1 Introduction

Vegetable consumption below recommended levels among children has long been an issue in many countries, including Finland (Hamner et al. 2023; Koivuniemi et al. 2022; Kyttälä et al. 2010; Lynch et al. 2014). Promoting the consumption of vegetables in early childhood is an important part of establishing lasting healthy eating habits. Therefore, the determinants of children's vegetable consumption are an important area of research. Parents shape both the physical and psychosocial home food environment and thus have the power to influence what, how much, when, and how children eat (Niermann et al. 2018; Ventura and Birch 2008; Wardle and Cooke 2008).

Constructs of parenting style and food parenting practices have been associated with preschool-aged children's (2–6 years) food consumption (Burnett et al. 2020; Yee et al. 2017). Research has identified numerous food parenting practices, which are specific actions parents implement within a food-related context, typically aimed at influencing children's eating (Kremers et al. 2013; Vaughn et al. 2016). Regarding preschool-aged children's vegetable consumption, controlling practices—such as pressure to eat—have been associated with lower consumption (Blissett 2011; Wardle et al. 2005). In contrast, practices like modelling healthy eating, making healthier foods available and accessible for the child at home, encouraging healthier intake, and involving the child in food preparation have been associated with higher consumption (Blissett 2011; Jarman et al. 2022; Kristiansen et al. 2017; Mazza et al. 2022; Yee et al. 2017). The evidence is more consistent for some practices than others, and most of the studies have been cross-sectional and relied on parent-report. Many of the practices associated with healthier food consumption are based on increasing exposure to healthier foods, and the effectiveness of repeated exposure is well-documented (Holley et al. 2017). Furthermore, food parenting practices targeted toward specific foods, such as vegetables (Baranowski et al. 2013), can be assessed. Parents may apply different practices depending on the food type, and certain practices might be more effective for promoting intake of some foods than others. However, it remains unclear whether examining food-specific practices provides added value compared to broader instruments that categorise foods into general groups such as ‘healthy’ and ‘unhealthy’ (Vaughn et al. 2016). There is a relative paucity of studies examining practices directed solely at vegetables and their association with children's vegetable consumption.

Parenting style refers to a more global measure of the nature of parent behaviour across different situations reflecting the parent's attitudes and beliefs towards the child (Darling and Steinberg 1993). It can be seen as the larger emotional context within which (food) parenting practices occur and in which the parent's behaviours are interpreted by the child (Darling and Steinberg 1993). Parenting style is commonly assessed based on two dimensions (warmth/responsiveness and control/demandingness), the combination of which results in four styles: authoritative (high warmth and high control), authoritarian (low warmth and high control), permissive (high warmth and low control), and disengaged (low warmth and low control) (Baumrind 1971; Power 2013). While there are some inconsistencies, the authoritative parenting style has been most often associated with healthier dietary intake, including higher fruit and vegetable consumption, among preschool-aged children (Blissett 2011; Burnett et al. 2020).

Recently, Sleddens et al. have developed a more detailed Comprehensive General Parenting Questionnaire (CGPQ) (Sleddens et al. 2014). It assesses five higher-order constructs of parenting (nurturance, structure, behavioural control, inappropriate control, and overprotection), and it was designed to capture different aspects of parenting—particularly forms of control—more extensively than previous instruments assessing parenting style (Sleddens et al. 2014). In short, nurturance refers to how responsive and supportive a parent is to their child's needs and autonomy. Structure describes the extent to which parents organise their child's environment in a consistent way to facilitate and foster child development, including setting boundaries. Behavioural control describes a form of control that is characterised by supervising and managing the child's activities in a neither over-controlling nor intrusive manner. Inappropriate control, in turn, is a form of control characterised by domination, intrusion, pressure and discouragement of child's autonomy. Overprotection, a third form of control, refers to excessive involvement or excessive nurturing that impedes the development of the child's autonomy. To our knowledge, only one study thus far has examined direct associations between parenting style assessed with the CGPQ and children's vegetable consumption. Among 288 children (6–12-year-olds) and their parents in Belgium, overprotection was correlated negatively with the child's vegetable consumption, although this association did not remain significant after adjusting for the child's age, child's gender, parental educational level, father's BMI, mother's BMI, child birth order and the other CGPQ constructs (Philips et al. 2014). Altogether, the associations between parenting style and children's vegetable consumption remain unclear, specifically regarding different forms of parental control.

