key: cord-0899616-ve8eh3so
authors: Philippe, Kaat; Chabanet, Claire; Issanchou, Sylvie; Monnery-Patris, Sandrine
title: Child eating behaviors, parental feeding practices and food shopping motivations during the COVID-19 lockdown in France: (how) did they change?
date: 2021-01-23
journal: Appetite
DOI: 10.1016/j.appet.2021.105132
sha: b84b40cedb98d3dbd7657cd988516fe21e50db39
doc_id: 899616
cord_uid: ve8eh3so

The COVID-19 pandemic caused France to impose a strict lockdown, affecting families’ habits in many domains. This study evaluated possible changes in child eating behaviors, parental feeding practices, and parental motivations when buying food during the lockdown, compared to the period before the lockdown. Parents of 498 children aged 3-12 years (238 boys; M=7.32; SD=2.27) completed an online survey with items from validated questionnaires (e.g., CEDQ, CEBQ, HomeSTEAD). They reported on their (child’s) current situation during the lockdown, and retrospectively on the period before the lockdown. Many parents reported changes in child eating behaviors, feeding practices, and food shopping motivations. When changes occurred, child appetite, food enjoyment, food responsiveness and emotional overeating significantly increased during the lockdown. Increased child boredom significantly predicted increased food responsiveness, emotional overeating and snack frequency in between meals. When parents changed their practices, they generally became more permissive: less rules, more soothing with food, more child autonomy. They bought pleasurable and sustainable foods more frequently, prepared more home-cooked meals and cooked more with the child. Level of education and increased stress level predicted changes in parental practices and motivations. This study provides insights in factors that can induce positive and negative changes in families’ eating, feeding and cooking behaviors. This can stimulate future studies and interventions.

The lockdown forced people to adapt their everyday behaviors to the new situation, including 47 their food-related behaviors. This particular situation stimulated many researchers to study the 48 impact of the lockdown on eating behaviors. Most studies have been conducted with 49 adolescents or adults. For example, Di Renzo and colleagues (2020) studied eating habits and 50 lifestyles changes during the lockdown among the Italian population (aged between 12-86 51 years). Marty and colleagues (2021) studied how changes in French adults' food choice 52 motives were related to changes in nutritional quality during the lockdown compared to the 53 period before the lockdown. Pietrobelli and colleagues (2020) conducted a study in Italy on 54 eating behavior with parents of children aged 6-18 years, but the sample was very small 55 (N=41) and the children all had obesity. 56

The current study is original and complementary to these researches as it focused specifically 57 on changes in children's eating behaviors and families' feeding practices during the 58 lockdown, compared to the period before the lockdown. 59

Since schools were closed and most people had to work from home or were technically 60 unemployed, many children and adults had to consume all their meals at home. Parents were 61 consequently responsible for their child's food intake throughout the whole day, and this 62 could be challenging in terms of time (additional meal planning, food shopping, food 63 preparation), especially for those parents who were still working. The pandemic also faced 64 some parents with changed accessibility and availability of foods and food insecurity, in 65 particular those parents who were financially vulnerable (Loopstra, 2020) . e.g., My child is always asking for food), and emotional overeating (four items; e.g., My child 146 eats more when anxious). Parents rated their agreement with each item on a five-point Likert-147 like scale (Never, Rarely, Sometimes, Often, Always), for both the period before and during 148 the lockdown. For emotional overeating, we also added a sixth answer option: not applicable, 149 as we were not sure if all children would have already presented all emotions (worried, 150 annoyed, anxious, boredom) during the lockdown. Higher scores indicated higher food 151 responsiveness and more emotional overeating. 152

