Development and validation of an illustrated questionnaire on healthy eating practices among Korean preschool children

Abstract

BACKGROUND/OBJECTIVES

An illustrated questionnaire, the Healthy Eating Practice Questionnaire (HEPQ), was developed and validated to assess healthy eating practices among Korean preschool children.

SUBJECTS/METHODS

The development process consisted of 4 phases: (1) item generation based on a literature review and focus group interviews, (2) item refinement through expert consultation, (3) illustration design, and (4) validation. One hundred and eighty-seven child–parent pairs and 24 teachers participated in the validation study. The reliability was examined using test–retest analysis, and the validity was assessed by comparing the children’s self-reports with those of their parents and teachers. The level of agreement was determined using the percentage agreement, kappa statistics, and intraclass correlation coefficient (ICC).

RESULTS

The final 30-item questionnaire consisted of 4 domains: food preferences, eating behaviors, hand/oral hygiene practices, and nutrition knowledge. Most items showed a percentage agreement above 60% and kappa coefficients exceeding 0.5 in test–retest analysis, indicating acceptable consistency. The child–caregiver agreement was moderate to substantial, with kappa values generally above 0.4. The domain-level ICCs indicated good reliability (ICC = 0.49–0.83) and validity (ICC = 0.33–0.77) across all domains and subgroups. The comparisons by age showed that 4-yr-olds provided valid and reliable responses comparable to 5-yr-olds. The parents and teachers confirmed that the illustrated format was easy for the children to understand and complete independently.

CONCLUSION

The HEPQ showed satisfactory reliability and validity for assessing healthy eating practices among preschool children. The illustrated format enhances comprehension and engagement, supporting its applicability for research and nutrition education in daycare settings.

INTRODUCTION

Adequate nutrition and healthy eating practices in early childhood are crucial for supporting optimal growth and development [1,2]. During this period, food preferences are established and eating habits begin to form, influencing the nutritional status and health outcomes across the lifespan [2,3]. Therefore, providing appropriate guidance for healthy eating practices during this developmental stage is essential.

Traditionally, children’s eating habits were shaped primarily within the family environment [4]. Nevertheless, the increasing enrollment of children in childcare facilities that provide meals has led to the significant influence of teachers and peers on children’s eating behaviors [5,6]. These societal changes have underscored the importance of implementing comprehensive nutrition education within childcare environments.

In South Korea, the Center for Children’s Foodservice Management (CCFSM) was established in 2011 to ensure hygiene and safety management in children’s foodservice facilities and support nutrition management through menu planning and nutrition education for preschool children in daycare centers (including kindergartens) [7]. The CCFSM had expanded to 236 centers nationwide by 2024. Nutrition education provided by CCFSM dietitians focuses on 3 key dietary tasks promoted by the Korean Ministry of Food and Drug Safety (MFDS): eating balanced meals, reducing food waste, and practicing proper handwashing [8]. Although improvements in hygiene management in children’s foodservice facilities have been confirmed through checklist-based evaluations [9], the evidence for the effectiveness of CCFSM nutrition education on children’s eating practices is limited, highlighting the need for appropriate nutritional evaluation tools.

Most previous studies relied on parent-reported questionnaires to assess young children’s eating behaviors [10]. Widely used examples include the Children’s Eating Behaviour Questionnaire (CEBQ) in the United Kingdom [11], the Nutritional Screening Tool for Every Preschooler (NutriSTEP) in Canada [12], and the Nutrition Quotient for Preschoolers in South Korea [13]. The CEBQ and NutriSTEP have been validated for reliability and validity across various countries and are considered useful [14,15,16,17,18,19], but parent-reported instruments have limitations. Specifically, they may not accurately capture children’s behavior in childcare settings where parents are not present [20]. Children’s eating behaviors in these environments may differ substantially because of peer influence, exposure to diverse foods, and varying mealtime structures [1]. These considerations emphasize the importance of developing age-appropriate tools that enable children to self-report their eating behaviors.

Considering the limited language and cognitive abilities of young children, visual approaches, such as picture- or photo-based assessments, are effective research methods for this age group [21]. Food photographs, in particular, have been used to evaluate food preferences and nutrition knowledge among preschool children [22,23,24,25,26]. Furthermore, recent studies have revealed the feasibility of illustrated dietary questionnaires for school-aged children, suggesting that they can reliably report their own eating behaviors [27,28]. Nevertheless, to the best of the authors’ knowledge, no validated illustrated instruments currently exist to assess the food preferences, eating behaviors, or nutrition knowledge of Korean preschool children.

Therefore, an illustrated questionnaire, the Healthy Eating Practice Questionnaire (HEPQ), was developed and validated. The HEPQ was designed to assess the eating behaviors among Korean preschool children and evaluate the effectiveness of CCFSM programs in promoting healthy eating practices. The specific objectives of this study were as follows: (1) describe the development process of the questionnaire, (2) examine its reliability through test–retest analysis, (3) assess its validity by comparing children’s responses with those of their caregivers, and (4) determine if the reliability and validity differed according to age and sex.

SUBJECTS AND METHODS

Procedures for questionnaire development

This study followed a systematic process for developing and validating a questionnaire [29,30]. The process consisted of 4 phases: (1) item generation based on a literature review and focus group interviews, (2) item refinement through expert consultation, (3) illustration design, and (4) validation, as shown in Fig. 1.

Fig. 1. Development process of the illustrated Healthy Eating Practice Questionnaire.

Phase 1: Item generation

Relevant literature and questionnaires from previous studies were collected and reviewed to identify dietary issues among preschool children. The main keywords included “preschool children,” “eating habits,” “eating behaviors,” “picky eating,” “food preferences,” “nutrition knowledge,” “nutrition education,” “handwashing,” and “toothbrushing.”

