A
Snack Preferences among School-Aged Children: Evidence from Discrete Choice ExperimentsHanyi Li1, Weiwei Zhou1, Xiaoyu Wu1, Fangxu Guan1,2, Xiaofang Jia1,2, Qiya Guo1,2, Jiguo Zhang1,2, Shuang Liu3, Xuefeng Zhong4, Wenwen Du1,2z,2*, Huijun Wang1,2*, Aidong Liu1,2*
1National Institute of Nutrition and Health, Chinese Center for Disease Control and Prevention, Beijing,China
2 Key Laboratory of Public Nutrition and Health, National Health Commission of the People’s Republic of China, Beijing 100050, China
3 Institute of Health Surveillance Analysis and Protection, Hubei Provincial Center for Disease Control and Prevention, Wuhan 430079, Hubei, China
4 School of Health Management, Anhui Medical University, Hefei, China
z,2 Corresponding author. National Institute of Nutrition and Health, Chinese Center for Disease Control and Prevention, Beijing10050,China.
E-mail address: duww@ninh.chinacdc.cn (Du Wenwen).
Abstract
Background
A
Accelerated economic development and innovative food industry trends have contributed to worsening dietary imbalanced among Chinese school-aged population.To investigate snack preferences among school-aged children, determine the relative importance of different snack attributes, and explore how these preferences vary by social-demographic characteristics to inform targeted nutritional interventions.Methods
A stratified cluster random sampling strategy was employed to recruit 854 junior school students (grades 7 and 8) from Hubei and Anhui provinces. An evaluation framework comprising six attributes (taste, nutrient claims, purchase location, price, package size, and social influence) was constructed based on a Discrete Choice Experiment (DCE). A D-optimal design was implemented using SAS 9.4 software to generate 16 choice sets (divided into two versions), with two-stage questions incorporated to mitigate bias. A mixed logit model was applied to calculate preference coefficients (β), relative importance (RI), and willingness-to-pay (WTP) for different attributes.
Results
All six included attributes demonstrated statistically significant effects on school-aged children's snack choices (all P < 0.05). Specifically, children showed stronger preferences for snacks sold in supermarkets (β = 0.439), those rich in dietary fiber (β = 0.611), with sweet taste (β = 0.471), and commonly consumed by family members (β = 0.452). They were willing to pay an additional 10.89 CNY for snacks rich in dietary fiber. Large package size (β=-0.112) and price (β=-0.013) showed slight negative associations. Heterogeneity analysis revealed that boys preferred snacks rich in dietary fiber (β = 0.554) and were more price-sensitive (β=-0.026), while girls prioritized low-fat options (β = 0.738) and showed lower willingness to choose large packages (β=-0.259). Children with less-educated parents placed greater emphasis on taste (RI = 25.33%), whereas those from highly-educated families valued package size more (RI = 10.43%). Overweight children preferred spicy snacks (β = 0.448) and were more price-sensitive (β=-0.115), while normal-weight children tended to choose sweet-tasting snacks (β = 0.510).
Conclusion
Snack preferences among school-aged children vary significantly by socio-demographic characteristics.
Keywords:
Child
Snack
Preferences༛ Discrete Choice Experiment
A
A
A
A
A
1 Introduction
In recent years, rapid socioeconomic development and dynamic transformations in the food industry have exacerbated the issue of imbalanced dietary patterns among school-aged children in China. Excessive consumption of snacks high in sugar, salt, and fat has not only accelerated the declining age trend of health problems such as obesity[1] and dental caries[2], but also emerged as a potential risk factor for chronic diseases including metabolic syndrome[3] and cardiovascular conditions[4]. Long-term intake of such snacks fosters a heightened preference for these products, reducing acceptance of healthy and balanced diets. Such dietary preferences may persist into adulthood, increasing the risk of chronic diseases in later life[5]. Although numerous studies have been conducted on children’s eating behaviors, existing research predominantly focuses on the isolated effects of single attributes (e.g., packaging[6]) on snack choices, with limited attention given to a systematic analysis of the formation mechanisms underlying school-aged children’s snack preferences from the perspective of multidimensional attribute interactions.
Against this backdrop, this study employed a DCE methodology combined with a stratified cluster random sampling strategy to recruit 854 junior high school students from Hubei and Anhui provinces.DCE is a quantitative research method grounded in the principles of utility maximization, attribute decomposition, and random utility theory, used to understand individual or group preferences in different choice scenarios[7]. As a flexible preference analysis tool, DCE enables dynamic simulation of multi-attribute interactions and has been widely applied across various fields such as health[8], marketing[9], and willingness-to-pay studies[10]. While most existing research has focused on analyzing adult food choice preferences using DCE[11, 12], this study extends its application to the domain of snack preferences among school-aged children. The findings not only provide actionable insights for snack manufacturers, retailers, and policymakers but also offer a scientific basis for developing integrated "knowledge-attitude-practice" intervention frameworks for student nutrition.
2 Methods
2.1 Data sources
A
A
The data of this study were derived from the "Research on Dietary Preferences and Intervention Strategies for School-Aged Children". Using a stratified cluster random sampling method, one urban middle school and one rural middle school were selected from both Hubei and Anhui provinces as survey sites. Each school was stratified by grade (Grades 7 and 8), with at least 100 students recruited from each stratum to ensure a minimum sample size of ≥ 200 per school. Paper-based questionnaires(Supplementary file 1–3) were administered for on-site data collection. A total of 863 questionnaires were returned, among which 854 were valid, yielding an effective response rate of 98.96%. Informed consent was obtained from all subjects and their parents or legal guardians involved in the study.The study protocol was approved by the Ethics Committee of the National Institute for Nutrition and Health, Chinese Center for Disease Control and Prevention (Ethics Approval No.A
A
: 2024-007).This research was performed in compliance with the World Medical Association’s Declaration of Helsinki, which is a statement of ethical principles for medical research involving human participants.2.2 Research Methods
The study comprised two components: a individual questionnaire survey (Supplementary file 1 ) and a snack preference questionnaire survey (Supplementary file 2 and 3). The discrete choice experiment (DCE) method was employed to design and conduct the investigation of snack preferences among school-aged children.