In France, the mid-afternoon snack ("goûter") is a common practice and is perceived as an 154 additional meal beside breakfast, lunch and dinner, especially in children (Francou & Hébel, 155 2017) . We therefore distinguished between the frequency of the mid-afternoon snack (which 156 usually also includes a drink) and the frequency of other snacks/drinks in between meals. We 157 clearly explained the difference between both types of snacking occasions to parents in the 158 instructions of the questions. For the mid-afternoon snack, parents were asked to rate the 159 child's frequency of this snacking occasion on a four-point scale (Less than once a week, 1-3 160 times per week, 4-6 times per week, Every day), for both the period before and during the 161 lockdown. For other snacks/drinks, parents rated the frequency on a seven-point scale (Less 162 than once a week, 1-3 times per week, 4-6 times per week, once per day, Twice a day, Three 163 times a day, 4 or more times a day), also for both the period before and during the lockdown. 164 We gave examples of possible snacks/drinks (e.g., candy, piece of bread, fruit, compote, 165 yoghurt, salty or sweet biscuits) to illustrate that any food and drink, except water, should be 166 counted as a snack/drink. 167

We asked parents as well about the types of foods their child usually consumed during snack 168 times: "When your child has a mid-afternoon snack or a snack/drink in between meals, how 169 often does (s)he consume the following types of foods and drinks?". The frequency of each 170 type of food/drink (Table 4 ) was rated on a five-point-Likert scale (Never, Rarely, Sometimes, 171

Often, Always), for both the period before and during the lockdown. The selection of the types 172 of foods and drinks was based on the food groups presented in a French food consumption 173 report (ANSES, 2017). 174

Parents were asked to report how often their child was bored at home on a five-point during the lockdown. Higher scores indicated the use of more soothing with foods, more child 197 autonomy, more rules and limits, a stricter meal setting, and a more positive meal atmosphere. 198

The items were translated from English to French by several researchers of the team, and 199 some questions were slightly modified; to adapt them to the French situation (e.g., mid-200 afternoon snack "goûter" vs. other snacks/drinks) or to be more uniform within the entire 201 questionnaire (Appendix A). 202

One additional feeding practice "Feeding on a schedule" was selected for this study. 

Scores were calculated for each dimension by averaging the scores of the corresponding 279 items, for the period of the lockdown, and for the period before the lockdown. For the 280 dimensions emotional overeating and soothing with food, the answer option "not applicable" 281 was coded as missing value. For emotional overeating, 22 parents responded with "not 282 applicable" to all corresponding items, and for soothing with food, six parents responded with 283 "not applicable" to all items. These parents thus did not report changes in this 284 behavior/practice during the lockdown compared to before the lockdown. Proportions of 285

individuals showing a change (score during lockdown -score before lockdown ≠ 0) were calculated for 286 each child behavior and each parental feeding practice. For those children/ parents for whom 287 changes were reported, paired-samples t-tests were conducted for each behavior/practice in 288 order to compare mean scores of both periods (M during lockdown -M before lockdown ). Simple 289 regressions were performed to study the effects of changes in level of child boredom at home, 290 child age, child sex, and child z-BMI (as a continuous variable) on changes in child eating 291 behaviors. Simple regressions were also used to study the effects of parental demographics 292 (parent's sex, BMI, relationship status, level of education, work status during lockdown, 293 perception of financial status) and changes in parental stress levels at home, on changes in 294 parental feeding practices, changes in parental motivations for buying foods, and on changes 295 in parental cooking behaviors. Whenever the results of these simple regressions indicated 296 multiple significant predictors for a given dependent variable, we subsequently performed a 297 multiple regression analysis to verify if the relations remained significant after controlling for 298 the effects of the other predictors. In all regression analyses, the dependent variables only 299 included the children/parents for whom changes in their behaviors, practices or motivations 300 were reported. This approach was chosen since this study was specifically designed to focus 301 on possible predictors of the observed changes, but also for statistical reasons (i.e., to meet the 302 assumption of normality, and to maintain a homogenous variance). The significance level was 303 set at p < 0.05 for all analyses. Our analytic plan was pre-specified in our study file and 304 submitted to the ethical committee before the data were collected. 305 3 Results 306