Focus group interviews were conducted with the mothers of children attending daycare centers registered with the CCFSM in 3 regions: a metropolitan area (Seoul), a large city (Incheon), and a rural county (Eumseong). A single focus group interview was held in each region with 4 to 5 mothers between August and October 2020. The interview in Eumseong was conducted in person, while those in Seoul and Incheon were held online via video conference because of the coronavirus disease 2019 pandemic.

Following the literature review, existing questionnaire analysis, and focus group interviews, 49 items across 4 domains—food preferences, eating behaviors, hand/oral hygiene practices, and nutrition knowledge—were developed. Simple iPad illustrations were incorporated to present the initial visual concepts in the preliminary questionnaire.

Phase 2: Item refinement

A mail survey was conducted in November 2020 with 45 experts to obtain feedback on the preliminary questions: 8 CCFSM directors, 28 CCFSM nutrition team leaders, 3 professors specializing in early childhood education, and 6 daycare teachers. The experts identified the appropriate age group for each item, rated its suitability on a 5-point Likert scale, and provided written comments. The Likert-scale ratings were converted to numerical scores for analysis. Based on expert feedback, the items were revised or removed, resulting in a 30-item questionnaire within the same 4 domains.

Phase 3: Illustration design

A professional illustrator created child-friendly images for each question and response option to ensure children’s comprehension and engagement. Facial expressions were used for the response options. The questionnaire was designed as a booklet, with 2 questions per A4-sized page. The front page included fields for the child’s name, sex, and age to facilitate matching with the child’s retest data and caregivers’ reports.

An instruction manual was developed for parents and teachers to ensure standardized administration of the questionnaire. The manual described the questionnaire’s purpose, target respondents, its structure, administration procedures, and guidance for assisting children’s responses. The instruction manual was also produced as a booklet, incorporating illustrations to guide administration and provide explanations.

Phase 4: Validation

Study design

Children completed the questionnaire twice, with a 2-week interval, to assess the test–retest reliability. The validity was evaluated by comparing the children’s self-reports with those of their parents and teachers, who have typically served as proxy respondents to assess the children’s eating behaviors [10,11,12,13,27,28,31]. Specifically, the children’s reports on food preferences and hygiene practices were compared with the parents’ reports, and the children's reports on eating behaviors at daycare were compared with the teachers’ reports.

Participants

Children aged 4 to 5 yrs attending daycare centers, along with their parents and teachers, were recruited in collaboration with the CCFSMs. Five CCFSMs were selected in proportion to the number of preschool children, based on city size: one metropolitan area (Seoul), one large city (Daegu), 2 medium-sized cities (Pocheon, Cheongju), and one rural county (Jincheon).

CCFSM staff received online video training on the survey procedures before participant recruitment. Each CCFSM recruited 4 daycare centers, with approximately 10 child–parent pairs and at least one teacher per daycare, resulting in 200 children, their parents, and 24 teachers. Trained CCFSM staff visited the participating daycare centers to explain the study and distribute survey materials. The teachers then informed the parents of the purpose and procedures of the study.

Questionnaires for parents and teachers

Parallel questionnaires were developed for parents and teachers to compare with children’s reports. The parent version contained identical items on the food preferences and hygiene practices, while the teacher version included items on the eating behaviors at daycare. For hygiene practices, the children’s questionnaire used 2 response options (yes or no), whereas the parent version used 4 options (always, sometimes, rarely, or never) to obtain more detailed data.

Data collection

The initial survey with children, parents, and teachers began in July 2021, and the retest for children was conducted approximately 2 weeks later. Data collection continued until October 2021 because of variations in the initial survey schedules across daycare centers.

The children’s questionnaire was administered either at home under parental supervision or at daycare under teacher supervision. The manual was provided to ensure standardized administration, and the children were encouraged to respond independently. The parents and teachers also completed their respective questionnaires based on the observations at home and daycare, and provided feedback on the children’s experience with the questionnaire. After excluding 13 incomplete cases, the final analysis included 187 child–parent pairs and data on 187 children reported by 24 teachers.

Ethical consideration

The Public Institutional Bioethics Committee approved the focus group interview and expert consultation studies (P01-202006-22-008), and written informed consent was obtained from all participants. The Institutional Review Board of Chungbuk National University approved the validation study (CBNU-202105-HR-0049). Written informed consent was obtained from all participating parents and teachers, and verbal assent was obtained from the children with parental confirmation.

Statistical analysis

Statistical analyses were conducted using the IBM SPSS Statistics 29 software (IBM Co., Armonk, NY, USA). The participants’ characteristics were described using frequencies and percentages. A χ² test was used to investigate the differences in sex and residential area by age. The Likert-scale items were converted to numerical scores, and the results are presented as the means and SDs. For the hygiene practice items, children had 2 response options, whereas the parents had 4. Therefore, “always” and “sometimes” in the parent questionnaire were coded as “yes,” and “rarely” and “never” were coded as “no.”

Agreement between the test–retest and child–caregiver reports was examined using percentage agreement, kappa statistics, and intraclass correlation coefficient (ICC) [32,33,34]. The percentage agreement indicates the proportion of cases classified into the same category. Although it is used widely and is easy to interpret, it does not correct for chance agreement [32,35]. The weighted kappa with quadratic weighting was applied for ordinal data on food preferences and eating behaviors [35]. For dichotomous data on hygiene practices and nutrition knowledge, the prevalence-adjusted and bias-adjusted kappa (PABAK) was used because of unbalanced response distributions. Unlike Cohen’s kappa, which is sensitive to prevalence extremes, PABAK is based solely on the observed agreement [32,34]. The kappa values were interpreted as follows: < 0 = poor; 0–0.20 = slight; 0.21–0.40 = fair; 0.41–0.60 = moderate; 0.61–0.80 = substantial; 0.81–1.00 = almost perfect agreement [35].