In addition to potentially having direct relationships with child food consumption, theoretical models suggest that parenting style may act as a moderator in associations between (food) parenting practices and the child's behaviour (e.g., food consumption) (Darling and Steinberg 1993; Niermann et al. 2018). Parenting style creates an emotional environment that partly determines how the child perceives the parenting practices (Darling and Steinberg 1993). Studies in both younger and older children suggest that these interactions may play a role in influencing children's food consumption (Blissett 2011; Yee et al. 2017). Overall, the findings have been rather mixed, and only a few studies have specifically examined the CGPQ constructs as moderators. For example, one Dutch study found that the CGPQ constructs moderated longitudinal associations between food parenting practices when the child was 6 years old and the child's snacking and consumption of sugar-sweetened drinks and fruits at age 8 (Sleddens et al. 2014).

The objective of this study was two-fold: (1) to examine direct associations of the CGPQ constructs and three vegetable-related parenting practice (VPP) constructs (‘encouragement and modelling’, ‘child involvement’, ‘enhanced availability and autonomy support’) with children's vegetable consumption, and (2) to explore moderation by the CGPQ constructs in associations between the VPPs and children's vegetable consumption. Concerning the direct associations, we hypothesised that all VPP constructs would have a positive association with vegetable consumption, and, out of the CGPQ constructs, nurturance, structure, and behavioural control would have a positive association whereas overprotection and inappropriate control would have a negative association with vegetable consumption. To our knowledge, this study is the first to examine the CGPQ constructs as potential moderators in the associations between the VPPs and vegetable consumption among young children.

2 Methods

3 Results

3.3 Parenting Style as a Moderator

The interaction terms for nurturance and overprotection with the VPP construct of child involvement were statistically significant in both the crude and the adjusted models (p < 0.10). Results for each tested interaction term between parenting style constructs and VPP constructs in association to children's vegetable consumption can be found in Supporting Information S2: Table S2. The significant interactions were probed using the adjusted models. As can be seen in Figure 1, at lower levels of nurturance, the positive association between child involvement and the child's vegetable consumption was stronger. The association was significant when nurturance was set at 3.71 (16th percentile) and at 4.14 (50th percentile), but the association was not significant when nurturance was set at 4.57 (84th percentile) (Table 3). Overprotection had a similar moderation effect to nurturance. Figure 2 visualises overprotection as a moderator; higher child involvement was more strongly associated with higher child vegetable consumption at lower levels of overprotection. The association was significant only when overprotection was set at 2.00 (16th percentile) but not when overprotection was set at 2.60 (50th percentile) or 3.40 (84th percentile) (Table 3).

TABLE 3. Conditional associations between the vegetable-related parenting practice construct of child involvement and the child's vegetable consumptiona at different values of nurturance and overprotection.

Moderator	Valueb	B	95% CI	p-value
Nurturance	3.71	0.180	0.038; 0.322	0.013
	4.14	0.103	0.005; 0.202	0.041
	4.57	0.027	−0.096; 0.149	0.670
Overprotection	2.00	0.160	0.024; 0.296	0.022
	2.60	0.090	−0.006; 0.186	0.066
	3.40	−0.003	−0.137; 0.131	0.965

• Note: Nurturance ranged from 1.7 to 5, and overprotection ranged from 1 to 4.6.
• ^a Child vegetable consumption (g/MJ) is a square-root-transformed variable.
• ^b Value of the variable at the 16th, 50th and 84th percentiles.

4 Discussion

In our sample of 3–6-year-olds, parents' greater use of practices pertaining to the enhanced availability and autonomy support construct was associated with the child consuming more vegetables. Moreover, a higher score on the parenting style construct of overprotection was associated with the child consuming less vegetables. Additionally, two exploratory moderation effects of parenting style were found. However, it is also worth noting that several of the associations studied were not statistically significant.

