertained

272

the percentage of consumption for each food item and the mean intake of the so called familiar

273

foods (orange juice, potato tortilla sandwich and chocolate cookie) and of the unfamiliar items

274

(apple/beetroot juice, zucchini tortilla sandwich and spinach cookie).

275

Liking of the food items

276

As for the liking of the food items eaten, children were requested to measure how much they

277

liked the food items tasted throughout the several stages of the session. A validated illustrative

278

5-point facial scale (de Ruyter, Katan, Kuijper, Liem, & Olthof, 2013; Leon, Couronne,

279

Marcuz, & Köster, 1999) that ranged from a broad smile (extremely liked) through a neutral

280

expression (maybe liked, maybe disliked) to a large frown (extremely disliked) was used to

281

measure liking. This scale was also presented on a laptop HP 250 G2 (15.6”). Finally, we

282

ascertained the mean liking of the so called familiar foods (orange juice, potato tortilla

283

sandwich and chocolate cookie) and for the unfamiliar ones (apple/beetroot juice, zucchini

284

tortilla sandwich and spinach cookie).

285

Statistical analysis

286

Statistical analyses were performed using JMP® 12.0.1 statistical software (SAS Institute Inc.,

287

NC, USA). Student's unpaired t-tests were carried out for the variables of age, food neophobia

288

score and the weekly frequency of food purchasing between the two groups for baseline values.

AC C

EP

TE D

M AN U

SC

RI PT

264

14

ACCEPTED MANUSCRIPT

Chi-square tests were performed for the other variables in order to detect differences between

290

the PURCHASE and the CONTROL conditions: gender distribution, cooking involvement,

291

frequency of eating in school restaurant, frequency of fruit intake, frequency of vegetable

292

intake, liking of zucchini, spinach and beetroot.

293

To assess the difference between the two groups for the percentage of children who

294

spontaneously chose at least one unfamiliar food item amongst the three proposed, a chi-square

295

test was used.

296

Separate Univariate Analysis of Co-Variance (ANCOVA) were performed to analyze the

297

differences between the PURCHASE and the CONTROL conditions for the following

298

variables: mean number of unfamiliar food items spontaneously chosen by the children, mean

299

number of unfamiliar food items children accepted to taste during the experimental session,

300

subjective hunger/satiety scores, percentages of food intake estimated by photography, and

301

liking of the food items. All ANCOVA models were adjusted for age. Likewise, the models for

302

liking of the three unfamiliar food items were also adjusted for initial liking of beetroot,

303

zucchini and spinach liking. The normality of the variables were previously checked using the

304

Shapiro–Wilk test, and the two variables: “mean number of unfamiliar food items

305

spontaneously chosen by the children” and “ mean number of unfamiliar food items children

306

accepted to taste during the experimental session” were not normally distributed. As a result,

307

for those two variables, we ran a non-parametric Wilcoxon-Mann-Whitney two sample rank

308

test to determine whether differences existed between the two conditions. Subsequently, we ran

309

a rank-based nonparametric analysis of covariance with age as a covariate, following the

310

Quade Method (Quade test) (Grouin, 2008).

311

Furthermore, Student's paired t-tests were carried out in each group to examine the variations

312

in hunger/satiety levels between T1 and T0 and between T2 and T1. As for the food intake

313

estimation, Inter-rater reliability (IRR) was assessed using Intraclass Correlation Coefficient

AC C

EP

TE D

M AN U

SC

RI PT

289

15

ACCEPTED MANUSCRIPT

(ICC) to analyze to what extent evaluators were consistent in their estimation. Mean values of

315

intake estimations obtained from both evaluators were used for ANCOVA intake analyses.

316

We also compared the effects of the food purchasing workshop with those of a cooking

317

workshop reported in a previous study (Allirot et al., 2016) (the 69 children in the previous

318

study participated in a cooking workshop: they prepared the same three unfamiliar food items

319

than those used in the present one). We used non-parametric Wilcoxon-Mann-Whitney two

320

sample rank tests and nonparametric ANCOVA based on Quade’s test, with age as a covariate

321

for comparing between the purchasing workshop and the cooking one (1) the mean number of

322

unfamiliar food items spontaneously chosen by the children and (2) the mean number of

323

unfamiliar food items children accepted to taste during the experimental session.

324

All tests were two-sided and differences were considered significant at P < 0.05. Values are

325

expressed as mean ± standard error of mean (SEM).

326

RESULTS

327

A total of 86 children (mean age = 9.0 ± 0.1) participated in the study, of whom 30% were

328

boys (n = 26) and 70% were girls (n = 60) with no difference in gender distribution between

329

groups (P = 1,00). Table 3 shows the differences in baseline variables of participants in the two

330

groups: PURCHASE and CONTROL. No differences were found between the two groups.