A sample of 498 parents of children aged 3.0-12.3 years (47.8% boys; M age = 7.3; SD = 2.2) 308 was retained for analyses after data cleaning. The demographics for the parents are presented 309 in Table 1 . According to parental reports of child weight and height, 8% of children aged 3.0-310 5.0 years had underweight (z-BMI < -2), 68% had a normal weight (-2 ≤ z-BMI < 1), 18% 311 were at risk for overweight (1 ≤ z-BMI < 2), 5% had overweight (2 ≤ z-BMI < 3), and 1% had 312 obesity (z-BMI > 3) (categories derived from WHO, 2006) . Among the children aged 5.1-313 12.3 years, 6% had underweight (z-BMI < -2), 69% had a normal weight (-2 ≤ z-BMI < 1), 314 15% had overweight (1 ≤ z-BMI < 2), and 9% had obesity (z-BMI > 2) ( Sixty percent of parents reported a change on at least one dimension of their child's eating 331 behaviors during the lockdown compared to the period before the lockdown. When looking 332 only at the children with changed behaviors, paired-samples t-tests resulted in a significant 333 increase for all behaviors but food pickiness ( Table 2 ). The highest increases in mean score 334 were observed for emotional eating (+0.61) and for food responsiveness (+0.44). 335 

In this study, two types of snacking were studied: the mid-afternoon snack (perceived as a 345 meal for children in France) and snacks/drinks in between meals. The frequency of the mid-346 afternoon snack increased in 15% of children (during versus before the lockdown), decreased 347 in 9%, and did not change in 76% of children. The majority of children already had a daily 348 mid-afternoon snack before the lockdown, and maintained this habit during the lockdown 349 (Table 3) . Parents reported an increase in snack frequency in between meals in 36% of 350 children, a decrease in 4% of children, and no change in 60% of children. 351 Table 3 352 Frequency of mid-afternoon snacks and of snacks/drinks in between meals for all children and all 353 parents (N = 498), before and during the lockdown. 

Concerning the types of foods consumed by the children during (mid-afternoon) snack 356 occasions, 66% of parents reported at least one change in consumption during the lockdown 357 versus before. When studying only the children with a change in their consumption, paired-358 samples t-tests resulted in a statistically significant increase in mean scores (M during lockdown -359 M before lockdown) for candy/chocolate, fruit juices, sodas, chips/salty biscuits, ice creams, 360 pastries/cake/sweet cookies, cream dessert, milks, yoghurt/cheese/quark, fresh and dried 361 fruits, and nuts. A significant decrease in the consumption of compote/fruits in syrup was 362 observed (Table 4) . 363 Table 4 364 Snacking frequency: percentage of total sample of parents (N = 498) reporting a change for their child 365 (%), mean scores before and during the lockdown (M before and M during) for these children with 

Forty-five percent of parents reported no change in their child's level of boredom at home 373 during the lockdown compared to the period before the lockdown, 53% reported an increase 374 in level of boredom, and 2% a decrease. A paired-samples t-test performed on the scores of 375 the children for whom changes were reported (n = 276) indicated a significant increase in 376 mean score of level of boredom (+1.20, t(275) = 26.82, p < 0.001; M before = 2.28, SD before = 377 0.67; M during = 3.48, SD during = 0.70). 378

Simple regressions indicated that a higher increase in children's level of boredom at home 379 (during vs. before lockdown) was significantly linked with a higher increase in emotional 380 overeating, in food responsiveness and in snack frequency in between meals (Table 5) . Simple 381 regressions also indicated that child age, child sex and child z-BMI were not significant 382 predictors for changes in child boredom levels, neither for changes in child (mid-afternoon) 383 snack frequency, nor for changes in child eating behaviors, except for a significantly lower 384 increase in food responsiveness in children with higher BMI z-scores (β = -0.07, t = -2.96, p < 385 0.001). The results of these regression analyses, significant and non-significant, can be found 386 in Appendix B.1. 387 

Sixty percent of parents reported at least one change in their feeding practices during 395 lockdown compared to the period before the lockdown. When including only the parents who 396 reported a change, paired-samples t-tests resulted in a significant increase in mean scores for 397 soothing with food, guided choices -when, what and amount, and meal atmosphere. A 398 significant decrease was observed for rules and limits around unhealthy foods, meal setting, 399 and feeding on a schedule ( Table 6 ). The highest increases in mean score were observed for 400 soothing with food (+0.43) and guided choices -when (+0.36), the highest decrease was 401 observed for feeding on a schedule (-0.40). 402 403 J o u r n a l P r e -p r o o f 