The ICCs and their 95% confidence intervals were calculated using a 2-way mixed-effects model with absolute agreement [34]. ICC, which is used for continuous variables, is mathematically equivalent to weighted kappa under certain conditions [36]. In this study, the weighted kappa and ICC values for food preferences and eating behaviors were similar. Therefore, the ICCs were presented solely to assess the agreement in the average scores across domains between test–retest and child–caregiver reports. ICC values < 0.5 are considered as poor, 0.5 to < 0.75 as moderate, 0.75 to < 0.9 as good, and values ≥ 0.90 as excellent agreement [34]. Paired t-tests were used to determine the differences in test–retest and child–caregiver scores within each domain. Differences in the average score of each domain by age and sex were analyzed using independent t-tests. A P-value of < 0.05 was considered significant.

RESULTS

Development of a preliminary questionnaire for expert consultation

The literature review showed that picky eating, particularly avoidance of vegetables, was the most prevalent issue among Korean preschool children. Nutrition education programs focused mainly on encouraging balanced eating and promoting vegetable consumption, as shown in Table 1. Their effectiveness was typically evaluated through nutrition knowledge, food preferences, and eating behaviors [37,38,39,40,41,42,43,44,45,46]. Nutrition knowledge and food preferences were mostly self-reported by the children, while parents usually reported the children’s eating behaviors. For the child respondents, data were collected through individual or small-group interviews, and food photographs or illustrations were used in the questionnaire, but none of the questionnaires had been validated.

Table 1. Evaluation domains and respondents in studies (2010–2019) on the effectiveness of nutrition education among Korean preschool children.

Author (yrs)	Topic of education	Age (yrs)	Respondents in each evaluation domain				Use of visual cues
			Nutrition knowledge	Food preferences	Eating behaviors	Handwashing/toothbrushing
Kim et al. (2010) [37]	Food groups	4	C	-	-	-	Food photographs (partially)
Hong et al. (2010) [38]	Healthy eating/table manners	5–6	C	-	P	-	Illustrations
Hong et al. (2010) [39]	Healthy eating/handwashing/toothbrushing	5–6	C	-	P	P	Illustrations
Oh et al. (2012) [40]	Vegetables/fruits/dairy products	6	C	C	C	-	Illustrations (partially)
Lee and Lee (2014) [41]	Vegetables	5–6	C	-	-	-	×
Jo et al. (2015) [42]	Handwashing/vegetables	3–5	-	-	P, T	P, T	×
Shin et al. (2015) [43]	Handwashing/toothbrushing/table manners/healthy eating	4	C	-	C, P	-	Illustrations
Yeom and Cho (2016) [44]	Sugar reduction	5	C	C	P	-	Food photographs, illustrations (partially)
Moon and Her (2017) [45]	Vegetables	5	C	C	-	-	Illustrations
Park and Kim (2018) [46]	Vegetables	3–5	-	-	P	-	×

C, children; P, parents; T, teachers; -, no questions.

Table 2 summarizes the findings of focus group interviews. The parents expected their children to eat balanced meals, consume adequate amounts, and enjoy mealtimes. The main nutritional concerns were picky eating and poor appetite. Commonly avoided foods included vegetables, hard-to-chew meats, and plain milk. Parents reported that children’s eating behaviors were influenced by the family, peers, and media exposure, and recognized the need for nutrition education, although its effects lasted only a few days. These findings highlight the importance of providing a variety of foods and encouraging children to consume them by offering frequent, engaging nutrition education at daycare to promote healthy food preferences and eating behaviors.

Table 2. Summary of focus group interview findings on children’s eating behaviors.

Questions	Responses
Desired eating behaviors	• Eating balanced meals, avoiding picky eating
	• Consuming adequate amounts
	• Enjoying mealtimes
Nutritional concerns	• Picky eating (avoiding vegetables, meats, white milk)
	• Poor appetite or low interest in foods (leading to underweight)
	• Frequent consumption of convenience/instant foods
	• Frequent consumption of sugar-containing snacks
Food preferences	• Avoided foods: vegetables, meats, white milk
	• Preferred foods: fruits, dumplings, noodles, instant foods, sweet snacks
Factors influencing eating behaviors	• Family environment (siblings, parents)
	• Peers’ eating behaviors
	• Media exposure (TV, smartphone)
Consumption of instant/processed foods	• Frequency: mostly 2–3 times per week
	• Commonly consumed foods: ramyeon, ham, chicken nugget, dumplings, fried rice
	• Reasons for providing: convenience of preparation, menu variety, children’s preferences
Parental recognition of nutrition education	• Recognized as necessary
	• Effective only for a few days
	• Need for more frequent education
	• Desired topics: balanced eating, reducing picky eating, limiting sweet foods, promoting healthy eating habits

Based on a literature review, the focus group findings, and the key dietary tasks promoted by the Korean MFDS, a preliminary 49-item questionnaire was developed across 4 domains: food preferences, eating behaviors, hand/oral hygiene practices, and nutrition knowledge. The response formats varied according to the domain: a 5-point scale for food preferences, a 3-point scale for eating behaviors, yes/no for hygiene practices, and single-choice questions from 5 options for nutrition knowledge.