The experimental design included the following components:
(1) Identification of attributes and levels: Through literature analysis, expert consultation, and group discussions, six snack attributes were identified (taste, nutrient claims, purchase location, price, package size, and social influence)[13–21]. Each attribute contained 3–4 levels[22].
(2) Discrete choice experiment design: The DCE, grounded in random utility theory and characteristic demand theory, is a key method in stated preference research[23]. By designing choice sets comprising different combinations of key attribute levels, respondents were able to make trade-offs based on their preferences. A fractional factorial design was used to optimize the choice sets and reduce the number of alternatives[24]. The %ChoicEff macro in SAS 9.4 was applied to implement a D-optimal design[25], generating 16 choice sets, each containing two alternatives. The 16 choice sets were divided into two versions (eight sets per version), each including two options per sets(See Supplementary file 2 and 3). Respondents were randomly assigned one version during the survey.
(3) Questionnaire content: The survey also collected general demographic information (sex, age, highest parental education level, primary caregiver, etc.). To avoid overestimating experimental attributes due to the lack of an opt-out option or insufficient preference information resulting from excessive opt-out selections[26], a two-stage question approach was adopted. Question 1 asked respondents to choose their preferred alternative, and Question 2 inquired whether they would actually choose that option in real life, thereby reducing biases caused by forced choices or opt-out preferences(Table 1).
(4) Sample size: Based on the rule of thumb for sample size calculation, N ≥ 500c/t×a[27], where c is the largest number of levels in any attribute, t is the number of choice tasks, and a is the number of alternatives per task, the required sample size was determined as N ≥ 125(with c = 4, t = 8, and a = 2 in this study). To meet the precision requirements for subgroup analyses (e.g., urban/rural residence, sex, obesity status), the sample size was increased to 800.
|
Snack Attributes
|
Snack A
|
Snack B
|
||
|---|---|---|---|---|
|
Taste
|
Savory
Flavor
|
 |
Spicy
Flavor
|
 |
|
Nutrient Claims
|
Low Fat
|
 |
Rich Minerals/
Vitamins
|
 |
|
Purchase Location
|
Convenience
Store
|
 |
Supermarket/
Hypermarket
|
 |
|
Price
|
5 ~ 9 ¥CNY
|
  |
≥ 10 ¥CNY
|
   |
|
Package Size
|
Large
|
 |
Medium
|
 |
|
Social Influence
|
Advertising/
Influencer Recommendations
|
 |
Peer Recommendations
|
 |
|
Which option would you be more willing to choose
|
Snack A ≤
|
Snack B ≤
|
||
|
Considering your daily diet, would you consume this snack in real life?
Yes≤ No≤
|
||||
3 Results
3.1 Participants characteristics
The study included a total of 854 participants, all of whom were junior high school students with a mean age of 13.3 years old. Demographic characteristics showed a balanced gender distribution: males accounted for 55.04% (n = 470), and females comprised 44.96% (n = 384). Age distribution was relatively homogeneous: Grade 7 students represented 50.23% (n = 429, mean age 12.79 ± 0.49 years), while Grade 8 students accounted for 49.77% (n = 425, mean age 13.80 ± 0.48 years). In terms of socioeconomic characteristics, the highest parental education level was predominantly junior school or below (37.59%, n = 321), with a small proportion holding postgraduate or higher degrees (5.97%, n = 51); data were missing for 9.37% (n = 80) of participants.
A
Regarding daily living arrangements, the primary caregivers were mostly parents (92.15%, n = 787), the students boarding at school constituted a minority (19.56%, n = 167). However, collective school dining was widely adopted (97.54%, n = 833). Economic support and nutritional awareness indicators revealed that most students had disposable pocket money (78.34%, n = 669) and were aware of nutrition labels on food packages (78.57%, n = 671).Table 2 .|
Variable
|
Number(n)
|
Proportion(%)
|
Mean ± SD
|
|---|---|---|---|
|
Gender
|
|||
|
Males
|
470
|
55.04
|
|
|
Females
|
384
|
44.96
|
|
|
Age(Years)
|
|||
|
Grade 7
|
429
|
50.23
|
12.79 ± 0.49
|
|
Grade 8
|
425
|
49.77
|
13.80 ± 0.48
|
|
School Location
|
|||
|
Rural
|
409
|
47.89
|
|
|
Urban
|
445
|
52.11
|
|
|
Highest Parental Education Level
|
|||
|
Junior School or Below
|
321
|
37.59
|
|
|
High School / Vocational School
|
220
|
25.76
|
|
|
University / College
|
182
|
21.31
|
|
|
Postgraduate or Above
|
51
|
5.97
|
|
|
Unknown
|
80
|
9.37
|
|
|
Primary Caregiver
|
|||
|
Parent(s)
|
787
|
92.15
|
|
|
Grandparent(s)
|
57
|
6.67
|
|
|
Nanny / Other
|
8
|
0.94
|
|
|
Unknow
|
2
|
0.23
|
|
|
Boarding at School
|
|||
|
Yes
|
167
|
19.56
|
|
|
No
|
687
|
80.44
|
|
|
Eating Collective Meals at School
|
|||
|
Yes
|
833
|
97.54
|
|
|
No
|
21
|
2.46
|
|
|
Having Discretionary Pocket Money
|
|||
|
Yes
|
669
|
78.34
|
|
|
No
|
185
|
21.66
|
|
|
Awareness of Nutrition Labels
|
|||
|
Yes
|
671
|
78.57
|
|
|
No
|
183
|
21.43
|
3.2 Snack preferences among school-aged children
The analysis included 854 participants, yielding 13,664 observed data points. The model demonstrated good overall fit (log-likelihood ratio=-4377.84; Likelihood Ratio χ2 = 244.37; P < 0.001), indicating its effectiveness in explaining school-aged children's snack choice behaviors.