332 333 334 335 336 337

SC

M AN U

TE D

EP

AC C

331

RI PT

314

16

ACCEPTED MANUSCRIPT 338

Table 3: Differences in baseline variables of participants in the two groups: PURCHASE (n =

339

43) and CONTROL (n = 43). P1

Age (years; mean ± SEM)

PURCHASE CONTROL Group Group 8.9 ± 0.1 9.1 ± 0.1

t or chi-square value1 t(84) = 1.68

0.09

Gender (% girls)

70

χ²(1) = 0.00

1.00

RI PT

70 23.7 ± 1.8

t(84) = 0.36

0.71

Weekly food purchasing frequency (mean 1.34 ± 0.63 ± SEM)

1.27 ± 0.60

t(84) = 0.55

0.58

Cooking involvement ≤ 1 time a month (%) 2 or 3 times a month (%) 1 time a week (%) ≥ 1 time a week (%) Frequency of eating in school restaurant Never (%) Once or twice a month (%) Once or twice a week (%) 3 or 4 times a week (%) Each school day (%)

39.0 36.6 19.5 4.9

χ2(3) = 4.56

0.21

26.8 2.4 9.8 0.0 61.0

χ2(4) = 5.73

0.22

9.5 14.3 33.3 21.4 21.4

4.9 29.3 22.0 31.7 12.2

χ2(4) = 5.61

0.23

Frequency of vegetable intake ≤ 1 time a week (%) 1 to 3 times a week (%) 4 to 6 times a week (%) 1 time a day (%) ≥ 1 time a day (%)

7.1 40.5 45.2 7.1 0.0

4.9 58.5 26.8 7.3 2.4

χ2(4) = 4.52

0.34

Liking zucchini Hates it (%) Do not like it (%) Likes it (%) Likes it very much (%) Never tasted (%) Do not know (%)

0.0 25.6 51.2 9.3 13.9 0.0

2.4 4.9 70.7 7.3 12.2 2.4

χ2(5) = 9.38

0.09

M AN U 35.7 4.8 16.7 4.7 38.1

AC C

EP

TE D

Frequency of fruit intake ≤ 1 time a week (%) 1 to 3 times a week (%) 4 to 6 times a week (%) 1 time a day (%) ≥ 1 time a day (%)

47.6 19.1 19.0 14.3

SC

Food Neophobia Score (score from 7 to 22.9 ± 1.7 49; mean ± SEM)

17

ACCEPTED MANUSCRIPT

0.0 30.9 7.1 0.0 30.9 30.9 results for the

0.0 26.8 4.9 χ2(5) = 4.04 0.26 0.0 51.2 17.1 test comparing PURCHASE and

CONTROL groups.

χ2(5) = 6.26

0.28

M AN U

341

2.4 26.8 51.2 2.4 12.2 4.1

RI PT

Liking beetroot Hates it (%) Do not like it (%) Likes it (%) Likes it very much (%) Never tasted (%) Do not know (%) 1 340 P, t-values and chi-square values are

0.0 30.9 38.1 4.8 7.1 19.0

SC

Liking spinach Hates it (%) Do not like it (%) Likes it (%) Likes it very much (%) Never tasted (%) Do not know (%)

342

Willingness to choose and taste unfamiliar foods containing vegetables

344

In the CONTROL group, 7 children (16%) spontaneously chose at least one unfamiliar food

345

item for the afternoon snack, while 20 children (46%) did the same in the PURCHASE group:

346

χ2(1) = 9.12; P = 0.0025.

347

Figure 1A represents the mean number of unfamiliar food items that spontaneously were

348

chosen by the children for the afternoon snack in the two conditions in the first selection they

349

made individually. Results of the Wilcoxon-Mann-Whitney test showed a statistically

350

significant difference between the two conditions U = 9.78, z = 3.12, P = 0.002, in the mean

351

number of unfamiliar food items chosen, being higher in the PURCHASE condition (0.70 ±

352

0.14) compared to CONTROL (0.19 ± 0.07). This difference was still significant with the

353

Quade test, after introducing the age as a covariate F(1, 84) = 9.14, P = 0.003. No effect of the

354

age was found F(1,84) = 2,477, P = 0.119. Furthermore, in the figure 1B, the mean number of

355

unfamiliar food items tasted by the children during all the experimental session in both

356

conditions is observed, summing the first selection made individually and the second

357

opportunity to taste in group the food that were not selected at first. Results of the Wilcoxon-

AC C

EP

TE D

343

18

ACCEPTED MANUSCRIPT

Mann-Whitney test showed no difference between the PURCHASE condition (2.16 ± 0.14)

359

and the CONTROL (2.00 ± 0.14), U = 1.06, z = 1.02, P = 0.304. With age as a covariate

360

(Quade test), there was also no difference F(1,84) = 1.24, P = 0.268, nor did we find an effect

361

of the age (F(1,84) = 0.212, P = 0.647).

362

We also compared the effects of the purchasing workshop with those of a cooking workshop

363

(Allirot et al., 2016). The children chose 0.74 ± 0.13 unfamiliar food items after the cooking

364

workshop, while 0.70 ± 0.14 were chosen after the food purchasing workshop. No difference

365

was found without any covariates U = 0.09, z = 0.30, P = 0.762 and with age as a covariate

366

F(1,110) = 0.27, P = 0.601). No effect of age was found F(1,110) = 1.89, P = 0.172).

367

The overall number of unfamiliar food items tasted by the children was slightly higher in the

368

cooking condition (2.43 ± 0.10) compared to the purchasing one (2.16 ± 0.14), but this

369

difference did not reach significance without any covariates U = 3.14, z = 1.77, P = 0.077 or

370

with age as a covariate F(1,110) = 3.06, P = 0.083. No effect of age was found F(1,110) = 0.39,

371

P = 0.532).