Eighty-five percent of parents reported at least one change in their motivations to buy and 415 prepare certain foods for their child(ren) during the lockdown compared to the period before 416 the lockdown. For each motivation dimension, proportions of parents who reported no 417 change, a decrease, or an increase are presented in Fig. 1 

were observed for buying pleasurable and sustainable foods. The highest decrease in 419 motivation was observed for buying convenient foods. 420 

The frequency of the mid-afternoon snack increased in 35% of parents (during versus before 426 the lockdown), decreased in 4%, and did not change in 61% of parents. Thirty-one percent of 427 parents reported an increase in their snack frequency in between meals, 8% reported a 428 decrease, and 62% no change. The frequencies of both snack occasions in parents before and 429 during the lockdown are presented in Table 3 . When asked if the lockdown and the 430 accompanying emotions (e.g., boredom, stress, anxiety) induced parents to have more, the 431 same or less desire to eat during the lockdown than before, 46% of parents answered that they 432 felt more like eating than before, 41% of parents reported no change, and 14% of parents 433 reported feeling less like eating than before. 434

When asked about the preparation of homemade dishes, 66% of parents reported preparing 435 more homemade dishes than before, 30% reported no change, and 4% of parents reported 436 preparing less homemade dishes. When asked about the preparation of comforting foods or 437 recipes, 57% of parents reported preparing more comforting foods or recipes, 40% reported 438 no change, and 3% reported preparing less. When asked about the time they spent cooking 439 with their child(ren), 71% of parents reported spending more time cooking with their 440 child(ren), 26% reported no change, and 2 % reported spending less time cooking together. 441

Effects of changes in parental stress level on parental feeding practices 443

Forty-four percent of parents reported no change in their level of stress at home during the 444 lockdown compared to the period before the lockdown. An increase in level of stress was 445 reported by 42% of parents and a decrease by 14%. A paired-samples t-test performed on the 446 scores of the parents with a change in their stress level (n = 280), indicated a significant 447 increase in mean score of stress level with +0.59 (t(279) = 7.70, p < 0.001; M before = 2.74, 448 SD before = 0.86; M during = 3.33, SD during = 0.93). 449

Simple regressions indicated that higher increases in stress level were linked with higher 450 increases in guided choice -amount (more autonomy for the child to decide the amount of 451 intake) (Table 7) : on average, guided choice -amount increased during the lockdown (Table  452 6), and this increase was even higher if stress level increased. Also, on average, the meal time 453

atmosphere quality improved during the lockdown (Table 6 ), but not for those parents who 454 became more stressed at home (Table 7) . More specifically, compared to the period before the 455 lockdown, there was no improvement in meal atmosphere quality if parents' stress level 456 increased by one unit, and there was a decrease in atmosphere quality if the stress level 457 increased by more than one unit. 458 

Effects of parental demographics on changes in parental feeding practices 464

Some parental demographics were also identified as significant predictors of changes in 465 parental feeding practices. Simple regressions indicated that the decrease in rules and limits 466 around unhealthy foods (Table 6) 

Some parental demographics were also identified as significant predictors of changes in 494 parental motivations for buying foods for their child(ren). Employment status during the 495 lockdown significantly predicted changes in the motivation to buy convenient foods: parents 496 who were working from home (β = -0.54, t = -3.18, p < 0.001) and parents who were at home 497 without work (β = -0.41, t = -2.41, p = 0.02) showed a significant decrease in this motivation, 498 while parents working outside the home showed no significant change in this motivation. In 499 simple regressions, parental level of education (β = -0.11, t = -2.18, p = 0.03) and parent BMI 500 (β = 0.03, t = 2.05, p = 0.04) also significantly predicted changes in the motivation for buying 501 convenient foods. However, in a multiple regression including these three predictors (work 502 status, level of education, parent BMI), only the effect of work status remained significant 503 when adjusted for the effects of these other predictors. 504