Overall, all six attributes demonstrated statistically significant effects on school-aged children's snack choices (P < 0.05). The relative importance of attributes in descending order was: nutrient claims (34.13%), taste (22.89%), social influence (19.72%), purchase location (16.17%), package size (3.70%), and price (3.39%).
Specifically, school-aged children demonstrated a preference for snacks that are commonly consumed by family members (β = 0.452, P < 0.001), purchased in supermarkets (β = 0.439, P < 0.001), with a sweet taste (β = 0.471, P < 0.001), and rich in dietary fiber (β = 0.611, P < 0.001) or vitamins/minerals (β = 0.463, P < 0.001). In contrast, package size and price had relatively smaller influences on preferences, with large package size (β=-0.112, P = 0.019) and price (β=-0.013, P = 0.007) showing slight negative associations. Additionally, the effects of salty taste and medium package size were not statistically significant (P > 0.05) (Table 3).
|
Attributes and levels
|
β(SE)
|
P
|
SD(SE)
|
P
|
RI(%)
|
|---|---|---|---|---|---|
|
Taste (Original Flavor a)
|
22.89
|
||||
|
Sweet Flavor
|
0.471(0.067)
|
< 0.001
|
1.048(0.085)
|
< 0.001
|
|
|
Savory Flavor
|
0.015(0.053)
|
0.771
|
0.393(0.100)
|
< 0.001
|
|
|
Spicy Flavor
|
0.399(0.072)
|
< 0.001
|
1.223(0.088)
|
< 0.001
|
|
|
Nutrient Claims (None a)
|
|||||
|
Low-fat
|
0.399(0.058)
|
< 0.001
|
0.035(0.276)
|
0.899
|
34.13
|
|
Rich Minerals / Vitamins
|
0.463(0.055)
|
< 0.001
|
-0.028(0.157)
|
0.857
|
|
|
High Dietary Fiber
|
0.611(0.055)
|
< 0.001
|
-0.135(0.136)
|
0.320
|
|
|
Purchase Location (Near School a)
|
16.17
|
||||
|
Convenience Store
|
0.098(0.050)
|
0.047
|
0.330(0.122)
|
0.007
|
|
|
Supermarket / Hypermarket
|
0.439(0.048)
|
< 0.001
|
0.372(0.121)
|
0.002
|
|
|
Social Influence (Advertising / Influencer Recommendations a)
|
19.72
|
||||
|
Peer Recommendations
|
0.253(0.041)
|
< 0.001
|
-0.149(0.124)
|
0.230
|
|
|
Family Consumption Habits
|
0.452(0.048)
|
< 0.001
|
-0.447(0.083)
|
< 0.001
|
|
|
Package Size (Small a)
|
3.7
|
||||
|
Medium
|
0.010(0.048)
|
0.830
|
0.099(0.176)
|
0.575
|
|
|
Large
|
-0.112(0.048)
|
0.019
|
-0.164(0.243)
|
0.501
|
|
|
Price
|
-0.013(0.005)
|
0.007
|
3.39
|
||
|
Sample Size
|
854
|
||||
|
Number of Observations
|
13664
|
||||
|
Log Likelihood
|
-4377.84
|
||||
|
LRchi2(12)
|
244.37
|
||||
|
Prob > chi2
|
0.00
|
||||
| a (ref) | |||||
3.3 Heterogeneity Analysis of Snack Preferences Among School-Aged Children with Different Characteristics
3.3.1 Analysis of Snack Preferences by Gender
The relative importance ranking of snack attributes was consistent across genders. However, compared to girls, boys showed stronger influences from social influence (RI = 21.01%) and price (RI = 1.04%), demonstrating a marked preference for snacks rich in dietary fiber (β = 0.554, P < 0.01). Package size did not significantly affect their preferences, while they exhibited a distinct negative preference for higher prices (β=-0.026, P < 0.01). In contrast, girls placed greater emphasis on package size (RI = 8.02%) and preferred low-fat snacks (β = 0.738, P < 0.01). They showed a significant negative preference for large package sizes (β=-0.259, P < 0.01), while price had no statistically significant effect on their snack choices. See Table 4.
|
Attributes and levels
|
Boys
|
Girls
|
||
|---|---|---|---|---|
|
β(SE)
|
RI(%)
|
β(SE)
|
RI(%)
|
|
|
Taste (Original Flavor a)
|
24.30
|
21.41
|
||
|
Sweet Flavor
|
0.423**(0.095)
|
0.540**(0.100)
|
||
|
Savory Flavor
|
0.037(0.072)
|
-0.030(0.082)
|
||
|
Spicy Flavor
|
0.334**(0.095)
|
0.521**(0.111)
|
||
|
Nutrient Claims (None a)
|
34.96
|
36.34
|
||
|
Low-fat
|
0.162*(0.078)
|
0.738**(0.097)
|
||
|
Rich Minerals / Vitamins
|
0.481**(0.075)
|
0.472**(0.085)
|
||
|
High Dietary Fiber
|
0.554**(0.076)
|
0.718**(0.087)
|
||
|
Purchase Location (Near School a)
|
15.97
|
16.37
|
||
|
Convenience Store
|
0.030(0.068)
|
0.187*(0.075)
|
||
|
Supermarket /Hypermarket
|
0.385**(0.065)
|
0.507**(0.078)
|
||
|
Social Influence (Advertising / Influencer Recommendations a)
|
21.01
|
17.83
|
||
|
Peer Recommendations
|
0.180**(0.055)
|
0.366**(0.065)
|
||
|
Family Consumption Habits
|
0.436**(0.065)
|
0.471**(0.072)
|
||
|
Package Size (Small a)
|
2.72
|
8.02
|
||
|
Medium
|
0.064(0.066)
|
-0.070(0.076)
|
||
|
Large
|
0.008(0.066)
|
-0.259**(0.074)
|
||
|
Price
|
-0.026**(0.007)
|
1.04
|
0.001(0.007)
|
0.03
|
|
Sample Size
|
470
|
384
|
||
|
Number of Observations
|
7520
|
6144
|
||
|
Log Likelihood
|
-2414.94
|
-1931.38
|
||
|
LRchi2(12)
|
142.75**
|
106.33**
|
||
| **p < 0.01,*p < 0.05 | ||||
| a (ref) | ||||
3.3.2 Analysis of Snack Preferences by Parental Education Level
Results revealed significant variations in snack preferences among children based on parental education levels. Children whose parents had a junior high school education or below placed greater emphasis on taste (RI = 25.33%) and nutrient claims (RI = 37.50%). For those with parents holding a high school or vocational secondary education, social influence (RI = 25.46%) was considered more important than taste (RI = 21.55%). In contrast, children of parents with college or higher education levels showed greater concern for package size (RI = 10.43%).