375 376 377 378 379 380 381 382

SC

M AN U

TE D

374

EP

373

AC C

372

RI PT

358

19

ACCEPTED MANUSCRIPT 383 384 385 386

RI PT

387 388 389

SC

390 391

M AN U

392 393 394 395

TE D

396 397 398

EP

399 400

Figure 1: A - Mean number of unfamiliar food items chosen by the children in the CONTROL

402

and in the PURCHASE conditions (1st choice made alone). B - Mean number of unfamiliar

403

food items tasted by the children (i.e. chosen for the afternoon snack or accepted to be tasted

404

after the end of the snack) (1st choice made alone + 2nd opportunity to taste in group) in the

405

CONTROL and in the PURCHASE conditions.

406

* P < 0.01.

407

AC C

401

20

ACCEPTED MANUSCRIPT 408

Subjective hunger and satiety

409 410 411

RI PT

412 413 414

SC

415 416

M AN U

417 418 419 420

TE D

421

Figure 2: Hunger/Satiety score at T0 (before the beginning of the workshop), T1 (at the end of

423

the purchasing or creative workshop) and T2 (after all the children of the group declared

424

having finished their snack)1,2.

425

1

426

Hunger/Satiety.

427

2

428

PURCHASE and CONTROL conditions at T0, T1 and T2.

EP

422

AC C

Each value is a 1 (very hungry/not satiated at all) to 5 (not hungry at all/very full) score for

NS means “No statistically Significant” where P values are results for the test comparing

429 430

In figure 2, hunger/satiety scores during the experimental session can be seen. No differences

431

were found between the two conditions PURCHASE and CONTROL without any covariates:

432

at baseline (T0) (F(1, 84) = 0.02, P = 0.89), after the workshop (T1) (F(1, 84) = 0.66, P = 0.42)

21

ACCEPTED MANUSCRIPT

and at the end of the afternoon snack (T2) (F(1,84) = 0.51, P = 0.48). Results were similar with

434

age as covariate: at T0 (F(1, 84) = 0.05, P = 0.82), at T1 (F(1, 84) = 0.70, P = 0.40) and at

435

T2(F(1,84) = 0.12, P = 0.72).

436

However, we observed a statistically significant reduction of these scores as a result of an

437

increase of hunger between baseline (T0) and after having finished the workshop (T1) in both

438

CONTROL (t(42) = 2.18, P = 0.03) and PURCHASE groups (t(42) = 4.38, P < 0.001). This

439

score decrease was not different between the two conditions (F(1, 84) = 0.89, P = 0.34).

440

Inversely, after having finished the workshop (T1) and once the afternoon snack was ended

441

(T2), a statistically significant increase of the mean scores was seen corresponding to a

442

decrease on the hunger level in both CONTROL (t(42) = 7.23, P < 0.001) and PURCHASE

443

groups (t(42) = 9.40, P < 0.001). However, this score increase was not different between the

444

two conditions (F(1,84) = 0.07, P = 0.78).

445

Food intake estimation

446

IRR was examined along with ICC. Excellent ICC values were observed: ICC Juice intake = 0.98 ,

447

ICC

448

adjustment between evaluators. For all subsequent analyses, mean values of intake estimations

449

obtained from both evaluators were used.

450

In table 4, the percentage of intake for each food item in each condition is provided. Regardless

451

of the children’s selections for each food group (i.e. tortilla sandwich, juice and cookie) there

452

were no differences on intake estimations between the two conditions with ANCOVA models

453

adjusted for age (P Tortilla sandwich intake = 0.704; P Juice intake = 0.804; P = 0.493; P Cookie intake = 0.493

454

and P Whole afternoon snack = 0.593) or without any covariates. If we look at each food item intake

455

estimation one by one, no differences were found between groups.

456

Concerning the mean consumption of the familiar items (orange juice, potato tortilla sandwich

457

and chocolate cookie), no difference was observed between the PURCHASE condition (97.4 ±

TE D

= 0.96 and ICC

Cookie intake

= 0.97. These values demonstrated a high level of

AC C

EP

Tortilla intake

M AN U

SC

RI PT

433

22

ACCEPTED MANUSCRIPT

1.2 %) and the CONTROL (91.1 ± 2.8 %), with age as a covariate (F(1, 81) = 3.18, P = 0.08)

459

or without any covariates (F(1, 81) = 2.30, P = 0.133).In addition, the mean intake estimation

460

of the unfamiliar food items (apple/beetroot juice, zucchini tortilla sandwich and spinach

461

cookie) was not different between the PURCHASE (76.0 ± 7.7 %) and CONTROL condition

462

(84.4 ± 10.5 %), with age as a covariate (F(1, 24) = 0.13, P = 0.716) or without any covariates

463

(F(1, 81) = 0.38, P = 0.543).

AC C

EP

TE D

M AN U

SC

RI PT

458

ACCEPTED MANUSCRIPT

23 Table 4: Mean percentage of food intake estimated by photography in the PURCHASE group (n = 43) and in the CONTROL group (n = 43).