Furthermore, in simple regressions, parents with a higher level of education showed a higher 505 increase in the motivation to buy healthy foods (β = 0.13, t = 3.25, p < 0.001), foods linked to 506 weight control (β = 0.12, t = 2.37, p = 0.02), comforting foods (β = 0.12, t = 2.28, p = 0.02), 507 and sustainable foods (β = 0.17, t = 5.04, p < 0.001) than parents with a lower level of 508 education. In a simple regression model, perceived financial status also significantly predicted 509 changes in the motivation to buy foods related to weight control (β = 0.13, t = 2.08, p = 0.04), 510 but in a multiple regression model, both the effects of level of education and financial status 511 became non-significant after adjustment for each other's effect. Also, in simple regressions, 512 parents with a more comfortable perceived financial status showed a higher increase in the 513 motivation to buy sustainable foods (β = 0.14, t = 3.19, p < 0.001) and single parents showed 514 a lower increase in this motivation (β = -0.37, t = -2.57, p = 0.01) compared to parents with a 515 less comfortable financial status and parents with a partner. In a multiple regression, level of 516 education and family situation ("single parent") remained significant predictors for 517 sustainability after adjusting for each other's effects, but not financial status. Finally, parents 518 with a higher BMI showed a lower increase in the motivation to buy foods that can easily be 519 This study wanted to evaluate possible changes in eating and feeding habits in families with 524 young children during the COVID-19 lockdown in France, versus the period before the 525 lockdown. The results showed that not all, but a majority of parents reported some changes in 526 their child's eating behaviors, in their feeding practices, their food shopping motivations, and 527 in their own eating and cooking behaviors. This clearly indicates that the lockdown had an 528 important impact on families' eating and feeding habits at home. 529

Children showed significant increases in "food approach" behaviors during the lockdown 530 (behaviors involving a movement toward or a desire for foods: i.e. food enjoyment, emotional 531 overeating, food responsiveness (Vandeweghe, Vervoort, Verbeken, Moens, & Braet, 2016; 532

Webber, Cooke, Hill, & Wardle, 2010)). Children's snack frequency in between meals also 533 increased significantly. Moreover, increases in emotional overeating, food responsiveness and 534 snack frequency were predicted by an increase in child boredom at home: children may have 535 tried to "fill up" their time with eating or found comfort and enjoyment in food during this 536 unusual, monotonous period. In children, the literature related to bored-eating is scarce and 537 the construct is often lumped together in questionnaires with emotional-and stress-eating 538 (e.g., in CDEBQ, CEBQ). In this study, we also studied emotional overeating in a more 539 general way with the CEBQ (four items studying overeating in response to both boredom, 540 anxiety, annoyment, and worry). However, recent studies have indicated that bored-eating is 541 viewed as a distinct construct by mothers, and may be a more common practice in children 542 than emotional-or stress-eating. Therefore, the authors suggested that it may be of interest to 543 present and to study bored-eating separately from other emotions ( boredom has previously been found to increase the desire to eat unhealthily (e.g., Moynihan et 546 al., 2015) . Similar to the results in adults, our results showed that increased boredom in 547 children was strongly related to increased food responsiveness, increased emotional 548 overeating and increased snack frequency. Our study thus showed that also in (young) 549 children boredom can play a role in their desire for foods. 550

Moreover, even though the COVID-19 lockdown was an unusual situation, the increased 551 manifestation of these food approach behaviors and their link with child boredom could be 552 cause for concern. It suggests that these children did not merely rely on their internal cues of 553 hunger and satiety when asking for foods/drinks (crucial for an optimal self-regulation of food boys' short-term energy compensation improved following a 4-week intervention. 574