Development of an illustrated questionnaire for validation

Forty-five experts reviewed the initial 49 items, and 20 were excluded based on appropriateness scores and qualitative feedback. Most of the remaining items achieved mean appropriateness scores above 4.0 (Table 3) and were refined to enhance comprehension for preschool children. For example, questions were simplified from “How do you feel when you eat ∼?” to “Do you like ∼?” Although the food groups for the food preference and eating behavior domains were derived from the Food Balance Wheel, modifications were made to reflect daycare menus better: kimchi was added, and beans were replaced with tofu. In the eating behavior domain, questions were rephrased from asking whether children ate all their food to whether they left certain food at daycare. The items in the hygiene domain were changed from first-person statements to direct questions for consistency, and the nutrition knowledge domain adopted more child-friendly terms, for example, “golden poop” instead of “pooping well.” The experts noted that the questionnaire was appropriate for children aged 4 and 5 yrs, but some items might be challenging for 3-yr-olds. The final questionnaire consisted of 30 items across 4 domains: food preferences (8 items), eating behaviors (9 items), hand/oral hygiene practices (8 items), and nutrition knowledge (5 items).

Table 3. Refined questions based on expert consultation.

Preliminary question				Appropriateness score1)	Refined question
Food preference
		1	How do you feel when you eat meat?	4.16	Do you like meat?
		2	How do you feel when you eat fish?	4.16	Do you like fish?
		3	How do you feel when you eat eggs?	4.16	Do you like eggs?
		4	How do you feel when you eat beans?	4.16	Do you like tofu?
		5	-		Do you like kimchi?
		6	How do you feel when you eat vegetables?	4.16	Do you like vegetables?
		7	How do you feel when you eat fruit?	4.14	Do you like fruit?
		8	How do you feel when you drink milk?	4.16	Do you like white milk?
Eating behaviors
		1	Do you eat all your rice at daycare?	4.29	Do you leave your rice at daycare (kindergarten)?
		2	Do you eat all your meat dishes at daycare?	4.36	Do you leave your meat dishes at daycare (kindergarten)?
		3	Do you eat all your fish dishes at daycare?	4.31	Do you leave your fish dishes at daycare (kindergarten)?
		4	Do you eat all your egg dishes at daycare?	4.36	Do you leave your egg dishes at daycare (kindergarten)?
		5	Do you eat all your bean dishes at daycare?	4.32	Do you leave your tofu dishes at daycare (kindergarten)?
		6	-		Do you leave your kimchi at daycare (kindergarten)?
		7	Do you eat all your vegetable dishes at daycare?	4.36	Do you leave your vegetable dishes at daycare (kindergarten)?
		8	Do you eat all your fruit at daycare?	4.36	Do you leave your fruit at daycare (kindergarten)?
		9	Do you drink all your milk at daycare?	4.33	Do you leave your white milk at daycare (kindergarten)?
Hygiene practices
	Hand washing
		1	I wash my hands before eating.	4.67	Do you wash your hands before eating?
		2	I wash my hands after using the bathroom.	4.69	Do you wash your hands when you come out of the bathroom?
		3	I wash my hands after blowing my nose or coughing	4.54	Do you wash your hands after blowing your nose?
		4	I wash my hands when I come home from daycare (or kindergarten), academy, or playground.	4.64	Do you wash your hands when you come home from outside?
		5	I use soap when I wash my hands.	4.67	Do you wash your hands using soap?
	Toothbrushing
		6	I brush my teeth after eating meals.	4.68	Do you brush your teeth after eating meals?
		7	I brush my teeth after eating snacks.	4.55	Do you brush your teeth after eating sweet snacks?
		8	I brush my teeth before going to bed.	4.68	Do you brush your teeth before going to bed
Nutrition knowledge
		1	Which food helps you poop well?	3.91	Which food helps you make golden poop?
		2	Which food is best for making your bones strong?	4.27	Which food helps make your bones strong?
		3	Which food is good for your eyes?	4.00	Which food helps keep your eyes healthy?
		4	Which food helps build your muscles?	3.58	Which food helps make your muscles strong?
		5	Which food makes your teeth go bad?	4.37	Which food is bad for your teeth?

Values are presented as the means.

¹⁾Scores were calculated based on a 5-point Likert scale (1 = highly inappropriate, 2 = inappropriate, 3 = neutral, 4 = appropriate, 5 = highly appropriate).

Each of the 30 finalized items was illustrated to make the questionnaire visually engaging and easy for children to understand. These child-friendly designs allowed the children to respond independently with minimal verbal explanation. Fig. 2 presents examples of the illustrated HEPQ.

Fig. 2. Examples of the illustrated Healthy Eating Practice Questionnaire. The illustrations were produced with the support of Storymaker Co., Ltd., which holds the copyright.

Validation of the illustrated questionnaire

One hundred and eighty-seven child–parent pairs and 24 teachers participated in the validation study. The proportions of 4- and 5-yr-olds were similar (52.4% and 47.6%), and the sex distribution and residential area did not differ according to the age group (Table 4). Most parent respondents were mothers (90.9%), and all participating teachers were female.

Table 4. General characteristics of child–parent pairs for validation.

Group	Variable		Total (n = 187)	4 yrs (n = 98)	5 yrs (n = 89)	P-value1)
Children	Sex					0.439
		Boy	98 (52.4)	54 (55.1)	44 (49.4)
		Girl	89 (47.6)	44 (44.9)	45 (50.6)
	Residential area					0.572
		Seoul/large city	79 (42.3)	40 (40.8)	39 (43.8)
		Medium/small city	70 (37.4)	40 (40.8)	30 (33.7)
		Rural county	38 (20.3)	18 (18.4)	20 (22.5)
Parents	Mother		170 (90.9)	87 (88.8)	83 (93.2)	0.287
	Father		17 (9.1)	11 (11.2)	6 (6.7)

Values are presented as number (%).

¹⁾P-value was determined using the χ² test.