PURCHASE group

CONTROL group

F

Juice1 Orange Apple/Beetroot

83.2 ± 4.8 (n = 43) 90.8 ± 4.8 (n = 30) 65.8 ± 9.2 (n = 13)

81.5 ± 5.4 (n = 43) 81.4 ± 6.1 (n = 36) 82.1 ± 11.8 (n = 7)

Tortilla sandwich1 Potato Zucchini

95.4 ± 2.8 (n = 43) 99.3 ± 0.7 (n = 36) 75.7 ± 15.8 (n = 7)

Cookie1 Chocolate Spinach

97.8 ± 2.2 (n = 43) 100.0 ± 0.0 (n = 35) 88.1 ± 11.8 (n = 8)

ANCOVA models without any covariates

RI PT

ANCOVA models adjusted for age

% of intake (mean ± SEM)

P

F

P

F(1, 84) = 0.06 F(1, 64) = 1.14 F(1, 18) = 0.49

0.804 0.289 0.630

F(1, 84) = 0.06 F(1, 64) = 1.38 F(1, 18) = 0.99

0.811 0.245 0.333

93.7 ± 3.0 (n = 43) 93.7 ± 3.0 (n = 43) (n = 0)

F(1, 84) = 0.14 F(1, 77) = 2.21 -

0.704 0.141 -

F(1, 84) = 0.18 F(1, 77) = 2.72 -

0.675 0.103 -

95.9 ± 2.5 (n = 43) 95.4 ± 2.6 (n = 41) 100.0 ± 0.0 (n = 2)

F(1, 84) = 0.47 F(1, 74) = 2.17 F(1, 8) = 0.80

0.493 0.145 0.401

F(1, 84) = 0.31 F(1, 74) = 2.28 F(1, 8) = 0.23

0.577 0.135 0.645

TE D

M AN U

SC

464

Whole afternoon snack1 92.2 ± 2.2 (n = 43) 90.4 ± 2.8 (n = 43) F(1, 84) = 0.29 0.593 F(1, 84) = 0.25 0.615 1 465 Data for the juice intake, the tortilla sandwich intake, the cookie intake and the whole afternoon snack are shown independent of the type of

EP

tortilla sandwich, the type of cookie and the type of juice chosen.

AC C

466

24

ACCEPTED MANUSCRIPT 467

Liking of the food items

468

In table 5, the mean liking for each food item in each condition is shown. Regardless of the

469

children’s selections for each food group (i.e. tortilla sandwich, juice and cookie) there were no

470

differences in liking between the two conditions with ANCOVA models adjusted for age

471

(P Tortilla

472

afternoon snack liking

473

The only food item that presented a borderline significant difference on liking between the two

474

conditions was the apple/beetroot juice (F(1, 65) = 4.12, P = 0.048) (tasted and evaluated by 32

475

children in the PURCHASE condition and by 35 in the CONTROL group). The liking of this

476

food item was higher in the PURCHASE group compared to the CONTROL group. This

477

difference was still significant using ANCOVA model without any covariates (F(1, 65) = 4.26,

478

P = 0.043). Regarding the other food items’ liking one by one, there were no differences

479

between groups.

= 0.453; P

Juice liking

= 0.288; P = 0.493; P

= 0.453 and P

Whole

AC C

EP

TE D

M AN U

SC

= 0.176) or without any covariates.

Cookie liking

RI PT

sandwich liking

ACCEPTED MANUSCRIPT

25 Table 5: Mean liking in the PURCHASE condition and in the CONTROL condition.

Juice1 Orange Apple/Beetroot Tortilla sandwich1 Potato Zucchini

4.49 ± 0.14 (n = 43) 4.71 ± 0.10 (n = 42) 2.33 ± 0.30 (n = 24)

4.67 ± 0.08 (n = 43) 4.67 ± 0.09 (n = 43) 1.86 ± 0.26 (n = 21)

F(1, 84) = 1.26 F (1, 83) = 0.02 F (1, 43) = 2.01

0.265 0.881 0.163

F(1, 84) = 1.20 F (1, 83) = 0.09 F (1, 43) = 2.01

0.276 0.768 0.163

Cookie1 Chocolate Spinach

4.77 ± 0.08 (n = 43) 4.93 ± 0.04 (n = 42) 3.79 ± 0.22 (n = 38)

4.86 ± 0.10 (n = 43) 4.86 ± 0.10 (n = 43) 3.60 ± 0.28 (n = 32)

F(1, 84) = 0.57 F (1, 83) = 0.34 F (1, 68) = 0.34

0.453 0.561 0.561

F(1, 84) = 0.54 F (1, 83) = 041 F (1, 68) = 0.31

0.464 0.523 0.577

P

F

P

F(1, 84) = 1.14 F(1, 81) = 0.0001 F(1, 65) = 4.12

0.288 0.994 0.048*

F(1, 84) = 1.02 F(1, 81) = 0.06 F(1, 65) = 4.26

0.316 0.813 0.043*

M AN U

SC

F

RI PT

ANCOVA models adjusted for age2

ANCOVA models without any covariates

Liking (score between 1 and 5) (mean ± SEM) PURCHASE CONTROL group group 4.16 ± 0.14 (n = 43) 4.37 ± 0.15 (n = 43) 4.32 ± 0.15 (n = 41) 4.26 ± 0.17 (n = 42) 3.72 ± 0.20 (n = 32) 3.26 ± 0.24 (n = 35)

TE D

480

482

of tortilla sandwich, the type of cookie and the type of juice chosen.

483

2

484

liking, initial zucchini liking and initial spinach liking, respectively.

485

* P < 0.05.

