A systematic review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (46).

To be eligible, studies had to be available in English and explore adolescents' (13–18 years old) views on what impacted their dietary habits using qualitative methods. For the purpose of this review, dietary habits were defined as the quantity/quality of foods eaten or meal-related aspects. When age was not available, studies using 7th to 12th graders were eligible, as this range often coincides with adolescence. Studies that included younger (11–12 years) or older youth (19 years) were eligible if the study defined its population as “adolescents,” reported separate results, or had a substantial proportion of participants who were 13–18 years old. Adults could have also been participants as long as the results were presented separately. Studies had to utilize a mixed-gender sample (i.e., boys and girls) and report at least one unique theme by gender. In the absence of gender, biological sex was used. Studies conducted on adolescents living with a diagnosed psychological disorder (e.g., depression and eating disorders), pregnancy, severe dietary restriction (intake of two or more food groups), or allergies were not eligible. Dissertations, reviews, opinions, and editorials were excluded but hand-checked by AD.

This study was registered prior to search syntax formation with the International Prospective Register of Systematic Reviews (PROSPERO ID: CRD42022298077). All returned studies from the search in each database were uploaded into Covidence (47) where three trained research students screened the initial titles and abstracts. AD and one additional study team member had to select each study as eligible for retention at this stage and during full-text review. Once retained from the full-text review, data were extracted by a research team member and checked for accuracy by AD.

The search syntax for each database is found in Table 1. It was developed and pilot tested by the first author (AD) in consultation with a trained librarian.

MEDLINE Ovid, Embase, Web of Science, and the Cumulative Index to Nursing and Allied Health Literature (CINHAL) were searched from conception to 17 December 2021. To be considered, studies had to discuss adolescents' motivations behind their food habits.

The three research assistants extracted study data (e.g., title, country, aims, data collection methodology, sample characteristics, and main themes). The fidelity of all extracted data was checked by AD.

A thematic analysis was led by AD in conjunction with CB (48). The two iteratively derived the codebook by reviewing all extracted themes and inductively grouping themes together (49). Discrepancies were resolved with TC. Themes could have been stated uniquely among boys in some studies, girls in others, or as non-gender unique (i.e., said by boys and girls in a study). To interpret this, this review analyzed differences in the themes uniquely reported by boys or girls in any of the 34 studies. A sub-analysis of themes across countries was also conducted by AD, and trends were considered by country.

The quality and potential for bias of all included studies were assessed using the Critical Appraisal Skills Programme (CASP) checklist (50, 51). This appraisal tool consists of 10 questions. In the appraisal, one research assistant and CB independently reviewed studies and chose “yes,” “no,” or “somewhat” to each question. Discrepancies between the two were resolved by AD. As the CASP program does not have a formal scoring system, this study created one (yes = 2, somewhat = 1, and no = 0). Calculated scores were divided by the maximum score of 20 to obtain a percent out of 100. Values > 70.0% were considered high quality, 40.0–60.0% as moderate, and <40.0% as low (52).

Six studies (17.6%) were conducted in the United States (53–58). Three were conducted in China (8.8%) (59–61). Australia (62, 63), Costa Rica (64, 65), Bangladesh (66, 67), Ireland (68, 69), and Sweden (70, 71) all had two studies (5.9%) each. Ethiopia (72), India (73), Indonesia (74), Iran (75), Morocco (76), Netherlands (77), Peru (78), Scotland (79), Slovakia (80), South Africa (81), United Arab Emirates (82), and United Kingdom (83) all had one study (2.9%) each. One study considered data from Canada and India, these countries were evaluated separately (84). Sub-analyses were grouped by continent and income level classified by the World Bank (85) as there were too few countries to draw country-level trends.

Sample sizes ranged from 12 to 141 participants, with the majority disclosing how many boys and girls participated (79.9%). Of the studies disclosing how many boys and girls participated, 55.2% included more girls than boys (55–59, 62, 68, 70–74, 77, 83, 84, 86) and 24.0% included an equal number of boys and girls participated (54, 60, 63, 67, 76, 79, 82). One study did not indicate specific participant numbers or genders (75).