Children whose parents had a junior high school education or below demonstrated a preference for sweet-tasting snacks (β = 0.485, P < 0.01) and those rich in dietary fiber (β = 0.596, P < 0.01), while showing a slight negative preference for price (β=-0.017, P < 0.05). In contrast, children with parents holding a high school or vocational secondary education preferred spicy-tasting snacks (β = 0.434, P < 0.01). Meanwhile, children of parents with college or higher education exhibited stronger preferences for sweet taste (β = 0.529, P < 0.01) and snacks rich in vitamins/minerals (β = 0.667, P < 0.01), along with a significant negative preference for large package sizes (β=-0.325, P < 0.01). Price did not significantly influence their preferences. Details are presented in Table 5.
|
Attributes and levels
|
Junior School or Below
|
High School/Vocational School
|
University/College or Above
|
|||
|---|---|---|---|---|---|---|
|
β(SE)
|
RI(%)
|
β(SE)
|
RI(%)
|
β(SE)
|
RI(%)
|
|
|
Taste (Original Flavor a)
|
25.33
|
21.55
|
20.84
|
|||
|
Sweet Flavor
|
0.485**(0.110)
|
0.429**(0.150)
|
0.529**(0.136)
|
|||
|
Savory Flavor
|
-0.047(0.088)
|
0.026(0.112)
|
-0.003(-0.108)
|
|||
|
Spicy Flavor
|
0.407**(0.116)
|
0.434**(0.144)
|
0.477**(0.158)
|
|||
|
Nutrient Claims (None a)
|
37.50
|
33.58
|
33.62
|
|||
|
Low-fat
|
0.354**(0.098)
|
0.319*(0.124)
|
0.494**(0.122)
|
|||
|
Rich Minerals / Vitamins
|
0.423**(0.091)
|
0.441**(0.118)
|
0.667**(0.123)
|
|||
|
High Dietary Fiber
|
0.596**(0.093)
|
0.651**(0.117)
|
0.644**(0.113)
|
|||
|
Purchase Location (Near School a)
|
16.55
|
17.39
|
16.58
|
|||
|
Convenience Store
|
0.041(0.088)
|
0.138(0.101)
|
0.101(0.100)
|
|||
|
Supermarket / Hypermarket
|
0.414**(0.078)
|
0.465**(0.113)
|
0.559**(0.111)
|
|||
|
Social Influence (Advertising / Influencer Recommendations a)
|
17.66
|
25.46
|
17.91
|
|||
|
Peer Recommendations
|
0.171*(0.070)
|
0.359**(0.087)
|
0.339**(0.084)
|
|||
|
Family Consumption Habits
|
0.378**(0.076)
|
0.602**(0.104)
|
0.489**(0.097)
|
|||
|
Package Size (Small a)
|
2.30
|
1.48
|
10.43
|
|||
|
Medium
|
0.066(0.082)
|
0.050(0.104)
|
-0.117(0.101)
|
|||
|
Large
|
-0.055(0.079)
|
-0.009(0.103)
|
-0.325**(0.100)
|
|||
|
Price
|
-0.017*(0.008)
|
0.65
|
-0.020*(0.010)
|
0.65
|
0.023(0.045)
|
0.63
|
|
Sample Size
|
321
|
220
|
233
|
|||
|
Number of Observations
|
5136
|
3520
|
3728
|
|||
|
Log Likelihood
|
-1649.82
|
-1120.30
|
-1174.74
|
|||
|
LRchi2(12)
|
98.21**
|
66.25**
|
88.66**
|
|||
| **p < 0.01,*p < 0.05 | ||||||
| a (ref) | ||||||
3.3.3 Analysis of Snack Preferences by BMI
Compared to overweight/obese children, those with normal weight placed greater emphasis on purchase location (RI = 19.01%) and preferred snacks with sweet taste (β = 0.510, P < 0.01) that were rich in dietary fiber (β = 0.612, P < 0.01). Package size and price did not significantly influence their preferences. In contrast, overweight/obese children were more influenced by nutrient claims (RI = 39.47%) and price (RI = 3.84%), selecting snacks that were similarly rich in dietary fiber (β = 0.684, P < 0.01) but with a spicy taste (β = 0.448, P < 0.01). They also exhibited a negative preference for price (β=-0.115, P < 0.01). See Table 6.