EP

481

Whole afternoon snack1 4.47 ± 0.09 (n = 43) 4.63 ± 0.08 (n = 43) F(1, 84) = 1.86 0.176 F(1, 84) = 1.72 0.193 1 Data for the juice liking, the tortilla sandwich liking, the cookie intake and the whole afternoon snack liking are given independent of the type

AC C

The models for apple/beetroot juice liking, zucchini tortilla sandwich liking and spinach cookie liking were also adjusted for initial beetroot

26

ACCEPTED MANUSCRIPT 486

Concerning the mean liking of the familiar items (orange juice, potato tortilla sandwich and

488

chocolate cookie), no difference was found between the PURCHASE condition (4.66 ± 0.07)

489

and the CONTROL one (4.60 ± 0.09), with age as a covariate (F(1, 84) = 0.07, P = 0.786) or

490

without any covariates (F(1, 84) = 0.26, P = 0.609). As well as that, the mean liking of the

491

unfamiliar food items (apple/beetroot juice, zucchini tortilla sandwich and spinach cookie) was

492

not different between the PURCHASE condition (3.43 ± 0.15) and the CONTROL condition

493

(3.14 ± 0.20), with age as a covariate (F(1, 78) = 2.60, P = 0.110) or without any covariates

494

(F(1, 78) = 1.65, P = 0.203).

495

DISCUSSION

496

The present study shows that a single session of food purchasing with children has effects on

497

subsequent food choices. Children who actively participated in the food purchasing activity

498

spontaneously chose more unfamiliar foods containing vegetables compared with a control

499

group. Involving children in food shopping can thus help in directing children’s choices

500

towards novel foods containing vegetables. The present study is the first to demonstrate, in an

501

experimental setting, a beneficial effect of involving children in food shopping on their

502

spontaneous food choices.

503

This result highlights the interest of involving children in the different steps of meal

504

preparation, as a strategy for changing their food choices and reducing food neophobia. The

505

importance of children’s involvement is commonly recognized by parents. In a study by Casey

506

& Rozin (1989), 89.5% of parents rated “involving the child in the preparation and serving” as

507

an effective method for changing children food preferences. Russell et al. (2015) demonstrated

508

that parents from a less neophobic group of children were used to inviting their children to help

509

them in preparing meals, while parents from a neophobic group of children rarely reported

510

these types of behaviours.

AC C

EP

TE D

M AN U

SC

RI PT

487

27

ACCEPTED MANUSCRIPT

Most of the studies assessing the links between meal preparation involvement and children diet

512

quality focused only on cooking (Chu et al., 2014; van der Horst et al., 2014). In one

513

observation study, Nozue et al. (2016) considered shopping (jointly with cooking) as one of the

514

food-related activities positively associated with more favourable food intake in children.

515

After having demonstrated experimentally in a previous study (Allirot et al., 2016) the interest

516

of cooking involvement, the present one shows that food shopping involvement alone can also

517

promote healthy food choices in children. As exactly the same design was used in both studies

518

we can compare the effects of involvement in cooking with involvement in food purchasing ..

519

Similar results were found in both studies regarding spontaneous food choices towards

520

unfamiliar foods. The overall number of unfamiliar food items tasted by the children was

521

slightly higher in the cooking condition compared to the purchasing one but this difference did

522

not reach significance. The comparison of the results of the two studies suggests that involving

523

children in cooking vs. food purchasing activities has similar beneficial effects on children

524

food choices and willingness to taste novel foods containing vegetables. Future studies should

525

examine if those effects are cumulative.

526

As already suggested for cooking involvement, being involved in food shopping might increase

527

the willingness of children to choose unfamiliar foods and/or foods containing vegetables

528

because of the “IKEA effect” (Allirot et al., 2016; Dohle, Rall, & Siegrist, 2014). This

529

phenomenon also referred to as the “I designed it myself” might be best explained as effort

530

justification (Norton, Mochon, & Ariely, 2012). Accordingly, people like self-created objects

531

more than objects created by someone else because they have put more effort in to these

532

objects (Dohle et al., 2014). In the case of preparing food, labour increases liking, even if the

533

labour required to prepare these foods is only modest (Norton et al., 2012). In the present

534

study, we confirm (1) that this effect can influence food choices and not only modify liking and

535

(2) that it works not only with cooking involvement but also with shopping involvement,

AC C

EP

TE D

M AN U

SC

RI PT

511

28

ACCEPTED MANUSCRIPT

another step of the food preparation process. This effect could be mediated by the fact that food

537

preparation involvement stimulates positive feelings such as the sense of ownership and pride

538

(Heim, Stang, & Ireland, 2009; van der Horst et al., 2014). Creating positive experience with

539

novel foods has been shown to be an effective strategy for reducing food neophobia in children

540

(Pliner & Salvy, 2006). The beneficial effects of shopping on food choices, shown in the

541

present study, might be explained by the enjoyable and positive time spent by the children

542

during the shopping experience.

543

The food shopping effect on the willingness to choose unfamiliar foods might also be

544

explained by the increase in the familiarity of the food items during the purchasing workshop

545

compared to the control one. The role of familiarity in dietary development in children has

546

been demonstrated previously (Aldridge, Dovey, & Halford, 2009). Particularly, increasing

547

familiarity by exposing children to novel foods is a well-known strategy for reducing food

548

neophobia in children (Pliner & Salvy, 2006). By shopping, the children can increase their

549

visual familiarity as well as their contextual and categorical familiarity with the novel food

550

items (Aldridge et al., 2009).