Table 5 presents the test–retest reliability of the individual items, revealing them to be acceptable. For food preferences, the percentage agreement ranged from 42.2% to 72.2% and the weighted kappa values ranged from 0.50 to 0.71, indicating moderate to substantial reliability. The eating behavior items showed 70.1–90.9% agreement, with kappa values of 0.46–0.69. The PABAK coefficients for hygiene practices and nutrition knowledge ranged from 0.57 to 0.98 and 0.49 to 0.96, respectively.

Table 5. Reliability of the individual questionnaire items.

Item			Score		% agreement	Kappa (95% CI)1)	Kappa value interpretation
			Children’s test (n = 187)	Children’s retest (n = 187)
Food preferences2)
	Meat		4.30 ± 0.88	4.40 ± 0.90	67.4	0.63 (0.51–0.76)	Substantial
	Fish		4.29 ± 0.92	4.30 ± 0.85	61.5	0.50 (0.35–0.65)	Moderate
	Egg		4.39 ± 0.96	4.48 ± 0.88	64.7	0.71 (0.60–0.83)	Substantial
	Tofu		3.95 ± 1.12	4.05 ± 1.09	50.3	0.55 (0.43–0.66)	Moderate
	Kimchi		3.29 ± 1.43	3.37 ± 1.37	48.7	0.66 (0.57–0.76)	Substantial
	Vegetables (except kimchi)		3.44 ± 1.17	3.37 ± 1.21	42.2	0.53 (0.40–0.66)	Moderate
	Fruits		4.51 ± 0.84	4.57 ± 0.83	72.2	0.66 (0.52–0.80)	Substantial
	White milk		4.28 ± 1.02	4.39 ± 0.89	61.5	0.55 (0.42–0.69)	Moderate
Eating behaviors3)
	Rice		2.78 ± 0.43	2.79 ± 0.44	85.0	0.60 (0.47–0.73)	Moderate
	Meat		2.77 ± 0.49	2.77 ± 0.48	80.8	0.46 (0.28–0.64)	Moderate
	Fish		2.80 ± 0.47	2.82 ± 0.45	82.4	0.50 (0.32–0.68)	Moderate
	Egg		2.86 ± 0.41	2.83 ± 0.45	88.2	0.60 (0.41–0.78)	Moderate
	Tofu		2.72 ± 0.55	2.72 ± 0.57	77.5	0.51 (0.34–0.67)	Moderate
	Kimchi		2.52 ± 0.68	2.48 ± 0.74	77.0	0.69 (0.59–0.80)	Substantial
	Vegetables (except kimchi)		2.49 ± 0.64	2.45 ± 0.70	70.1	0.63 (0.53–0.73)	Substantial
	Fruits		2.80 ± 0.47	2.84 ± 0.43	88.2	0.63 (0.47–0.80)	Substantial
	White milk		2.90 ± 0.34	2.91 ± 0.32	90.9	0.51 (0.25–0.76)	Moderate
Hygiene practices4)
	Hand washing
		Before eating meals	0.94 ± 0.25	0.93 ± 0.26	93.1	0.86 (0.79–0.94)	Almost perfect
		After using the restroom	0.91 ± 0.29	0.86 ± 0.35	86.6	0.73 (0.63–0.83)	Substantial
		After blowing your nose	0.49 ± 0.50	0.52 ± 0.50	79.2	0.58 (0.47–0.70)	Moderate
		When come home from outside	0.98 ± 0.13	0.99 ± 0.07	97.9	0.96 (0.92–1.00)	Almost perfect
		With soap	0.97 ± 0.16	0.98 ± 0.15	96.3	0.93 (0.87–0.98)	Almost perfect
	Tooth brushing
		After eating meals	0.81 ± 0.39	0.85 ± 0.36	85.6	0.71 (0.61–0.81)	Substantial
		After eating sweet snacks	0.45 ± 0.50	0.53 ± 0.50	78.6	0.57 (0.45–0.69)	Moderate
		Before going to bed	0.98 ± 0.13	0.98 ± 0.13	98.9	0.98 (0.95–1.00)	Almost perfect
Nutrition knowledge5)
	Have a healthy poop		0.95 ± 0.21	0.91 ± 0.29	91.4	0.83 (0.75–0.91)	Almost perfect
	Make your bones strong		0.79 ± 0.41	0.87 ± 0.34	83.4	0.67 (0.56–0.78)	Substantial
	Keep your eyes healthy		0.75 ± 0.43	0.74 ± 0.44	79.7	0.59 (0.48–0.71)	Moderate
	Make your muscles strong		0.52 ± 0.50	0.60 ± 0.49	74.3	0.49 (0.36–0.61)	Moderate
	Bad for your teeth		0.98 ± 0.15	0.98 ± 0.15	97.9	0.96 (0.92–1.00)	Almost perfect

Values are presented as mean ± SD.

CI, confidence interval.

¹⁾Weighted kappa for food preferences and eating behaviors, and prevalence-adjusted and bias-adjusted kappa for hygiene practices and nutrition knowledge.

²⁾Scores were calculated based on a 5-point Likert scale (1 = dislike a lot, 2 = dislike, 3 = Neither like nor dislike, 4 = like, 5 = like a lot).

³⁾Scores were calculated based on a 3-point Likert scale (1 = leave all, 2 = leave a little, 3 = do not leave any).

⁴⁾Scores were calculated based on a 2-point Likert scale (0 = No, 1 = Yes).

⁵⁾Scores were calculated based on 5 responses (0 = incorrect answers or ‘do not know,’ 1 = correct answer).

