Several Non-Communicable Diseases are associated with diet behaviors ( Almoraie et al., 2021) that include snacking. Snacking is largely considered unhealthy ( Shatwan et al., 2022) but also partly healthy ( Damen et al., 2020; Schlinkert et al., 2020). It is highly prevalent among young adults around the world. For instance, college students in Asia and the Middle Eastern countries are fond of snacking ( Shatwan et al., 2022). As young adults, behaviors learned at this stage tend to linger on throughout their adulthood ( Rodrigues et al., 2019).

Scholars have tried to understand individual behavioral choices through different theoretical explanations. The Social Cognitive Theory (SCT), for instance, asserts that people acquire new behaviors through observing and imitating others, within their social context ( Bandura, 1971). When an observed action is considered rewarding, the observer will replicate the action later. This theoretical explanation suggests that young adults’ eating behaviors result from listening and watching feeding habits of role models ( Powell et al., 2021). However, individuals may mentally learn new habits but not immediately display the learned habits. Individuals may have intentions to display the learned behavior at a later time. Schwarzer’s (2008) Health Action Process approach (HAPA), as another theoretical explanation, posits that planning is a crucial mediator between individual intentions and actual behavior. Planning details the boundaries of when, where, and how the behavior will be done ( Domke et al., 2021). One manifestation of planning is Meal planning ( Alm & Olsen, 2017). It is a skill of scheduling meals before consumption time involving, the meal location, meal time and food kind ( Domke et al., 2021). The HAPA theoretical explanations suggest that when the youth watch the eating habits of role models, they schedule to eat similar foods. This meal planning then leads to taking a diet similar to their role models. Theory, therefore, suggests that learning may lead to preparation to do the learned behavior ( Bandura, 1971), which preparation then translates to replicating the learned behavior ( Parkinson et al., 2023).

Existing research appears not to have extensively examined the connection between role modeling, meal planning, and snack choices ( Hamilton et al., 2017). Previous studies have primarily focused on the influence of role modeling on snack choices, with limited attention given to the impact of TV cooking shows ( Ngqangashe et al., 2018). Additionally, the majority of research in this area has concentrated on children and adolescents, with minimal investigation within African sub-populations. There is also a lack of studies examining the relationships between role modeling and meal planning ( Yee et al., 2017), as well as meal planning and snack choices ( Ducrot et al., 2017). To address these gaps, this study aims to investigate the mediating role of meal planning in the link between role modeling and snack choices among young adults in Uganda.

The specific study objectives are to examine the relationship between: i)

Role Modeling and Snack Choices

Results show that the study sample had 197 males (48.9%) and 206 females (51.1%). The respondents’ mean age was 21 years. 89.8% of the respondents were pursuing a bachelor’s degree. Many respondents resided in off-campus residences (61.5%), and most of them (58.8%) resided alone in their residences.

We carried out CFA to examine whether the item measures were consistent with the construct nature ( Hair et al., 2014). Table 1 (Extended data) ( Kiggwe et al., 2024) shows several commonly used fit indices we employed to assess the overall model fit ( Hu & Bentler, 1999). The comprehensive goodness-of-fit indices produced a Chi-square of 458.888 and Chi-Square/DF=1.779.

We considered the comparative fit index (CFI) value of 0.947, incremental fit index (IFI) value of 0.947, and Tucker-Lewis index (TLI) value of 0.938, as very good fits to the model based on Hu and Bentler (1999) assertion that for these indices a value of 0.7 and above is satisfactory, 0.8 and above is good, and 0.9 and above is very good. The root mean square error of approximation (RMSEA) value was 0.044 indicating a close fit. Thus, the results in the measurement model (see Figure 1) indicate an acceptable fit.