551

The present study demonstrated that food shopping activities can increase children’s liking of

552

foods. Indeed, we showed higher liking in the PURCHASE group compared to the CONTROL

553

one, for one of the three unfamiliar foods: apple/beetroot juice. This finding is in line with

554

results found with cooking activities (Allirot et al., 2016). However, the effect on liking is

555

lower for shopping involvement compared to cooking involvement, for which liking increased

556

for two of the three unfamiliar foods and one of the three familiar foods. As for cooking

557

activities, we hypothesize that the pride of having participated in one step of the preparation

558

(Heim et al., 2009) and the positive experience of shopping together could explain such results,

559

as the context or atmosphere in which the food exposure takes place is an important factor in

560

the development of food likes and dislikes (Aldridge et al., 2009).

AC C

EP

TE D

M AN U

SC

RI PT

536

29

ACCEPTED MANUSCRIPT

In the present study, we did not demonstrate any effect of the shopping activity on appetite and

562

on the percentage of intake of the whole afternoon snack and of each of the food item. In our

563

previous study (Allirot et al., 2016), the children who cooked had a higher spinach cookie

564

intake than the others, for those who chose the spinach cookie for their afternoon snack. While

565

cooking and shopping involvement had quite similar results on spontaneous food choices, we

566

showed a greater effect of cooking on intake and liking, compared to shopping. This might be

567

explained by the greater direct contact children have with foods in the cooking activities. In the

568

shopping workshop, children had physical contacts with the real food items only at the end of

569

the activity.

570

One limitation of our study is that it involves an online food purchase rather than in-shop due

571

to the experimental setting. This could have led to a decrease in the different observed effects,

572

due to the low interaction with foods in an online context. We hypothesize that the effects on

573

food choices and liking would have been increased in a shop context. Future studies could

574

address this specific question. Moreover, it might be less likely that parents involve their

575

children when they are shopping online than when they are physically attending a supermarket.

576

Thus, even if positive results were found for exposing children to online shopping for

577

vegetables, parents should be encouraged to purchase foods with their children in shop

578

contexts, in order to increase children’s exposure to a wide variety of food items. However, as

579

online grocery shopping is rapidly growing and is predicted to increase in the future (Benn,

580

Webb, Chang, & Reidy, 2015), it is relevant to use this tool also in research. Some limitations

581

to our study may have occurred due to the fact that, for logistical reasons, randomisation was

582

done by group and not by individual. However, we calculated the design effect for the initial

583

Food Neophobia Score (Food Neophobia Score was correlated to the willingness to taste

584

unfamiliar foods (data not shown)) and we found that the study design did not lead to an

585

increase of variance on the initial Food Neophobia Score comparing to a design of same

AC C

EP

TE D

M AN U

SC

RI PT

561

30

ACCEPTED MANUSCRIPT

sample size with strictly independent observations. Socio-economic aspects have not been

587

addressed in the present study. The children were recruited in three schools with very similar

588

socioeconomic profiles (middle class), and there were no difference between the two groups

589

regarding the school of provenance of the children (data not shown). The study design does not

590

enable us to assess the impact of socio-economic status on the results. Moreover, the present

591

study should be considered as a proof of concept, as the research methods and the data do not

592

allow for extrapolation to a wider population.

593

The fact that the experimental sessions have occurred in a university in gastronomic science is

594

another limitation of the present study. A previous study (Allirot et al., 2013) compared

595

appetite scores in similar situations but diferent settings (culinary art school vs. hospital),

596

demonstrating that, even if hunger scores and desire to eat savoury foods were higher in the

597

culinary art school, the trends were exactly the same in both places. In any case, effort should

598

be focused on further studies to confirm the present results in more familiar settings (at home

599

or in a scholar environment). Because of the particular context of the study (university in

600

gastronomic science), it is relevant to propose a control activity also related to food, as it would

601

not be understood by children and/or parents to come to the BCC for a non-food related

602

activity. In other contexts, another type of control activity could have been chosen, such as

603

online shopping with non-food items. However, even if it could be questionable to expose

604

children in the control group to the food and nutrition thematic, it makes sense to compare the

605

effects on food choices of a novel approach for changing behaviours (food shopping) with the

606

one more traditionally used (nutrition education). As familiarity has been reported as the

607

percentage of children answering Yes to the closed question “Have you ever tasted this food

608

before today?”, we cannot exclude a bias in this measure due to social desirability or simply

609

because the child thought that this was the response they should give. In order to limit this

610

possible bias in the measure of food familiarity, some other measures could have been used,

AC C

EP

TE D

M AN U

SC

RI PT

586

31

ACCEPTED MANUSCRIPT

such as the current frequency of consumption (Hardman, McCrickerd, & Brunstrom, 2011) or

612

a questionnaire sent to parents with 5-point scales (Mustonen, Oerlemans, & Tuorila, 2012).

613

However, considering the process we followed for selecting familiar and unfamiliar food items,

614

involving chefs, parents and children of the Basque Country (Allirot et al., 2016), we can

615

consider that the 3 a-priori familiar food items are culturally very familiar in the Basque

616

Country and inversely for the 3 a-priori unfamiliar food items. Another limitation of the

617

present study is the fact that only short term effects were considered. Long-term effects of

618

involving children in food shopping on their food choices should be assessed in further studies.