The agreement between children’s and caregivers’ reports was moderate to substantial for most items (Table 6). The weighted kappa values for food preferences ranged from 0.53 to 0.74, and the PABAK coefficients for hygiene practices ranged from 0.30 to 0.96. The child–teacher agreement on eating behaviors showed weighted kappa values ranging from 0.35 to 0.59.

Table 6. Validity of the individual questionnaire items.

Item			Score		% agreement	Kappa (95% CI)2)	Kappa value interpretation
			Children’s test (n = 187)	Caregivers’ report1) (n = 187)
Food preferences3)
	Meat		4.30 ± 0.88	4.22 ± 0.83	63.1	0.66 (0.57–0.76)	Substantial
	Fish		4.29 ± 0.92	4.18 ± 0.79	61.0	0.53 (0.41–0.65)	Moderate
	Egg		4.39 ± 0.96	4.28 ± 0.94	65.8	0.74 (0.64–0.84)	Substantial
	Tofu		3.95 ± 1.12	3.79 ± 0.97	55.1	0.60 (0.49–0.72)	Moderate
	Kimchi		3.29 ± 1.43	3.05 ± 1.22	54.0	0.74 (0.66–0.82)	Substantial
	Vegetables (except kimchi)		3.44 ± 1.17	3.16 ± 0.98	48.1	0.58 (0.48–0.69)	Moderate
	Fruits		4.51 ± 0.84	4.26 ± 0.89	62.0	0.64 (0.51–0.76)	Substantial
	White milk		4.28 ± 1.02	3.92 ± 1.09	62.6	0.63 (0.51–0.75)	Substantial
Eating behaviors4)
	Rice		2.78 ± 0.43	2.67 ± 0.50	81.3	0.58 (0.45–0.71)	Moderate
	Meat		2.77 ± 0.49	2.73 ± 0.50	73.8	0.37 (0.23–0.52)	Fair
	Fish		2.80 ± 0.47	2.84 ± 0.41	82.4	0.42 (0.22–0.61)	Moderate
	Egg		2.86 ± 0.41	2.84 ± 0.40	85.0	0.49 (0.29–0.68)	Moderate
	Tofu		2.72 ± 0.55	2.68 ± 0.52	69.5	0.35 (0.19–0.52)	Fair
	Kimchi		2.52 ± 0.68	2.37 ± 0.70	66.3	0.59 (0.48–0.70)	Moderate
	Vegetables (except kimchi)		2.49 ± 0.64	2.26 ± 0.70	62.6	0.49 (0.37–0.61)	Moderate
	Fruits		2.80 ± 0.47	2.72 ± 0.55	82.4	0.58 (0.42–0.74)	Moderate
	White milk		2.90 ± 0.34	2.89 ± 0.39	89.3	0.41 (0.15–0.67)	Moderate
Hygiene practices5)
	Hand washing
		Before eating meals	0.94 ± 0.25	0.90 ± 0.30	89.8	0.80 (0.71–0.88)	Substantial
		After using the restroom	0.91 ± 0.29	0.93 ± 0.26	85.0	0.70 (0.60–0.80)	Substantial
		After blowing your nose	0.49 ± 0.50	0.55 ± 0.50	65.2	0.30 (0.17–0.44)	Fair
		When come home from outside	0.98 ± 0.13	0.98 ± 0.13	97.9	0.96 (0.92–1.00)	Almost perfect
		With soap	0.97 ± 0.16	0.99 ± 0.10	96.3	0.93 (0.87–0.98)	Almost perfect
	Tooth brushing
		After eating meals	0.81 ± 0.39	0.83 ± 0.38	78.1	0.56 (0.44–0.68)	Moderate
		After eating sweet snacks	0.45 ± 0.50	0.56 ± 0.50	69.0	0.38 (0.25–0.51)	Fair
		Before going to bed	0.98 ± 0.13	0.98 ± 0.13	97.9	0.96 (0.92–1.00)	Almost perfect

Values are presented as mean ± SD.

CI, confidence interval.

¹⁾Parents (n = 187) reported on food preferences and hygiene practices, while 24 teachers reported on eating behaviors of 187 children, with each teacher evaluating no more than 10 children.

²⁾Weighted kappa for food preferences and eating behaviors, and prevalence-adjusted and bias-adjusted kappa for handwashing/toothbrushing practices.

³⁾Scores were calculated based on a 5-point Likert scale (1 = dislike a lot, 2 = dislike, 3 = Neither like nor dislike, 4 = like, 5 = like a lot).

⁴⁾Scores were calculated based on a 3-point Likert scale (1 = leave all, 2 = leave a little, 3 = do not leave any).

⁵⁾Scores were calculated based on a 2-point Likert scale (0 = No, 1 = Yes).

Table 7 lists the mean score of each domain and the ICCs for test–retest reliability and child–caregiver agreement according to the children’s age and sex. No significant differences were found between the children’s test and retest mean scores, except for food preferences and hygiene practices among 4-yr-olds (P < 0.05). Children tended to report higher scores for food preferences and eating behaviors than caregivers (all P < 0.05). In addition, 5-yr-olds showed significantly higher scores for eating behaviors at the children’s and caregivers’ reports (all P < 0.001) and lower scores for hygiene practices at children’s retest (P < 0.05) than 4-yr-olds. However, no significant sex differences were observed in the average scores of each domain.

Table 7. Mean score of domains and ICC between test and retest and between child and caregivers’ report by children’s age and sex.