All standardized factor loadings should be at least 0.5 and statistically significant to indicate that observed indicators are strongly related to their associated constructs ( Hair et al., 2014). In the model, all standardized factor loadings were above 0.5 and significant (see Figure 1). These findings indicate that there was a satisfactory fit between the proposed model and the data. The test for convergent validity revealed that the composite reliability coefficients for all constructs were above 0.9. In addition, the Average Variance Extracted (AVE) for all constructs met the threshold of 0.5 ( Hair et al., 2014). There was discriminant validity in our instrument based on Fornell and Larcker (1981) criterion that when the correlation values among the latent variables are less than the square root of AVE, then there is discriminant validity (see Table 2). Finally, the values of Skewness and Kurtosis did not indicate any serious violations of normality, as all the coefficients were below ±2 (see Table 3).

To explore the association between role modeling, meal planning, and snack choice behavior, we estimated the Pearson correlation coefficient, which indicated a significant correlation between the variables (r = -0.02, p > 0.05) (see Table 2).

We carried out Structural Equation Modeling (SEM) to evaluate whether role modeling predicted food choices. We used SEM as it provides for the control of the measurement error.

Table 4 indicates that the relationship between role modeling and snack food choice is not significant (β = - 0.138, t-value = - 0.973, p-value = 0.331); the relationship between role modeling and meal planning is significant (β = 0.785, t-value = 5.455, p-value = 0.000) and the relationship between meal planning and snack food choice is significant (β = 0.401, t-value = 3.696, p-value = 0.00). The squared Multiple Correlation is 0.104 (See Figure 1). This means role modeling and meal planning, approximately, explain 89.6 percent of the variance in Snack food Choice. In other words, the error variance of role modeling and meal planning is approximately 89.6 percent of the variance of snack food choices.

We employed a bootstrap procedure to test for mediation effects and determine the level of significance of the mediation effects. The bootstrap approach simultaneously tests for both the direct and indirect relationships in a single model. Table 5 depicts the mediation results.

The bootstrap result revealed a significant full mediation (β = 0.315, p = 0.000). This means that changes in role modeling positively affect variations in Meal Planning, which in turn cause changes in Snack Choice. This implies that when parents, friends, television cooking shows and celebrities influence the youth, they go ahead and plan for where they shall have their meals, what meals they shall eat, and with whom they shall eat. This meal scheduling consequently leads to the youth consuming more fruits and more snacks.

This study examined the mediating role of meal planning in the relationship between role modeling and snack choice. The study’s results revealed a full mediation. This suggests that modeling does not have a direct impact on food consumption behavior. Instead, meal planning serves as a pathway through which modeling influences food consumption behavior. This implies that when parents encourage the youth to eat certain snacks it does not directly translate into the youth consuming those foods. This was a surprising finding. The plausible explanation is that at this life stage, the youth have started making autonomous consumption decisions. Parents no longer have as much control as when they were younger. So, even when the parents urge the youth to eat healthy the youth may not follow the advice. Further to this is that these youths largely reside alone in off-campus residences. Parents are rarely eating companions of the youth during their stay at university. Even when friends encourage the youth to eat snack foods It may not directly lead to consumption. this may be due to the familiarity of eating different foods than those suggested by friends. This familiarity with eating these foods may be caused by the cultural upbringing of the youth that suggests they eat other foods. This finding contradicts Roy et al. (2021) assertion that advice from friends and family significantly influences healthy eating choices. The finding also means that even when television celebrities endorse healthy foods, it does not directly influence the youth’s eating habits. Perhaps the youth consider this food quite expensive and therefore do not purchase it. This finding contradicts Folkvord et al. (2020) assertion that watching cooking programs that promoted healthy foods was highly likely to lead to healthy food choices.

The findings, additionally, indicate that when the youth, in Uganda, watch television celebrities, that endorse healthy foods, the youth make detailed eating prearrangements. This may be due to the youth’s admiration of the celebrity’s accomplishments like body appearance and fame. The youth will hence desire to uphold the celebrity’s eating values and beliefs. These meal plans may then aid the youth achieve their goals by developing an eating routine given their busy study schedules. This finding agrees with Ngqangashe and De Backer’s (2021) conclusion that showing videos depicting fruits and vegetables to adolescents resulted in them preparing to eat the demonstrated foods.