Four studies reported only participants grades (11.8%) (53, 56, 60, 82). All other studies reported age ranges from 8 to 19 years (25, 54, 55, 58, 59, 61–66, 68–73, 76–81, 83, 84, 86). Of these, 14 studies (46.7%) included ages below 13 years (25, 58, 59, 62–64, 68–70, 72, 80, 81, 83, 84), one study included adolescents over 18 years, and 4 studies (13.0%) included older and younger ages (57, 67, 74, 75).

The majority of studies (67.7%) used focus groups (53, 55–60, 62, 64–66, 68–73, 75–77, 81, 83, 84), with 30.4% of these conducting mixed-gender groups (55, 56, 58, 60, 70, 71, 77). Parents were included in 26.1% of studies (53, 65, 66, 72, 73, 76, 81), but only present with adolescents during data collection in one study (53). Semi-structured interviews (26.5%) (25, 54, 61, 63, 67, 78, 79, 82, 86), open ended text (2.9%) (80), and immersed field observation (2.9%) (74) were also used. Less than half of the studies reported the theoretical lens used to interpret the collected data (41.2%) with a Socio-Ecological Model used most (20.6%) (25, 53, 58, 61, 64, 76, 78).

Quality appraisal scores ranged from 45 to 85%, with a mean score of 66%. Thirteen studies (38%) ranked as high quality (scores ≥ 70%) (56, 58, 59, 62, 66, 70–72, 74, 78–81), while all other studies ranked as moderate. Most studies did not report details on researcher relationships.

Two overarching themes dictated adolescents' dietary habits, including what, when, or how much to eat, regardless of gender. These two themes included Self-motivators and Uncontrollable factors. Self-motivators spoke to adolescents identified intrinsic and extrinsic motivators that they perceived some level of control over. Six sub-themes (Manage health, Enjoyment, Fit in with peers, Manage performance, Control appearance, and Form a food identity) arose. Adolescents also recognized Uncontrollable factors, factors they perceived they had no control over that limited their dietary habits. This theme included two sub-themes: Cognitive control and Logistics. A summary and explanation of these themes are found in Table 2.

In studies reporting non–gender-unique influences (i.e., said by boys and girls in the same study), Logistics was found to be the largest factor impacting dietary habits, accounting for 28.0% of the stated themes in the 34 studies. This was followed by Enjoyment (20.0%), Form a food identity (19.0%), Manage health (16.0%), Fit in with peers (9.0%), Manage performance (4.0%), Control appearance (2.0%), and Cognitive control (2.0%). This hierarchy was not the same among boys and girls when it came to unique gender themes stated in individual studies (see Figure 2).

Boys and girls both uniquely identified Fit in with peers the most (26.0% of themes in each group). Among boys, this was closely followed by Enjoyment (19.0%), Mange performance (17.0%), Manage health (14.0%), Control appearance (12.0%), Form a food identity (10.0%), and Logistics (2.0%). Cognitive control was not uniquely stated by boys in any of the 34 studies. This differed from girls, where Control appearance was the second most stated motivator (24.0%) followed by Manage health (20.0%), Enjoyment (13.0%), Form a food identity (7.0%), Cognitive control (4.0%), Logistics (4.0%), and Manage performance (2.0%).

Unique gender differences in stated sub-themes are discussed below, with an exploration by country. A summary of unique gender sub-themes by country is found in Table 3.

Though boys and girls both highlighted the importance of Fitting in with peers, they did so in different ways, a finding that is similar to other reviews (16, 88). In this study, we found that girls discussed how peers contributed to eating healthier and how they altered their dietary habits in front of boys to appear attractive. The boys did not mention these influences. It is possible that girls are more aware of gender connotations surrounding certain foods (e.g., healthy food as being more feminine), and thus consume them to alter social perceptions more so than boys (89). Yet, only three studies in this review saw girls explicitly acknowledge gender norms surrounding food, and furthermore, boys in this review also identified this (n = 4 studies). This may suggest instead that adolescent girls' dietary motivation is less about appearing feminine and more about fitting in socially compared to boys, which matches adult literature where women express greater importance on aligning with social norms compared to men (7). However, girls' emphasis on Concern for health (n = 11 studies) or Food quality in the context of health matches historical gender norms in other literature (4, 34, 90, 91). Taken together with the motivation to fit in by eating healthier with peers, it seems girls may actually still be attempting to conform to a version of femininity, it is just not acknowledged overtly (3). Gender literature has called for more public awareness of how gender norms still influence many behavioral facets, such as careers and parenting roles (3, 4), and it appears dietary habits are no exception. Future research needs to untangle the extent to which gender norms covertly impact boys' and girls' motivation behind their dietary habits (e.g., are girls more concerned about health or do they just say this because they feel they “should”?) to address how gender norms may be controlling eating (3, 4).