Although the variety of instruments used to assess food parenting practices and parenting style makes cross-study comparisons challenging, our results seem partly in line with previous research and our hypotheses. Regarding the VPP constructs examined, the positive association between enhanced availability and autonomy support and vegetable consumption in this study aligns with the broader and rather consistent evidence on higher availability and accessibility of healthier foods at home being associated with children's healthier food consumption (Blissett 2011; Yee et al. 2017). For example, among 395 Norwegian 3–5-year-olds, higher availability and accessibility of vegetables at home were associated with higher frequency and amount of vegetables consumed by the child (Kristiansen et al. 2017). That study had similar accessibility items to our enhanced availability and autonomy support items. Moreover, parents making healthy foods available and unhealthy foods less accessible at home were associated with higher vegetable consumption among Norwegian 1–6-year-olds (Mazza et al. 2022). The potential positive impact of availability and accessibility have been attributed to the fact that children will mostly eat what is available to them (Yee et al. 2017) and that repeated exposure (including only sight) can increase the familiarity of and preference for food items (Holley et al. 2017; Jago et al. 2007). The items of the enhanced availability and autonomy support construct in the current study also included an aspect of allowing the child some autonomy in situations where vegetables are served to them. Research has shown it is likely beneficial for children to have some control, age appropriately, over their food consumption, especially regarding how much they eat (Lohse and Mitchell 2021; Vaughn et al. 2016). Our result seems to support this. However, the potential benefits of allowing children control over their food consumption may depend on the type and extent of control. For example, higher ‘child control’ assessed with the Comprehensive Feeding Practices Questionnaire has been associated with lower vegetable consumption in some studies (e.g., Quah et al. 2018).

Somewhat unexpectedly, the constructs of child involvement and encouragement and modelling, were not associated with children's vegetable consumption. Parents modelling healthy eating has been associated with healthier food consumption among children, including higher vegetable consumption, in several previous studies (Blissett 2011; Vaughn et al. 2016; Yee et al. 2017). Involving the child in planning and preparing meals and encouraging healthier consumption have also previously been associated with higher vegetable consumption in some studies involving young children (Blissett 2011; de Jong et al. 2015; Kristiansen et al. 2017; Vaughn et al. 2016). However, null results also exist (Mazza et al. 2022; Shim et al. 2016). In the current study, the distribution of the encouragement and modelling variable was quite narrow, skewed towards high scores, which may have reduced the ability to detect an association.

In terms of parenting style, the overprotection construct was negatively associated with child vegetable consumption. This is in line with a negative correlation between overprotection and children's vegetable consumption found among 6–12-year-olds in Belgium (Philips et al. 2014). However, the association did not remain after adjusting for confounders in that study. In the short version of the CGPQ we used in the current study, overprotection consists of items pertaining to the subdimension of excessive involvement (example items: “I always help my child with everything he/she does”, and “I carefully plan my child's day so that he/she has enough activities to keep him/her busy”). Thus, an overprotective parent is highly involved with and supervises over the child, which discourages independent behaviour to the extent that the child's autonomy development can be impaired (Sleddens et al. 2014). As the authors of the Belgian study (Philips et al. 2014) also speculated, overprotective parenting might create a pressuring atmosphere at mealtimes, which could consequently lead to lower vegetable consumption. On the other hand, we hypothesise that overprotective parents may be more likely to cater to the child's food preferences, which often do not include many vegetables, to avoid upsetting the child. Furthermore, parenting style may influence children's food consumption through differential use of food parenting practices (van der Horst and Sleddens 2017).

Nurturance, structure, behavioural control and inappropriate control had no associations with vegetable consumption. Nurturance, structure and behavioural control are comparable to aspects of an authoritative parenting style (a combination of high levels of warmth/responsiveness with high levels of control/demandingness) (van der Horst and Sleddens 2017), which has previously been associated with higher vegetable consumption in many studies among young children (Burnett et al. 2020). Thus, associations with vegetable consumption could have been expected for nurturance, structure and behavioural control, even though different instruments are used to assess the authoritative parenting style. However, as parenting style is considered a distal predictor of children's food consumption (Niermann et al. 2018), associations between parenting style and children's food consumption are expected to be very small, if they exist. Moreover, nurturance, structure, and behavioural control showed relatively low variability, as most parents scored high on these constructs.