The results of our study further showed that when feeding practices were adapted, there was a 575 significant trend to more permissive, child-centered and pleasure-oriented practices: parents 576 reported less rules and limits, more soothing with food and gave more autonomy to the child 577 in deciding when, what, how much and where to eat. Regarding the types of foods offered 578 during snacking, we also observed increased intake of so-called "comfort foods". The theory 579 of division of autonomy states that parents should be mainly responsible for what, when and 580 where the child eats, but the child for the amount of food eaten (Satter, 1990; Vaughn et al., 581 2015) . Here, we could thus argue that parents may have become a bit too permissive 582 regarding the types of foods offered during the lockdown, there was also a significant 583 decrease in structure of the meals (timing of meals, place). By contrast, the increases in 584 guided choices (i.e., more child autonomy) may indicate that parents had the opportunity to 585 listen better to children's needs and demands, and to respond to them in a more responsive 586 way (even though we are aware that these child demands were not only based on children's 587 internal cues, as discussed above). Interestingly, our results also showed that parental level of 588 stress played a role in changes in parental feeding practices during the lockdown: higher 589 increases in stress predicted higher increases in giving autonomy to the child regarding the 590 amount to eat, and no improvement in meal atmosphere quality (in contrast to parents with no 591 increases in stress). 592 Furthermore, parents showed many changes in their motivations when buying foods for their 593 children. Highest increases in motivations were observed for buying pleasurable foods, 594 sustainable foods, natural foods and healthy foods. These findings are in accordance with the 595 findings of a French survey that was carried out by Ipsos during the lockdown in April 2020 596 for L'Observatoire E.Leclerc des nouvelles consommations: they found that French 597 consumers aged 16-75 years turned more to products of French origin (45%), fresh products 598 (37%) or products from short circuits (37%). Sixty-three percent of consumers claimed that 599 they consumed more local products in order to support the local economy during the 600 lockdown. For the parents in our study, pleasure also became an important motivation, and 601 this is in line with the observed increases in snack frequency in both parents and children, 602 increased emotional eating in both, and the increase in the preparation of comforting 603 foods/recipes during the lockdown. From a cultural point of view, family meals in France 604

were already known to be strongly pleasure-oriented (Lhuissier et al., 2013) , and the 605 lockdown seemed to have reinforced this. Convenience became less important for many 606 parents, which can be supported by their reported increase in the preparation of home-cooked 607 meals and their increase in time cooking with their children. Di Renzo and colleagues (2020) 608 also observed this increase in homemade recipes during the lockdown in Italy. 609

In the present study, parental motivations for buying foods for their child(ren), changes in 610 parental feeding practices and parental cooking behaviors were significantly predicted by 611 parental characteristics. We observed that especially a higher level of education was linked to 612 some more favorable changes in behaviors: for example, maintaining to eat at set times, 613 buying more sustainable and healthy foods, more cooking with the child, preparing more 614 homemade dishes (marginal effect: p = 0.06). These results may imply that it is of interest to 615 take into account parental level of education when planning interventions to improve parental 616 feeding behaviors. Parents with different levels of education may experience different barriers 617 and facilitators for changing their behaviors. It seems that, during the lockdown, increased 618 time at home could have played a role in facilitating cooking with the child, preparing 619 homemade dishes and buying more local, sustainable foods, but more particularly for parents 620 with higher levels of education. Previous studies have already shown that parental education 621 level is linked to differences in parental feeding practices and in parental motivations when 622 buying foods for their child. For instance, parents with lower levels of education tend to be 623 less concerned by health and more concerned by children's preferences when buying foods 624 (Rigal, Champel, Hébel, Lahlou, 2019), they serve larger portion sizes (Hébel, 2017; Rigal et 625 al., 2019) and are less likely to restrict their child's intake of unhealthy foods (Wijtzes,  626 Jansen, Jansen, Jaddoe, Hofman, Raat, 2013). 627

The COVID-19 pandemic has changed our habits in many ways during the lockdown, but 628 even after months, we have not gone back to the situation "before the pandemic". As we are 629 still reshaping some of our habits, we suggest that future research and policy makers also 630 focus on the implications for the food domain in all its facets, this by also taking into account 631 possible facilitators and barriers linked to people's socio-demographic characteristics. 632