|
Attributes and levels
|
Normal BMI
|
Overweight / Obesity
|
||
|---|---|---|---|---|
|
β(SE)
|
RI(%)
|
β(SE)
|
RI(%)
|
|
|
Taste (Original Flavor a)
|
24.00
|
21.60
|
||
|
Sweet Flavor
|
0.510**(0.087)
|
0.434**(0.117)
|
||
|
Savory Flavor
|
0.032(0.064)
|
-0.035(0.099)
|
||
|
Spicy Flavor
|
0.396**(0.093)
|
0.448**(0.119)
|
||
|
Nutrient Claims (None a)
|
33.94
|
39.47
|
||
|
Low-fat
|
0.381**(0.075)
|
0.441**(0.104)
|
||
|
Rich Minerals / Vitamins
|
0.443**(0.072)
|
0.557**(0.100)
|
||
|
High Dietary Fiber
|
0.612**(0.072)
|
0.684**(0.101)
|
||
|
Purchase Location (Near School a)
|
19.01
|
9.58
|
||
|
Convenience Store
|
0.165**(0.062)
|
-0.092(0.091)
|
||
|
Supermarket / Hypermarket
|
0.491**(0.062)
|
0.333**(0.085)
|
||
|
Social Influence (Advertising / Influencer Recommendations a)
|
19.59
|
19.70
|
||
|
Peer Recommendations
|
0.284**(0.052)
|
0.205**(0.074)
|
||
|
Family Consumption Habits
|
0.449**(0.060)
|
0.488**(0.084)
|
||
|
Package Size (Small a)
|
2.70
|
5.81
|
||
|
Medium
|
0.039(0.061)
|
-0.032(0.084)
|
||
|
Large
|
-0.101(0.06)
|
-0.158(0.088)
|
||
|
Price
|
-0.023(0.027)
|
0.76
|
-0.115**(0.039)
|
3.84
|
|
Sample Size
|
546
|
308
|
||
|
Number of Observations
|
8736
|
4928
|
||
|
Log Likelihood
|
-2789.44
|
-1575.38
|
||
|
LRchi2(12)
|
157.86
|
100.81
|
||
| **p < 0.01,*p < 0.05 | ||||
| a (ref) | ||||
3.4 Willingness-to-Pay Analysis
Substantial heterogeneity was observed in children's willingness-to-pay for different snack attributes. Specifically, snacks rich in dietary fiber commanded the highest premium (WTP = 10.89 CNY). Regarding taste, children were willing to pay 8.40 CNY more for sweet-tasting snacks compared to plain alternatives. For social influence, snacks commonly consumed by family members yielded a 8.05 CNY higher WTP than those promoted by advertisements or internet celebrities. In terms of purchase location, children demonstrated a 7.81 CNY greater WTP for snacks available in supermarkets compared to those sold near schools. It should be noted that package size did not exert a statistically significant influence on WTP (P > 0.05). SeeFigure 1.
A
Fig. 1
Willingness-to-Pay Analysis of Snack Preferences (n = 854)
3.5 Sensitivity Analysis
Model I included all 13,664 observations, while Model II excluded 3,276 observations from respondents who opted out (selecting "would not choose") in Question 2. Mixed logit regression analyses were conducted for both models, revealing statistically significant goodness-of-fit for each (P < 0.001). Furthermore, the regression coefficients for snack attribute levels in both models were consistent in direction, with negligible differences in magnitude and statistical significance. This indicates that the exclusion of opt-out observations did not substantially alter the overall regression results. Collectively, these findings demonstrate the robustness and validity of all analytical outcomes.See Table 7.
|
Attributes and levels
|
Model Ⅰ
|
Model Ⅱ
|
||
|---|---|---|---|---|
|
β(SE)
|
SD(SE)
|
β(SE)
|
SD(SE)
|
|
|
Taste (Original Flavor a)
|
||||
|
Sweet Flavor
|
0.471**(0.067)
|
1.048(0.085)
|
0.626** (0.078)
|
1.044 (0.106)
|
|
Savory Flavor
|
0.015(0.053)
|
0.393(0.100)
|
0.060 (0.064)
|
0.517 (0.129)
|
|
Spicy Flavor
|
0.399**(0.072)
|
1.223(0.088)
|
0.528**(0.083)
|
1.279 (0.109)
|
|
Nutrient Claims (None a)
|
||||
|
Low-fat
|
0.399**(0.058)
|
0.035(0.276)
|
0.482**(0.071)
|
0.060 (0.315)
|
|
Rich Minerals / Vitamins
|
0.463**(0.055)
|
-0.028(0.157)
|
0.450** (0.066)
|
-0.071 (0.139)
|
|
High Dietary Fiber
|
0.611**(0.055)
|
-0.135(0.136)
|
0.599**(0.067)
|
-0.035 (0.203)
|
|
Purchase Location (Near School a)
|
||||
|
Convenience Store
|
0.098*(0.050)
|
0.330(0.122)
|
0.112 (0.058)
|
0.379 (0.144)
|
|
Supermarket / Hypermarket
|
0.439**(0.048)
|
0.372(0.121)
|
0.444** (0.057)
|
0.351 (0.146)
|
|
Social Influence (Advertising / Influencer Recommendations a)
|
||||
|
Peer Recommendations
|
0.253**(0.041)
|
-0.149(0.124)
|
0.338** (0.050)
|
0.068 (0.127)
|
|
Family Consumption Habits
|
0.452**(0.048)
|
-0.447(0.083)
|
0.577**(0.057)
|
0.383 (0.121)
|
|
Package Size (Small a)
|
||||
|
Medium
|
0.010(0.048)
|
0.099(0.176)
|
0.015 (0.056)
|
-0.017 (0.262)
|
|
Large
|
-0.112*(0.048)
|
-0.164(0.243)
|
-0.069 (0.057)
|
0.336 (0.137)
|
|
Price
|
-0.013**(0.005)
|
-0.016** (0.006)
|
||
|
Sample Size
|
854
|
854
|
||
|
Number of Observations
|
13664
|
10388
|
||
|
Loglikelihood
|
-4377.84
|
-3289.73
|
||
|
LRchi2(12)
|
244.37
|
172,79
|
||
|
Attributes and levels
|
0.00
|
0.00
|
||
| **p < 0.01,*p < 0.05 | ||||
| a (ref) | ||||
Model I comprised observations from the full sample, while Model II included only those observations that selected "Yes" in Question 2.
4 Discussion
This study systematically assessed the key attribute levels and heterogeneity in snack preferences among school-aged children through a DCE, identifying nutrient claims, taste, and social influence as core drivers influencing children's choices. The findings align with existing research[28, 29]while also revealing characteristics specific to the Chinese context, thereby providing new perspectives for designing targeted interventions.