Regarding the moderation by parenting style, our result suggests that the potential positive impact of involving the child in choosing vegetables and preparing them may be dependent on the parenting style context. We found that greater use of the child involvement practices was associated with higher children's vegetable consumption only when parents scored lower on nurturance or overprotection. We are not aware of previous studies that have examined the interaction between parenting style and VPPs using the same instruments to assess them as we did. Hence, direct comparisons are not possible. In one previous study using the CGPQ, encouragement and covert control practices were positively associated with children's fruit consumption and negatively with children's snacking when overprotection was lower (Sleddens et al. 2014). Another previous study found that when nurturance was lower, higher availability of healthy foods and lower availability of unhealthy foods at home was associated with the child consuming fewer pastries and sweets (Gubbels et al. 2020). Many studies have not found moderation by parenting style in associations between food parenting practices and children's vegetable consumption (Gubbels et al. 2020; Langer et al. 2017; Zhang et al. 2019). Hence, at this point, the possible mechanisms behind the interactions found in the current study are unknown. It could be speculated that when a parent is highly overprotective, they may be controlling in situations where the involving practices are used (perhaps overly controlling how the child completes the food preparation tasks, for instance), which consequently could impede the potentially positive impacts. The moderation effect of nurturance might indicate that the child involvement practices are more important for children who have parents with less optimal levels of nurturance, although, in the current study, the level of nurturance where the association was no longer significant was still very high. These findings regarding moderation by parenting style should be interpreted cautiously, given that a more lenient alpha threshold was used for the interaction terms (p < 0.10).

Overall, the literature on parenting and children's food consumption is complicated by a diverse set of methodologies in assessing parenting as well as food consumption. This makes it difficult to compare and pinpoint the sources of differing study results. Especially for assessing parenting, developing consensus on standardised methodologies would be beneficial for future research. Nevertheless, this study adds to the important evidence base for designing interventions to promote vegetable consumption among children as well as for healthcare professionals who work with children and their parents to give advice on parenting. It is not entirely clear how malleable parenting style is (Morales et al. 2024), and as parenting style is a more distal factor associated with children's food consumption, it may not be as effective target for interventions as food parenting practices. Still, it may be an important aspect to consider when designing interventions that target food parenting practices to increase children's vegetable consumption, especially given its potential moderating effects. Tailoring interventions based on the parenting style of a parent might be an effective strategy.

However, future research is needed to clarify the findings of this study, particularly regarding parenting style, in terms of causality, mechanisms, and clinical relevance. Furthermore, since a single dimension/construct does not describe the overall parenting style of a parent, and food parenting practices are not used in isolation (Sleddens et al. 2014; Vaughn et al. 2016), future research could examine profiles of parenting style derived from the CGPQ and profiles of VPP use (e.g., cluster analyses) and how they are associated with children's vegetable consumption. On a broader level, parenting styles and food parenting practices considered ‘positive’ have been associated with overall healthier dietary patterns in children—for example, lower consumption of sugary foods and beverages, alongside higher vegetable consumption (Burnett et al. 2020; Ventura and Birch 2008; Yee et al. 2017). However, there are important nuances. For example, a specific practice may be associated with the consumption of healthier foods but show no association with the consumption of unhealthier foods (e.g., Gubbels et al. 2020). Hence, future research could explore how the CGPQ constructs and the VPPs relate to other key food groups—particularly to assess whether VPPs may influence broader dietary intake despite being targeted specifically at vegetables.

This study has several strengths, including the use of the CGPQ, which comprehensively captures different aspects of parental control over the child. However, it has been rarely used thus far. Another strength of this study is examining vegetable consumption separately from fruit consumption. Generally, preference for vegetables tends to be lower than preference for fruits among children (especially younger children) (Cooke and Wardle 2005), and it seems that vegetable consumption may have at least partly different determinants than fruit consumption (Mazza et al. 2022). In addition, the actual amount of vegetables consumed was examined, not just frequency, and a 3-day food record was used, which minimises recall bias (J.-S. Shim et al. 2014). However, future studies could examine not only the quantity but also the variety of vegetable consumption as an outcome, as both dimensions are important for health (Rosell and Fadnes 2024). Among the strengths of this study is also the age group, which is considered an important age for building healthy eating habits (Hodder et al. 2024).