Domain			Score					ICC (95% CI)
			Children’s test	Children’s retest	P-value1)	Caregivers’ report2)	P-value3)	Test–retest	Child–caregiver
Food preferences
	Total (n = 187)		4.06 ± 0.48	4.12 ± 0.49	0.020	3.86 ± 0.50	< 0.001	0.72* (0.64–0.78)	0.60* (0.41–0.73)
	Age
		4 yrs (n = 98)	4.03 ± 0.47	4.12 ± 0.49	0.018	3.82 ± 0.48	< 0.001	0.65* (0.52–0.75)	0.57* (0.34–0.71)
		5 yrs (n = 89)	4.09 ± 0.49	4.11 ± 0.48	0.475	3.90 ± 0.51	< 0.001	0.79* (0.70–0.86)	0.64* (0.43–0.77)
	Sex
		Boy (n = 98)	4.07 ± 0.53	4.12 ± 0.52	0.137	3.86 ± 0.54	< 0.001	0.77* (0.67–0.84)	0.59* (0.38–0.73)
		Girl (n = 89)	4.04 ± 0.42	4.11 ± 0.45	0.071	3.86 ± 0.45	< 0.001	0.64* (0.50–0.75)	0.63* (0.39–0.77)
Eating behaviors
	Total		2.74 ± 0.31	2.73 ± 0.30	0.866	2.67 ± 0.33	< 0.001	0.80* (0.75–0.85)	0.72* (0.63–0.79)
	Age
		4 yrs	2.64 ± 0.35§§	2.64 ± 0.35§§	0.914	2.58 ± 0.34§§	0.006	0.83* (0.76–0.88)	0.77* (0.67–0.84)
		5 yrs	2.84 ± 0.21	2.84 ± 0.20	0.894	2.77 ± 0.30	0.003	0.63* (0.48–0.74)	0.55* (0.38–0.68)
	Sex
		Boy	2.73 ± 0.32	2.73 ± 0.30	0.853	2.65 ± 0.36	< 0.001	0.83* (0.75–0.88)	0.73* (0.59–0.82)
		Girl	2.74 ± 0.30	2.74 ± 0.30	0.687	2.69 ± 0.30	0.035	0.78* (0.68–0.85)	0.72* (0.60–0.81)
Hygiene practices
	Total		0.82 ± 0.16	0.83 ± 0.16	0.119	0.84 ± 0.15	0.065	0.65* (0.56–0.73)	0.43* (0.30–0.54)
	Age
		4 yrs	0.84 ± 0.16	0.86 ± 0.15§	0.046	0.85 ± 0.15	0.331	0.76* (0.66–0.84)	0.43* (0.25–0.58)
		5 yrs	0.79 ± 0.16	0.80 ± 0.17	0.620	0.82 ± 0.15	0.100	0.53* (0.36–0.66)	0.41* (0.23–0.57)
	Sex
		Boy	0.81 ± 0.16	0.82 ± 0.16	0.631	0.83 ± 0.15	0.221	0.67* (0.54–0.76)	0.52* (0.36–0.65)
		Girl	0.82 ± 0.16	0.85 ± 0.16	0.086	0.85 ± 0.15	0.171	0.64* (0.50–0.75)	0.33* (0.14–0.50)
Nutrition knowledge
	Total		0.80 ± 0.20	0.82 ± 0.20	0.098	-4)		0.58* (0.48–0.67)	-
	Age
		4 yrs	0.79 ± 0.20	0.81 ± 0.21	0.266	-		0.62* (0.49–0.73)	-
		5 yrs	0.81 ± 0.20	0.84 ± 0.19	0.224	-		0.54* (0.37–0.67)	-
	Sex
		Boy	0.80 ± 0.21	0.82 ± 0.20	0.339	-		0.49* (0.33–0.63)	-
		Girl	0.79 ± 0.19	0.82 ± 0.20	0.132	-		0.70* (0.57–0.79)	-

Values are presented as mean ± SD.

ICC, intraclass correlation coefficient; CI, confidence interval.

¹⁾By paired t-test between children’s test and retest.

²⁾Parents (n = 187) reported on food preferences and hygiene practices, while 24 teachers reported on eating behaviors of 187 children, with each teacher evaluating no more than 10 children.

³⁾By paired t-test between children’s test and caregivers’ report.

⁴⁾Not investigated.

^*P < 0.001 by ICC agreement analysis.

^§P < 0.05, ^§§P < 0.001 by t-test between 4-yr-olds and 5-yr-olds.

Domain-level analysis showed that the ICCs for the test–retest reliability ranged from 0.49 to 0.83, indicating moderate to good reliability across all domains and all subgroups. The highest ICC was found for eating behaviors (0.80), followed by food preferences (0.72), hygiene practices (0.65), and nutrition knowledge (0.58, all P < 0.001). For the child–caregiver agreement, the ICCs ranged from 0.33 to 0.77, with the highest concordance observed for eating behaviors (0.72), followed by food preferences (0.60) and hygiene practices (0.43, all P < 0.001). There were no significant sex differences in the ICCs for the test–retest reliability and child–caregiver agreement across all domains. The ICCs for the test–retest reliability in eating behaviors and hygiene practices tended to be higher among 4-yr-olds than 5-yr-olds. Overall, 4- and 5-yr-olds showed reliable and valid responses, indicating that the questionnaire is suitable for preschool children.

Parents and teachers reported that the illustrated questionnaire was easy for their children to understand and complete independently. The use of facial expressions facilitated comprehension and made the survey engaging and age-appropriate. The questionnaire also encouraged caregivers to observe the children’s eating behaviors more closely. Nevertheless, some limitations were noted, including the difficulty assessing preferences when the children showed inconsistent responses within the same food group, and the tendency for children to be served only small portions of disliked foods at daycare, resulting in minimal food waste.

DISCUSSION

This study developed and validated an illustrated questionnaire to assess healthy eating practices among Korean preschool children. The questionnaire demonstrated acceptable test–retest reliability and validity. These results support the feasibility of using illustrated self-report tools for eating practice assessment in children aged 4 to 5 yrs.