A
Nutrient claims demonstrated the highest relative importance in snack selection, with snacks rich in dietary fiber and vitamins/minerals being particularly preferred. This finding aligns with Batista’s [30] theory that "positive nutrition labels serve as drivers for choosing healthier foods." However, our study revealed that the premium effect for dietary fiber (WTP 10.89 CNY) exceeded that for vitamins/minerals, suggesting that school-aged children may prioritize gut health over macronutrient supplementation. Both sweet and spicy tastes exhibited positive preferences, while salty taste showed no significant effect. This may be related to children’s innate preference for sweet taste[31, 32], whereas the prominence of spicy preference might be attributed to the widespread availability of spicy foods (e.g., spicy strips, puffed snacks) within China’s local dietary culture[33, 34]. Regarding social influence, frequent family consumption and peer recommendations reflected the social shaping of children’s dietary choices[35]. However, the influence of family consumption substantially surpassed that of peer recommendations, indicating that the home environment remains a critical leverage point for cultivating healthy eating behaviors among school-aged children[19, 36].From the perspective of subgroup heterogeneity, boys preferred snacks rich in dietary fiber with larger package sizes, while girls paid greater attention to low-fat options and smaller packages. Studies indicate that people generally hold both implicit and explicit associations linking males with large portions and females with small portions[37]. Such gender stereotypes may originate from sociocultural expectations regarding eating behaviors for men and women. Females often face pressure to maintain slender figures, making them particularly conscious of weight management and dietary choices to avoid guilt associated with overeating. Males’ preference for larger portions may be related to energy requirements[38] and evolutionary psychology[39], though the risks of excessive consumption should not be overlooked. Children with less-educated parents were more influenced by taste, whereas those from highly-educated families focused more on package size. This may occur because education can indirectly influence food choices by altering health consciousness[40].Overweight/obese children showed a significantly stronger preference for spicy snacks than their normal-weight counterparts. Although capsaicin may aid weight control by influencing energy metabolism and appetite regulation, the preference for spicy snacks in this group may be more closely related to seeking sensory stimulation[41].
The Snack Guidelines for Chinese Children and Adolescents (2018)[42] provide specific recommendations for school-aged children: snacks should supplement main meals and be consumed in small quantities (not exceeding 10% of daily energy intake); snacks during breaks should prioritize fruits, dairy, and nuts (avoiding high-sugar beverages and fried foods); high-salt, high-sugar, and high-fat snacks should be limited, and alcohol- or caffeine-containing beverages should be avoided; oral hygiene should be maintained (rinsing or brushing after snacking), and no snacks should be consumed within one hour before bedtime. The General Standard for the Labelling of Prepackaged Foods[43] mandates that all prepackaged food labels must include core information: the product name accurately reflecting its characteristics, an ingredient list in descending order of proportion, net content, name and address of the manufacturer, country of origin (if omission would be misleading), clear date marking ("best before" or "use by" date) with corresponding storage conditions. A key emphasis of the 2018 revision is mandatory allergen labeling, requiring clear identification of eight major allergens: gluten-containing grains, milk, eggs, fish, peanuts, soybeans, tree nuts, crustaceans, and high-concentration sulfites. These regulations enable parents and children to identify healthier snacks through nutrition labels. Additionally, schools should integrate food safety education into daily teaching activities. Through parent-teacher meetings, public health classes, and other channels, students and parents should be educated on how to appropriately use nutrition labels or nutrient claims to make informed snack choices and develop healthy eating habits. Simultaneously, snack manufacturers and retailers must strictly comply with labeling regulations, refrain from false advertising (e.g., exaggerated nutrient claims), avoid using inductive language to attract children, and voluntarily accept oversight from market regulatory authorities. In summary, integrating the findings of this study, we recommend that student nutrition intervention strategies incorporate regional factors, snack attributes, economic incentives (e.g., discounts or rewards via payment platforms), and health education (e.g., family empowerment). These strategies should be tailored to children’s gender, physical status, and family background. Multichannel approaches–such as public lectures, short videos, educational games, parent-teacher meetings, and handcrafted health posters–should be employed to engage both children and parents, enhance participation and educational outcomes, and reduce the purchase and consumption of unhealthy snacks.
4.1 Limitations
The survey was limited to Hubei and Anhui provinces, resulting in limited geographical representativeness;Attributes such as food additives, packaging style, and brand were not included, potentially underestimating children’s attention to technological food properties;The hypothetical bias inherent in DCE may overestimate the actual selection probability of healthy attributes[44]. Subsequent studies should combine multiple methods such as eye-tracking experiments[45] for validation.
4.2 Future Research
Future studies could expand to regions with significant variations in economic development levels and dietary habits to explore the impact of various emerging attributes on snack choices among school-aged children. Simultaneously, neuroimaging techniques[46] could be integrated to elucidate the neural mechanisms underlying children’s dietary decision-making, providing biological evidence to support intervention frameworks.
Abbreviations
DCE
Discrete Choice Experiment
WTP
Willingness-to-Pay
RI
Relative Importance
β
Beta coefficient
P
P-value
SE
Standard Error
SD
Standard Deviation
BMI
Body Mass Index
Ref
Electronic Supplementary Material
Below is the link to the electronic supplementary material
A
Author Contribution
HL drafted and revised the manuscript. WZ, XW and XZ collected and analyzed the data. FG, XJ, QG, JZ, and SL provided guidance and technical support. WD conceived the study, obtained funding, and critically reviewed the manuscript. HW and AL provided support and guidance. All authors read and approved the final manuscript.
A
Data Availability
The datasets generated and/or analysed during the current study are not publicly available due [REASON WHY DATA ARE NOT PUBLIC] but are available from the corresponding author on reasonable request.
A
Acknowledgement
We would like to thank all the school-aged children who participated in this study for their time and cooperation. We are also grateful to the teachers and school staff who assisted with the data collection process.
Reference
Ethics approval and consent to participate
Informed consent was obtained from all subjects and their parents or legal guardians involved in the study.The study protocol was approved by the Ethics Committee of the National Institute for Nutrition and Health, Chinese Center for Disease Control and Prevention (Ethics Approval No. 2024-007).This research was performed in compliance with the World Medical Association’s Declaration of Helsinki, which is a statement of ethical principles for medical research involving human participants.
Consent for publication
Not Applicable.
Availability of data and materials
The data of this study were derived from the "Research on Dietary Preferences and Intervention Strategies for School-Aged Children". The datasets generated and/or analysed during the current study are not publicly available due [REASON WHY DATA ARE NOT PUBLIC] but are available from the corresponding author on reasonable request.