The study findings also indicated that youths, in Uganda, who prepare what to eat in advance eat the foods that they prearranged. In other words, whenever a youth makes detailed plans regarding, say, eating fruits then s/he consumes fruits at the slated time. This could result from time restrictions that may hinder youth from eating during study periods. Students may, also, find eating fruits a convenient option since they lack enough time to cook. The findings conform to Domke et al. (2021) whose report indicated that German adults who planned their meals on precisely when, where, what, and how they ate consumed more fruits and vegetables. Therefore, when young people admire their role models’ eating habits it leads them to plan their meals and choose similar snacks.

The study indicates that role modeling is not an immediate antecedent of food consumption behaviors of the youth in Uganda. From a Social Cognitive Theory ( Bandura, 1977) perspective, it may be due to the lack of physical proximity to the role model. The youth largely lived alone mainly away from their parents and hence little physical interaction at meal times. This meant that there was little cognitive rehearsal or reminders from parents regarding the eating of healthy foods. The youths were largely unmarried which meant that no one was around during meal times to remind youths to eat these healthy foods. This is because spouses are known to encourage each other to eat healthy ( Kabwama et al., 2019). That is, as people who may be living together, they remind themselves (cognitive rehearsal) of what to eat. Surprisingly, one would expect a youth to eat the same food as the food that the television cooking shows or celebrities promote but this, based on the current findings, was not the case. This may be explained by the fact that according to the SCT, there needs to be proximity to the role models. This proximity means that the youth identify with the celebrity, which may have been lacking. Even constant reminding (cognitive rehearsal) is necessary for behavior to be copied from the model. These cooking shows may be rarely aired on television.

The other reason could be that even when the youth learn from screen celebrities regarding eating particular foods, the youth’s demonstration of the learnt behavior may not be immediate. The display of this learnt behavior could happen later when a different drive (e.g., financial capacity, health triggers etc.) makes it possible. Given that our study was cross-sectional, capturing the association between learning and acting on the learnt behavior (that may be displayed eventually) may not have been possible. Therefore, the fact that Bandura developed the theory using an experiment that involved children could strengthen this argument as these children immediately reproduced similar behaviors. In our case, the youth may be capable of learning an eating behavior from the role model but then delay acting on the learnt behavior.

The findings, also, lend support to the Health Action Process Approach ( Schwarzer, 2008), which asserts that planning serves as an operative mediator between intentions and behaviour. This view is based on the fact that when an individual admires a person’s diet, he makes an intention to consume the same food as the role model at a later time, which leads to scheduling that food ultimately leading to the consumption of that food.

At a policy level, governments should carry out mass awareness campaigns to encourage the youth to schedule what and when they are to eat their meals. The ministry could perhaps implement this campaign using youth opinion leaders that include movie/television celebrities. This prearrangement of meals would enable the youth to manage their busy work schedules and create time for the consumption of healthy meals.

The present study is not without limitations but in turn, offers the opportunity for future studies. The results of this study are an outcome of a case study of young adults in Uganda. There is a need for further extension of these studies to other countries and populations before researchers generalize the findings. The results of this study opened a way for assessing the mediating role of meal planning in the relationship between role modeling and snack consumption among young adults in Uganda. Therefore, future studies should consider studying the effect that physical or virtual proximity of the role model has on snack choices among young people in other contexts.

The study was cross-sectional in nature. Capturing the long-run effect of role modeling on food consumption was not possible. Given that learning from a role model may happen at an earlier time and display of the learned behavior at a far later time, it calls for a longitudinal study to ascertain this relationship in future studies.

In conclusion, the environment in which young people, in Uganda, grow up greatly influences their eating habits. They observe and admire the eating behaviors of their parents, friends, or even celebrities. This admiration leads to the pre-planning of their meals, deciding what to eat and when to eat it. As a result, young people, in Uganda, often choose to consume foods that include fruits and dairy products, which are like the foods eaten by their role models.