The HEPQ consisted of 4 domains: food preferences, eating behaviors, hand/oral hygiene practices, and nutrition knowledge. The literature review and focus group results indicated that the primary goal for children was to promote eating balanced meals. Items related to food preferences and eating behaviors toward basic food groups were included in the questionnaire because food preferences and early exposure are among the strongest predictors of what young children eat [47,48]. The questionnaire emphasized eating healthy foods rather than restricting unhealthy ones because limiting young children’s consumption of sweet and palatable foods is greatly influenced by environmental factors such as the home setting, peers, and broader social context [49]. Handwashing and toothbrushing items were incorporated because these behaviors are typically performed before and after eating for hygiene purposes [50] and are key topics in the nutrition education provided by the CCFSMs [42]. A nutrition knowledge domain was also included because knowledge is linked to healthier food preferences and intake among young children [47,51].

In the expert assessment, the mean appropriateness scores for most items exceeded 4.0, except for 2 nutrition knowledge questions, confirming the overall content validity of the questionnaire. The term “muscles” was identified as difficult for young children to understand. Accordingly, carefully designed illustrations were used to convey its meaning in an age-appropriate manner. The mean knowledge score for this item was 0.52, with moderate reliability (κ = 0.49).

The overall test–retest reliability indicated that children’s responses were stable over time. For most items, the percentage agreements were above 60% and the kappa coefficients exceeded 0.5, suggesting acceptable consistency. Among the 4 domains, food preferences and eating behaviors showed higher ICCs, suggesting that preschool children can consistently report their food preferences and the degree of food left when appropriate visual cues are provided. Lower but acceptable ICC for hygiene practices was found. The lower ICC may reflect the limited response variability due to the binary yes/no formats. More than 90% of children answered “yes” to 5 hygiene items, resulting in highly skewed distributions. Using 3 response options (e.g., always, sometimes, or never) may enhance sensitivity and discriminative power in future applications.

The validity assessment showed that the child–caregiver agreement of individual items ranged from fair to almost perfect (κ = 0.30–0.96), and the agreement across domains was moderate to good, except for hygiene practices, with the highest concordance observed for eating behaviors. Caregivers tended to rate the children’s behaviors less favorably than the children themselves, particularly for food preferences and eating behaviors. Similar results have been reported [52,53]. Nevertheless, food preference items showed strong agreement, suggesting that the preferences of young children can be assessed accurately through self-report surveys.

Domain-level analysis showed that the test–retest agreement was higher than the child–caregiver agreement. Vereecken et al. [54] also reported high test–retest ICCs for fruit and vegetable preferences (0.74 and 0.75, respectively) among children aged 4 to 6 yrs, which were significantly higher than the child–parent agreement ICCs (0.48 and 0.41, respectively). ICC comparisons in food preferences and nutrition knowledge across age and sex groups showed no significant differences, consistent with the previous findings [54,55]. Interestingly, the reliability in eating behaviors and hygiene practices tended to be slightly higher among 4-yr-olds than 5-yr-olds, contrary to the assumption that older children would perform better. These results confirmed that even 4-yr-olds can provide valid and reliable responses, supporting the appropriateness of the questionnaire for both age groups targeted.

The feedback from parents and teachers further supported the usability of the illustrated format. They reported that the illustrations, including facial expressions, helped children understand the questions better and maintained their engagement. Such visual approaches have been recognized as effective in bridging communication gaps in early childhood assessments [21]. Previous studies have used food photographs and neutral emoji-like faces to measure the food preferences and nutrition knowledge among young children [24,25,26]. This study used original, child-friendly illustrations rather than photographs, and incorporated 5 facial expressions alternating between boys and girls. Children aged 4 and above can express their degree of liking on a 5-point hedonic scale using visual scales with verbal descriptors [56].

This study has several strengths. First, the HEPQ was developed systematically through a literature review, focus group interviews, and expert consultation, followed by refinement, to ensure developmental appropriateness and contextual relevance for Korean daycare settings. Second, the HEPQ enabled children to make their own choices using an illustrated format with child-friendly facial expressions. The acceptable consistency between the children’s and caregivers’ reports suggests that this self-report questionnaire is suitable for assessing eating practices and can be administered in daycare settings with minimal adult assistance. Third, multiple statistical methods were used to assess reliability and validity, including the percentage agreement, weighted kappa, PABAK, and ICC. Finally, the HEPQ incorporated 3 key dietary tasks promoted by the Korean MFDS: eating balanced meals, reducing food waste, and practicing proper handwashing. Therefore, it is applicable for evaluating the outcomes of CCFSM programs.

This study had several limitations. First, despite the efforts to recruit validation participants from multiple regions with varying population sizes, the sample may still not fully represent all Korean preschool children. Future large-scale and longitudinal studies will be needed to confirm the generalizability of the findings and to monitor the changes in children’s eating practices over time. Second, the validity of children’s responses was evaluated based on caregivers’ reports. Nevertheless, it is unclear if caregivers can accurately represent children’s actual food preferences and eating behaviors [52,53,54,57,58]. Third, the questionnaire may not capture variations in individual foods within the same food group because it assesses the preferences and eating behaviors at the food group level. Developing expanded versions that assess preferences and eating behaviors for specific food items could improve the applicability of the questionnaire across diverse populations.

In conclusion, the 30-item illustrated HEPQ showed satisfactory reliability and validity in assessing healthy eating practices, including food preferences, eating behaviors, hand/oral hygiene practices, and nutrition knowledge, among Korean preschool children. This practical and engaging tool can be applied effectively in daycare and nutrition education settings to evaluate and promote healthy eating habits from early childhood.