Competing interests
The authors declare that they have no competing interests or personal relationships that could have appeared to influence the work reported in this paper.
Funding
China Student Nutrition and Health Promotion Association - Mead Johnson Academic Excellence Nutrition Research Fund Project (CASNHP-MJN2023-18).
Authors' contributions
HL drafted and revised the manuscript. WZ, XW and XZ collected and analyzed the data. FG, XJ, QG, JZ, and SL provided guidance and technical support. WD conceived the study, obtained funding, and critically reviewed the manuscript. HW and AL provided support and guidance. All authors read and approved the final manuscript.
Acknowledgements
We would like to thank all the school-aged children who participated in this study for their time and cooperation. We are also grateful to the teachers and school staff who assisted with the data collection process.
References 1. Zhen S, Ma Y, Zhao Z, Yang X, Wen D: Dietary pattern is associated with obesity in Chinese children and adolescents: data from China Health and Nutrition Survey (CHNS). Nutr J 2018, 17(1):68. 2. Mahmoud SA, El Moshy S, Rady D, Radwan IA, Abbass MMS, Al Jawaldeh A: The effect of unhealthy dietary habits on the incidence of dental caries and overweight/obesity among Egyptian school children (A cross-sectional study). Front Public Health 2022, 10:953545. 3. Fiore G, Pascuzzi MC, Di Profio E, Corsello A, Agostinelli M, La Mendola A, Milanta C, Campoy C, Calcaterra V, Zuccotti G et al: Bioactive compounds in childhood obesity and associated metabolic complications: Current evidence, controversies and perspectives. Pharmacol Res 2023, 187:106599. 4. Vos MB, Kaar JL, Welsh JA, Van Horn LV, Feig DI, Anderson CAM, Patel MJ, Cruz Munos J, Krebs NF, Xanthakos SA et al: Added Sugars and Cardiovascular Disease Risk in Children: A Scientific Statement From the American Heart Association. Circulation 2017, 135(19):e1017-e1034. 5. Koopman KE, Caan MW, Nederveen AJ, Pels A, Ackermans MT, Fliers E, la Fleur SE, Serlie MJ: Hypercaloric diets with increased meal frequency, but not meal size, increase intrahepatic triglycerides: a randomized controlled trial. Hepatology 2014, 60(2):545–553. 6. Dixon HG, Tran A, Allan MO, Capes HR, Hayward CL: Seeing is believing: How front-of-pack food imagery shapes parents' perceptions of child-oriented snack foods. Appetite 2025, 206:107804. 7. Prosser LA: Statistical Methods for the Analysis of Discrete-Choice Experiments: A Report of the ISPOR Conjoint Analysis Good Research Practices Task Force. Value Health 2016, 19(4):298–299. 8. Sturkenboom R, Keszthelyi D, Masclee AAM, Essers BAB: Discrete Choice Experiment Reveals Strong Preference for Dietary Treatment Among Patients With Irritable Bowel Syndrome. Clin Gastroenterol Hepatol 2022, 20(11):2628–2637. 9. Maró ZM, Balogh P, Czine P, Török Á: The roles of geographic indication and ethnocentrism in the preferences of Central European spirit consumers: The case of pálinka. Food Quality and Preference 2023, 108:104878. 10. Chen Y-H, Qiu K-H, Liu KE, Chiang C-Y: Are Consumers Willing to Pay a Premium for Pure Rice Noodles? A Study of Discrete Choice Experiments in Taiwan. Sustainability 2020, 12(15):6144. 11. Kamphuis CBM, de Bekker-Grob EW, van Lenthe FJ: Factors affecting food choices of older adults from high and low socioeconomic groups: a discrete choice experiment. The American Journal of Clinical Nutrition 2015, 101(4):768–774. 12. Livingstone KM, Abbott G, Lamb KE, Dullaghan K, Worsley T, McNaughton SA: Understanding Meal Choices in Young Adults and Interactions with Demographics, Diet Quality, and Health Behaviors: A Discrete Choice Experiment. The Journal of Nutrition 2021, 151(8):2361–2371. 13. Afshin A, Peñalvo JL, Del Gobbo L, Silva J, Michaelson M, O'Flaherty M, Capewell S, Spiegelman D, Danaei G, Mozaffarian D: The prospective impact of food pricing on improving dietary consumption: A systematic review and meta-analysis. (1932–6203 (Electronic)). 14. Corazza I, Pennucci F, De Rosis S: Promoting healthy eating habits among youth according to their preferences: Indications from a discrete choice experiment in Tuscany. (1872–6054 (Electronic)). 15. Rusmevichientong P, Jaynes J, Chandler L: Understanding influencing attributes of adolescent snack choices: Evidence from a discrete choice experiment. Food Quality and Preference 2021, 92:104171. 16. Sadeghirad B, Duhaney T, Motaghipisheh S, Campbell NR, Johnston BC: Influence of unhealthy food and beverage marketing on children's dietary intake and preference: a systematic review and meta-analysis of randomized trials. (1467-789X (Electronic)). 17. Hartmann M, Cash SB, Yeh CH, Landwehr SC, McAlister AR: Children's purchase behavior in the snack market: Can branding or lower prices motivate healthier choices? Appetite 2017, 117:247–254. 18. Svisco E, Byker Shanks C, Ahmed S, Bark K: Variation of Adolescent Snack Food Choices and Preferences along a Continuum of Processing Levels: The Case of Apples. Foods 2019, 8(2). 19. Damen FWM, Luning PA, Fogliano V, Steenbekkers BLPA: What influences mothers’ snack choices for their children aged 2–7? Food Quality and Preference 2019, 74:10–20. 20. Khomsan A, Anwar F, Riyadi H, Navratilova HF: Children’s food habits, consumption, and food safety of popular snacks in school environment in Indonesia. International Journal Of Community Medicine And Public Health 2022, 10(1):119–125. 