This study also has limitations to consider. Our cross-sectional data can only demonstrate associations, not causal relationships. It is likely that the associations between parenting and children's food consumption are bidirectional (Vaughn et al. 2016). It should also be noted that no multiple testing adjustments were adopted. Not accounting for vegetables contained in composite dishes is another limitation. However, this is unlikely to have substantially influenced the ranking of individuals (Lentjes et al. 2014). Additionally, the VPPs examined primarily target discrete vegetable consumption rather than the vegetables incorporated as ingredients in composite dishes. Thus, the potential associations can be expected to be detectable in the chosen variables. Nevertheless, future studies could strengthen these findings by also including vegetables consumed as part of composite dishes. One important consideration is that vegetable consumption in the ECEC centres was included. It is not directly controlled by parents and is influenced by early educators' practices, for instance (Lehto et al. 2019). This may have diluted the associations observed in this study. Although adapted from a validated measure, the VPP constructs used in this study lack formal validation. Moreover, Cronbach's alpha was moderately low for some of the parenting style and VPP constructs, although Cronbach's alpha over 0.5 can be considered acceptable (Portney and Watkins 2000). Assessing parenting behaviours by direct observation could provide valuable complementary insights in future research. The sample of this study might be biased to some extent, as indicated by the relatively low participation rate, the overrepresentation of higher educated parents compared to the general population of the same age in Finland (Statistics Finland 2023), and the higher likelihood of missing data on parenting style variables and children's vegetable consumption among less educated participants. As higher socioeconomic status has been associated with better dietary quality (Jarman et al. 2022) and more favourable parenting (Ayoub and Bachir 2023), it is also plausible that, relative to non-participants, participants may have had higher vegetable consumption and more favourable parenting behaviours. Additionally, since the parenting variables and children's vegetable consumption (outside preschool hours) were parent-reported, the data may be subject to social desirability bias. These factors limit the generalisability of the findings and may have attenuated the associations. However, we do not believe they affected the direction of the associations. A further limitation of the sample is that the parents were predominantly mothers. In addition, since parenting is a culturally bound construct, these associations need to be investigated in culturally/ethnically diverse samples.

5 Conclusions

Our finding that children consumed more vegetables when parents made vegetables more available at meals, while allowing the child to choose among them and portion them, supports previous evidence on the importance of availability and accessibility at home for children's vegetable consumption. Additionally, a relatively novel finding was that if parents were more overprotective, their children consumed less vegetables. Furthermore, our results provide partial support for theoretical models suggesting that parenting style may moderate the association between food parenting practices and children's food consumption. Specifically, two exploratory moderation effects were observed: higher use of practices that involve the child in choosing and preparing vegetables was associated with greater vegetable consumption only when parents scored lower on nurturance or overprotection. More research to elucidate the moderation effects of parenting style is necessary, as findings published so far have been mixed. All things considered, in addition to food parenting practices, parenting style may be an important factor to consider in efforts to promote young children's vegetable consumption, especially as a contextual factor given its potential moderation effects.

Author Contributions

Eva Roos is the principal investigator for the DAGIS project. Eva Roos and Maijaliisa Erkkola have been responsible for the funding of the project and collected data. Carola Ray, Reetta Lehto, and Henna Vepsäläinen were involved in collecting the data and/or conducting data management. Jenna Rahkola and Carola Ray conceptualised the present study. Jenna Rahkola performed the statistical analyses and drafted the manuscript. Eva Roos, Carola Ray, Maijaliisa Erkkola, Reetta Lehto, Alissa J. Burnett, Estervan der Borgh-Sleddens, Jessica Gubbels, Henna Vepsäläinen, Anna M. Abdollahi, Josefine Björkqvist, Emmi Tilli, and Nithya Serasinghe provided feedback and edited the manuscript. All authors read and approved the final manuscript.

Conflicts of Interest

The authors declare no conflicts of interest.