We acknowledge that there were several limitations to this study. First, parental practices and 633 behaviors were self-reported in this study and may be subject to social desirability bias even 634 though the questionnaires were anonymous. The children's eating behaviors and level of 635 boredom were also parent-reported and thus reflected the parent's perception. Second, the 636 data obtained about the period before the lockdown was reported retrospectively, possibly 637 leading to a recall bias that can threaten the internal validity of our study (Delgado-638

Rodriguez & Llorca, 2014; Hassan et al., 2005 ). Yet, recall accuracy diminishes with 639 increasing time gap, and as the time gap in this study was very small (max. eight weeks), we 640 think the recall bias was limited here. Here, we also want to note that we did not define "the 641 period before the lockdown" for the parents. It is therefore possible that parents interpreted 642 this period in different ways (more or less broad) and thus responded differently based on 643 their own interpretation, with possible corresponding effects on our results. We hope, 644 however, that the differential interpretations would be limited because of the high contrast 645 between the two periods parents needed to report on: the "normal" life and related general 646 habits right before the lockdown versus those during the lockdown. 647

Meanwhile, this study also has several strengths. To our knowledge, it is the only study that 648 looked in a more systemic way at changes in families' food habits during the COVID-19 649 lockdown, including eating and cooking behaviors, parental feeding practices and parental 650 motivations when buying foods for the family. Other studies tend to focus uniquely on adults 651 or on children. Our sample may not be entirely representative of the national population in 652

France: there was for example a relatively small sample of parents with a low level of 653 education (33.5% in our sample compared to approximately 55% in the French population 654 (Insee, 2016) ), and the majority of our participants were female (71.7%). However, we 655 managed to recruit parents with diverse profiles, also in terms of work status, perceived 656 financial situation, relationship status, and BMI categories (of both children and adults) that 657

were very close to representativeness in the French population (Argouarc'h & Picard, 2018; 658 Verdot, Torres, Salanave, Deschamps, 2017). This enabled us to obtain a broad idea of the 659 changes in eating and feeding habits in young children and their parents in France, and of the 660 parental characteristics that were linked to these changes. 661

This study provided unique insights into how a drastic change in habits is accompanied by 663 changes in eating and feeding habits both on parent and child level. The unusual situation 664 drove some parents to turn a blind eye to the usual feeding rules, and to privilege enjoyment 665 and comfort at home. Changes in child boredom and parental stress were found to influence 666 eating and feeding behaviors, and some parental characteristics were identified as possible 667 barriers and facilitators for eating, feeding and cooking behaviors. These insights could be 668 useful for future studies and interventions, and could be of interest to policy makers. 669

Qualitative studies that reflect the experiences of parents and children during the lockdown 670 could also be interesting to complement our results. They could provide us, for example, with 671 more insights into reasons why eating behaviors, feeding practices and food shopping 672 motivations have changed or not, and if the lockdown and the accompanying changes have 673 had an impact on families' food habits on a longer term and why. 674

The authors would like to thank the independent ethical committee Inserm for accepting to 676 review their study protocol in a very limited time span. They also thank the participants for 677 their interest in the study. Answer options ranged from "never" to "always" for all items in the French questionnaire. Items in italics are the original English items, unless described otherwise. ; (R) = reversed item Cronbach's alphas were acceptable if >0.60, borderline if between 0.50-0.60, and low if <0.50 All the items have a p-value ≤ 0.001 except the item Amount2bef, the p-value of which is 0.021. a For the dimension Rules and limits around unhealthy foods, the original item "How often do you restrict (or try to restrict) your child's access to fruit juice?" was deleted from the start because fruit juices are viewed as ambiguous in terms of healthiness in France. b For the dimension Guided choices: amount of food eating, the original item "I know better than my child if she or he is hungry or full" (R) was deleted from the start.; The CFAs also indicated that it was better to delete item Amount3. c For the dimension Feeding on a schedule, equal loadings were imposed in order to avoid a non-admissible solution (loadings higher than 1 or lower than -1, and negative variances work refers to those parents who were at home without work; e.g., those who are technically unemployed due to the lockdown, parents on parental leave, students, etc. b Perceived financial status ranges from less to more comfortable. c No longer significant after adjustment for the other significant predictors (multiple regression). * Remains significant after adjustment for the other significant predictors (multiple regression). 