21. Benito-Ostolaza JM, Echavarri R, Garcia-Prado A, Oses-Eraso N: Using visual stimuli to promote healthy snack choices among children. Soc Sci Med 2021, 270:113587. 22. de Bekker-Grob EW, Ryan M Fau - Gerard K, Gerard K: Discrete choice experiments in health economics: a review of the literature. (1099 − 1050 (Electronic)). 23. Johnson BJ, Golley RK, Zarnowiecki D, Hendrie GA, Huynh EK: Understanding the influence of physical resources and social supports on primary food providers' snack food provision: a discrete choice experiment. Int J Behav Nutr Phys Act 2020, 17(1):155. 24. Orme B: Getting Started with Conjoint Analysis: Strategies for Product Design and Pricing Research(3rd Edition). Research Publishers LLC.; 2014. 25. Alamri AS, Georgiou S, Stylianou S: Discrete choice experiments: An overview on constructing D-optimal and near-optimal choice sets. Heliyon 2023, 9(7):e18256. 26. Dhar R, Simonson I: The Effect of Forced Choice on Choice. Journal of Marketing Research 2003, 40(2):146–160. 27. Orme BK: Getting Started with Conjoint Analysis: Strategies for Product Design and Pricing Research. In: 2005; 2005. 28. McCann J, Woods J, Mohebbi M, Russell CG: Regulated nutrition claims increase perceived healthiness of an ultra-processed, discretionary toddler snack food and ultra-processed toddler milks: A discrete choice experiment. Appetite 2022, 174:106044. 29. Baenziger PS, Frels K, Greenspan S, Jones J, Lovegrove A, Rose D, Shewry P, Wallace R: A stealth health approach to dietary fibre. Nat Food 2023, 4(1):5–6. 30. Batista MF, de Carvalho-Ferreira JP, Thimoteo da Cunha D, De Rosso VV: Front-of-package nutrition labeling as a driver for healthier food choices: Lessons learned and future perspectives. Compr Rev Food Sci Food Saf 2023, 22(1):535–586. 31. Schiff HC, Kogan JF, Isaac M, Czarnecki LA, Fontanini A, Maffei A: Experience-dependent plasticity of gustatory insular cortex circuits and taste preferences. Sci Adv 2023, 9(2):eade6561. 32. Lif Holgerson P, Hasslöf P, Esberg A, Haworth S, Domellöf M, West CE, Johansson I: Genetic Preference for Sweet Taste in Mothers Associates with Mother-Child Preference and Intake. Nutrients 2023, 15(11). 33. Wen Q, Wei Y, Du H, Lv J, Guo Y, Bian Z, Yang L, Chen Y, Chen Y, Shi L et al: Characteristics of spicy food consumption and its relation to lifestyle behaviours: results from 0.5 million adults. Int J Food Sci Nutr 2021, 72(4):569–576. 34. Sun D, Lv J, Chen W, Li S, Guo Y, Bian Z, Yu C, Zhou H, Tan Y, Chen J et al: Spicy food consumption is associated with adiposity measures among half a million Chinese people: the China Kadoorie Biobank study. BMC Public Health 2014, 14:1293. 35. Zhou M, Bian B, Huang L: Do Peers Matter? Unhealthy Food and Beverages Preferences among Children in a Selected Rural Province in China. Foods 2023, 12(7). 36. Mahmood L, Flores-Barrantes P, Moreno LA, Manios Y, Gonzalez-Gil EM: The Influence of Parental Dietary Behaviors and Practices on Children's Eating Habits. Nutrients 2021, 13(4). 37. Méjean C, Deglaire A, Kesse-Guyot E, Hercberg S, Schlich P, Castetbon K: Association between intake of nutrients and food groups and liking for fat (The Nutrinet-Santé Study). Appetite 2014, 78:147–155. 38. Liang S, He J, Deng W, Cao P, Teng L, Tang Y, Lu X, Hu F, Zhang T, Li J: The Effect of Unhealthy Food Packaging Information Boundaries on Consumer Purchasing Intentions. Foods 2024, 13(15). 39. Duszka K, Hechenberger M, Dolak I, Kobiljak D, König J: Gender, Age, Hunger, and Body Mass Index as Factors Influencing Portion Size Estimation and Ideal Portion Sizes. Front Psychol 2022, 13:873835. 40. Teng CC, Lu CH: Organic food consumption in Taiwan: Motives, involvement, and purchase intention under the moderating role of uncertainty. Appetite 2016, 105:95–105. 41. Frederiksen C, Byrne DV, Andersen BV: Sensing the Snacking Experience: Bodily Sensations Linked to the Consumption of Healthy and Unhealthy Snack Foods-A Comparison between Body Mass Index Levels. Foods 2024, 13(3). 42. Society. CN: Chinese Children and Adolescents Snack Guide (2018). In. Beijing: People's Medical Publishing House, 2018. 43. Commission. CA. In: General Standard for the Labelling of Prepackaged Foods (CXS 1-1985). Rome: FAO/WHO; 1985. 44. Veldwijk J, Marceta SM, Swait JD, Lipman SA, de Bekker-Grob EW: Taking the Shortcut: Simplifying Heuristics in Discrete Choice Experiments. The Patient - Patient-Centered Outcomes Research 2023, 16(4):301–315. 45. Soekhai V, de Bekker-Grob EW, Ellis AR, Vass CM: Discrete Choice Experiments in Health Economics: Past, Present and Future. PharmacoEconomics 2018, 37(2):201–226. 46. Pyatin V, Maslova O, Romanchuk N, Volobuev A, Bulgakova S, Romanov D, Sirotko I: Neuroimaging: Structural, Functional, Pharmacological, Bioelementology and Nutritionology. Bulletin of Science and Practice 2021, 7(10):145–184.
1
* Corresponding author. National Institute of Nutrition and Health, Chinese Center for Disease Control and Prevention, Beijing10050, China.
E-mail address: liuad@ninh.chinacdc.cn (Liu Aidong).
* Corresponding author. National Institute of Nutrition and Health, Chinese Center for Disease Control and Prevention, Beijing10050, China.
E-mail address: wanghj@ninh.chinacdc.cn (Wang Huijun).