619

However, the first time trial of an unfamiliar healthy food product is very important and is

620

considered as a crucial step in the development of liking and acceptance of healthy alternative

621

products (Birch, 1999). As a result, developing strategies for stimulating a first exposure to

622

healthy foods is of high importance specifically for high neophobic people (Schickenberg, van

623

Assema, Brug, & de Vries, 2011). Even if it was not the case for the present study, we

624

recommend designing future meal preparation studies in line with theoretical underpinnings, as

625

highlighted by Hollywood et al. (2017) in a recent review. This could help in better

626

understanding the underlying mechanisms explaining the effects of meal preparation

627

interventions.

628

CONCLUSION

629

The present study demonstrates experimentally that involving children in food purchasing

630

activities can help in directing their spontaneous food choices towards novel foods and

631

enhancing their diet quality by promoting consumption of foods containing vegetables. Those

632

results support the idea that shopping with children needs to be promoted and that shopping

633

skills are essential food skills that should be included in healthy eating programs. Though

634

further studies are necessary in other contexts, the present results do provide encouragement

AC C

EP

TE D

M AN U

SC

RI PT

611

32

ACCEPTED MANUSCRIPT 635

for parents that purchasing foods with their children can promote the selection and the

636

appreciation of healthy foods.

637 ACKNOWLEDGEMENTS

639

We would like to acknowledge the support from the Basque Government. We thank all the

640

children who participated in the study as well as their parents, principals and members of the

641

three primary schools who facilitated the recruitment. We thank the Chief Iñigo Cojo for his

642

technical help in the elaboration of the recipes. We are also grateful for the support from the

643

members of Basque Culinary Center who facilitated the different steps of the study.

644

M AN U

SC

RI PT

638

645

CONFLICT OF INTEREST

646

The authors declare no conflict of interest.

TE D

647

REFERENCES

649

Aldridge, V., Dovey, T. M., & Halford, J. C. (2009). The role of familiarity in dietary

650

development. Developmental Review, 29(1), 32-44.

651

Allirot, X., da Quinta, N., Chokupermal, K., & Urdaneta, E. (2016). Involving children in

652

cooking activities: A potential strategy for directing food choices toward novel foods

653

containing vegetables. Appetite, 103, 275-285.

654

Allirot, X., Saulais, L., Seyssel, K., Roth, H., Charrié, A., et al. (2013). An isocaloric increase

655

of eating episodes in the morning contributes to decrease energy intake at lunch in lean men.

656

Physiology & Behavior, 110-111, 169-178.

AC C

EP

648

33

ACCEPTED MANUSCRIPT Benn, Y., Webb, T. L., Chang, B. P., & Reidy, J. (2015). What information do consumers

658

consider, and how do they look for it, when shopping for groceries online? Appetite, 89, 265-

659

273.

660

Bennett, C. & Blissett, J. (2014). Measuring hunger and satiety in primary school children.

661

Validation of a new picture rating scale. Appetite, 78, 40-48.

662

Birch, L. L. (1999). Development of food preferences. Annual Review of Nutrition, 19, 41-62.

663

Bucher, T., Siegrist, M., & van der Horst, K. (2014). Vegetable variety: an effective strategy to

664

increase vegetable choice in children. Public health nutrition, 17(6), 1232-1236.

665

Casey, R. & Rozin, P. (1989). Changing children's food preferences: parent opinions. Appetite,

666

12(3), 171-182.

667

Chu, Y. L., Storey, K. E., & Veugelers, P. J. (2014). Involvement in meal preparation at home

668

is associated with better diet quality among Canadian children. Journal of Nutrition Education

669

and Behavior, 46(4), 304-308.

670

Cooke, L., Carnell, S., & Wardle, J. (2006). Food neophobia and mealtime food consumption

671

in 4-5 year old children. The international journal of behavioral nutrition and physical activity,

672

3, 14.

673

Cooke, L., Wardle, J., & Gibson, E. L. (2003). Relationship between parental report of food

674

neophobia and everyday food consumption in 2-6-year-old children. Appetite, 41(2), 205-206.

675

de Ruyter, J. C., Katan, M. B., Kuijper, L. D., Liem, D. G., & Olthof, M. R. (2013). The effect

676

of sugar-free versus sugar-sweetened beverages on satiety, liking and wanting: an 18 month

677

randomized double-blind trial in children. PLOS One, 8(10), e78039.

AC C

EP

TE D

M AN U

SC

RI PT

657

34

ACCEPTED MANUSCRIPT DeCosta, P., Moller, P., Frost, M. B., & Olsen, A. (2017). Changing children's eating

679

behaviour - A review of experimental research. Appetite, 113, 327-357.

680

Dohle, S., Rall, S., & Siegrist, M. (2014). I cooked it myself: Preparing food increases liking

681

and consumption. Food Quality and Preference, 33, 14-16.

682

Falciglia, G. A., Couch, S. C., Gribble, L. S., Pabst, S. M., & Frank, R. (2000). Food

683

neophobia in childhood affects dietary variety. Journal of the American Dietetic Association,

684

100(12), 1474-1481.