Overall, boys perceived self-motivators that were more intrinsically focused than girls, a finding that aligns with adult literature (8). For example, boys' desire to grow, manage physical performance, eat foods they personally enjoy, and form a food identity that is autonomous suggests more intrinsic motivation. These contrast girls wanting to be thin, managing aesthetics, and altering dietary habits in front of peers/boys to be accepted. Though the trends in body shape found in this review are not original (27, 29, 92, 93), our study contributes to understanding the larger picture of adolescents' motivation behind eating based on gender. Specifically, we call on researchers who create dietary interventions targeting adolescents to consider focusing on using language and content that speaks to gender-diverse motivators, including body shape, fitting in, and pursuing autonomy (e.g., focus on autonomy and personal control among boys vs. highlighting social relationships and pressures among girls) (6, 7, 10, 23, 88). We also call on research to clarify the role of athletic performance in adolescents' dietary choices, as this review only identified this influence among boys. Other literature has suggested that dietary impacts among boys and girls vary based on the level of competition (21, 94, 95), whereas literature among adults has suggested that men may be more encouraged to alter their diet in regard to physical activity in less competitive contexts (8). Confirming the relationship between physical activity, competition level, and diet will help inform if interventions seeking to alter these health behaviors should combine them or keep them separate, and for whom.

Boys and girls did not differ in their unique perception of the role of logistics outside of their control. This is somewhat surprising given that literature suggests boys are more often granted more behavioral autonomy compared to girls (3, 43, 96). The findings of this review may thus indicate that parent's food-related activities (e.g., rules, what is brought into the home) may not be based on their child's gender as much as previously thought.

Our findings that only girls identified Cognitive eating (e.g., emotional eating) as an uncontrollable motivator may suggest that biological differences play a significant role. Research examining appetite (97, 98) and coping strategies for stress (99, 100) has found sex differences, largely through hormones. These differences in hormones can impact eating through the gut–brain axis, having a role in intentional eating for cognitive aspects compared to hunger (101). A recent cross-sectional study exploring mothers' and fathers' parenting practices on boys' and girls' dietary habits further supports some role for differences in cognitive factors, as it found that adolescents' dietary behaviors were dependent on their gender only through differences in their motivation and self-efficacy (28). Additional research in adults has found that attitudes toward emotional eating vary between women and men, with women expressing more enactment (5). Future sex and gender-based analyses are needed to establish if gender and social norms, cognitive or hormonal factors rooted in biology (or both), ultimately guide dietary habits. This knowledge could inform future dietary interventions focused on health promotion through diet (10, 23).

Major trends across countries themselves did not arise. This is likely because a limited number of countries were present in this review, and even less had multiple studies conducted. This matches data from a single study exploring the underlying motivation behind adults' dietary habits, which found prevailing trends across the 23 included countries between men and women (34). However, across continents, some small trends did emerge and may reflect trends based on a country's overall income level.

Studies from Africa and East Asia in this review did not uniquely identify any influences related to Forming a food identity, Managing performance, or Fitting in with peers. This may suggest that boys and girls do not experience pressures to be different (e.g., autonomy among boys) or match a group (e.g., match peers) based on typical gender trends seen in other countries. It is not clear why, but it could be attributed to differences in cultural norms across regions. Emotional eating also arose only unique among girls in high-income countries in Europe and North America. This could suggest that eating based on emotions is tied to aspects of socio-economic status, such as greater food availability or marketing. More studies and an exploration of dietary habits across diverse countries are needed to solidify our understanding of these concepts.

Many studies in this review focused on eating intentions, and this could have created a gap in understanding adolescents' practiced habits. Furthermore, most studies were from high-income white nations. The findings could overrepresent the lens of these groups. They may also underrepresent the influences that impact boys' dietary habits, as majority of studies recruited more girls than boys. Most studies (76.7%) relied on focus groups, which are subject to responder bias. Finally, literature thus far often has relied on dichotomized sex (e.g., male/female), leaving out the experiences of gender-diverse folks (102). Due to this, we are unable to explore any trends in motivation and dietary habits among gender-diverse adolescents.