Dealing with problematic eating behaviour

The effects of a mindfulness-based intervention on eating behaviour, food cravings, 689 dichotomous thinking and body image concern

Troisième étude individuelle nationale des consommations alimentaires

Actualisation de la base de données des consommations alimentaires et de 693

estimation des apports nutritionnels des individus vivant en France

Travail (ANSES)

Les niveaux de vie en 2016. La prime d'activité soutients 697

Maternal feeding practices and beliefs and their relationships 701 to overweight in early childhood

Development of food preferences. Annual review of nutrition

A pilot evaluation of appetite-706 awareness training in the treatment of childhood overweight and obesity: A preliminary 707 investigation

A new incremental fit index for general structural equation models

Two novel treatments to reduce overeating in overweight children: A randomized 713 controlled trial

Appetite awareness training: A cognitive behavioral 716 intervention for binge eating

Development of a WHO growth reference for school-aged children and adolescents

Eating 724 habits and lifestyle changes during COVID-19 lockdown: an Italian survey

Feeling bad or feeling good, 727 does emotion affect your consumption of food? A meta-analysis of the experimental 728 evidence

Relationship 731 between Portion Size and Energy Intake among Infants and Toddlers: Evidence of Self-732

Le goûter en perte de vitesse et loin des recommandations

Consommation et Modes de Vie

Recall bias can be a threat to retrospective and prospective research 737 designs

Low-income women's 739 conceptualizations of emotional-and stress-eating

Nouvelles données sur les déterminants des quantités consommées

France, portrait social, édition 2016. Insee Références, édition 744 2016

Behavioral and emotional disorders in children during the COVID-747 19 epidemic

Improving preschoolers' self-regulation of energy intake

Confirmatory factor analysis of the child-feeding questionnaire among parents 752 of adolescents

Caloric compensation and eating in the absence of hunger in 5-to 12-y-old weight-755 discordant siblings

Mindfulness-based eating awareness training for 758 treating binge eating disorder: the conceptual foundation

Who still eats three meals a day? Findings from a quantitative 762 survey in the Paris area

May 6) COVID-19 et 764 consommation : 57% des Français accordent davantage d'importance au prix

Vulnerability to food insecurity since the COVID-19

Improving self-regulation for obesity prevention in head 773 start: a randomized controlled trial

Stress, emotional eating behaviour and dietary 776 patterns in children

Development of a new questionnaire to assess the links between children's self-779 regulation of eating and related parental feeding practices

Eaten up by boredom: Consuming food to escape awareness of the bored 783 self

lockdown on lifestyle behaviors in children with obesity living in

Trends in energy intake among US children by eating 789 location and food source

R: A language and environment for statistical computing. R Foundation 792 for Statistical Computing

Development 794 and preliminary testing of a technology-enhanced intervention to improve energy intake 795 regulation in children

Links between maternal 797 feeding practices and children's eating difficulties. Validation of French tools

Food portion at ages 8-11 and 800 obesogeny: The amount of food given to children varies with the mother's education and 801 the child's appetite arousal

Food craving: new contributions on its 804 assessment, moderators, and consequences

The feeding relationship: problems and interventions. The Journal of 807 pediatrics

Food approach 809 and food avoidance in young children: Relation with reward sensitivity and punishment 810 sensitivity

Development of a comprehensive assessment of food parenting practices: The home 813 self-administered tool for environmental assessment of activity and diet family food 814 practices survey

Fundamental constructs in food parenting practices: a content map to guide future 819 research

Cronbach's alphas for dimensions parental feeding practices (before and during lockdown) and final item loadings in confirmatory factor analyses (CFA) (before and during lockdown).

Loading "before lockdown"Loading "during lockdown" Soothing with food (αbef = 0.81 (acceptable); αdur = 0.77 (acceptable)) Sooth1. Quand mon enfant s'ennuie ou s'inquiète, je lui propose de manger ou boire quelque chose même si je sais qu'il/elle n'a pas faim. (