685

Fernández-Ruiz, V., Claret, A., & Chaya, C. (2013). Testing a Spanish-version of the food

686

neophobia scale. Food Quality and Preference, 28(1), 222-225.

687

Fordyce-Voorham, S. (2009). Essential food skills required in a skill-based healthy eating

688

program. Journal of the Home Economics Institute of Australia, 16(2), 16-20.

689

Fordyce-Voorham, S. (2011). Identification of essential food skills for skill-based healthful

690

eating programs in secondary schools. Journal of nutrition education and behavior, 43(2), 116-

691

122.

692

Grouin, J. (2008). Rank‐Based Nonparametric Analysis of Covariance. Wiley Encyclopedia of

693

Clinical Trials, , 1-10.

694

Hardman, C. A., McCrickerd, K., & Brunstrom, J. M. (2011). Children's familiarity with snack

695

foods changes expectations about fullness. The American Journal of Clinical Nutrition, 94(5),

696

1196-1201.

AC C

EP

TE D

M AN U

SC

RI PT

678

35

ACCEPTED MANUSCRIPT

Heim, S., Stang, J., & Ireland, M. (2009). A garden pilot project enhances fruit and vegetable

698

consumption among children. Journal of the American Dietetic Association, 109(7), 1220-

699

1226.

700

Hollywood, L., Surgenor, D., Reiks, M., McGowan, L., Lavelle, F., et al. (2017). Identification

701

of Behaviour Change Techniques applied in interventions to improve Cooking Skills and Food

702

kills among adults. Critical reviews in food science and nutrition,

703

doi:10.1080/10408398.2017.1344613.

704

Jaime, P. C. & Lock, K. (2009). Do school based food and nutrition policies improve diet and

705

reduce obesity? Preventive medicine, 48(1), 45-53.

706

Larson, N. I., Story, M., Eisenberg, M. E., & Neumark-Sztainer, D. (2006). Food preparation

707

and purchasing roles among adolescents: associations with sociodemographic characteristics

708

and diet quality. Journal of the American Dietetic Association, 106(2), 211-218.

709

Leech, R. M., McNaughton, S. A., Crawford, D. A., Campbell, K. J., Pearson, N., & Timperio,

710

A. (2014). Family food involvement and frequency of family dinner meals among Australian

711

children aged 10-12years. Cross-sectional and longitudinal associations with dietary patterns.

712

Appetite, 75, 64-70.

713

Leon, F., Couronne, T., Marcuz, M., & Köster, E. (1999). Measuring food liking in children: a

714

comparison of non verbal methods. Food Quality and Preference, 10(2), 93-100.

715

Mustonen, S., Oerlemans, P., & Tuorila, H. (2012). Familiarity with and affective responses to

716

foods in 8-11-year-old children. The role of food neophobia and parental education. Appetite,

717

58(3), 777-780.

AC C

EP

TE D

M AN U

SC

RI PT

697

36

ACCEPTED MANUSCRIPT Nelson, S. A., Corbin, M. A., & Nickols-Richardson, S. M. (2013). A call for culinary skills

719

education in childhood obesity-prevention interventions: current status and peer influences.

720

Journal of the Academy of Nutrition and Dietetics, 113(8), 1031-1036.

721

Norton, M. I., Mochon, D., & Ariely, D. (2012). The'IKEA effect': When labor leads to love.

722

Journal of Consumer Psychology, 22(3), 453-460.

723

Nozue, M., Ishida, H., Hazano, S., Nakanishi, A., Yamamoto, T., et al. (2016). Associations

724

between Japanese schoolchildren's involvement in at-home meal preparation, their food

725

intakes, and cooking skills. Nutrition research and practice, 10(3), 359-363.

726

O'Dougherty, M., Story, M., & Stang, J. (2006). Observations of parent-child co-shoppers in

727

supermarkets: children's involvement in food selections, parental yielding, and refusal

728

strategies. Journal of nutrition education and behavior, 38(3), 183-188.

729

Pliner, P. & Hobden, K. (1992). Development of a scale to measure the trait of food neophobia

730

in humans. Appetite, 19(2), 105-120.

731

Pliner, P. & Salvy, S. (2006). Food neophobia in humans. In R. Shepherd & M. Raats (Eds.),

732

The psychology of food choice (pp. 75-92). UK: CABI.

733

Russell, C. G. & Worsley, A. (2013). Why don't they like that? And can I do anything about it?

734

The nature and correlates of parents' attributions and self-efficacy beliefs about preschool

735

children's food preferences. Appetite, 66, 34-43.

736

Russell, C. G., Worsley, A., & Campbell, K. J. (2015). Strategies used by parents to influence

737

their children's food preferences. Appetite, 90, 123-130.

AC C

EP

TE D

M AN U

SC

RI PT

718

37

ACCEPTED MANUSCRIPT Schickenberg, B., van Assema, P., Brug, J., & de Vries, N. K. (2011). Product samples

739

stimulate choice of unfamiliar healthful food products. Appetite, 57(1), 197-201.

740

van der Horst, K. (2012). Overcoming picky eating. Eating enjoyment as a central aspect of

741

children's eating behaviors. Appetite, 58(2), 567-574.

742

van der Horst, K., Ferrage, A., & Rytz, A. (2014). Involving children in meal preparation.

743

Effects on food intake. Appetite, 79, 18-24.

SC

RI PT

738

AC C

EP

TE D

M AN U

744