Skip to main content

Table 1 Study characteristics

From: Determinants of fruit and vegetable consumption among children and adolescents: a review of the literature. Part II: qualitative studies

Study by first author

[Ref. ID]

Country setting

Pheno-menon of interest (outcome)

Sampling and participants

Child socio-demographics:

A: sex. B: age.

C: school year. D: SEP. E: ethnic background.

Data collection methods and no. of FGs or interviews

Theoretical framework

Analytical method/approach

Main topics related to FV intake

Baranow-ski et al. 1993 [29]

US

FV intake

School-based (1 school): 235 schoolchildren, 15 parents, 8 teachers, 4 school food service workers.

A: no info. B: no info. C: 4 & 5. D: pre-dominantly lower SEP. E: more than 50% Afro-Americans and the rest mostly Anglo-Americans.

FG discussions: 5 year 4 schoolchild FGs, 5 year 5 schoolchild FGs, 2 parent FGs, 2 teacher FGs, 1 school food service worker FG.

Social cognitive theory: reciprocal determinism.

No clear description of analytical procedures. Theory-based interpretation of data. Results are categorised by aspects of reciprocal determinism.

Home and school FV availability, access to unhealthy food in school, sensory attributes (taste, appeal, appearance, smell, mouth feel), methods of preparation, preferences/liking, outcome expectancies, acceptance of national recommendations, food categorisation, preparation skills.

Bauer et al. 2004 [57]

US

Healthy nutrition (and physical activity)

School-based (2 schools): 26 schoolchildren and 23 faculty and staff members.

A: mixed. B: no info. C: 7 & 8. D (school level): mixed composition. E: mixed composition (80% White, 20% either Asian- or African- Americans).

FG discussions: 7 grade- and gender-homogeneous schoolchild groups, 3 faculty and staff member groups and 10 individual interviews with key informants (e.g. school nurse, cafeteria manager, administratives).

Ecological models by Bronfenbrenner (1979), Stokols (1996), Story, Neumark-Stzainer & French (2002).

Grounded theory: 1st step: systematic coding of themes, 2nd step: identification of 3 mechanisms of influences on eating within the school environment based on data-developed concepts and theoretical framework.

School FV availability (quantity, variety, quality), School availability of unhealthy competitive food choices.

Booth et al. 2008 [49]

Australia

Healthy food (Perceived causes of overweight and obesity)

School-based (3 secondary schools): 58 schoolchildren.

A: mixed, B: 12-17. C: 7-11. D (area level): areas selected to reflect a wide range of SEP differences. E: No info.

9 gender- and school year-homogeneous FGs (year 7+8, year 9+10, year 11).

No info.

No clear description of analytical procedures. Coding of themes.

School FV availability (price, quality, presentation).

Campbell 2009 [30]

US

Dietary choices

School-based (1 school): 12 schoolchildren.

A: mixed. B: 14-16. C: 9 & 10. D: pre-dominantly low-income families. E: mixed (Hispanic, African-American, Eurasian and combination of these).

FG discussions: one group interviewed twice, during lunch and immediately after.

Developmental psychology by Piaget and Erikson.

Content analysis, but no clear description of analytical procedures.

Home and school FV availability, parental influence, availability, liking, methods of preparation, knowledge, food categorisation.

Cullen et al. 1998 [25]

US

FV con-sumption

Community-based: 99 urban boy scouts and 39 parents.

A: boys. B: 10-14. C: elementary school. D: no info. E: mainly African-American (88%).

13 FGs with boy scouts and

7 FGs with parents.

Social cognitive theory concept of reciprocal determinism.

No clear description of analytical procedures. Transcripts were coded and quantified.

Preferences, outcome expectancies, sensory attributes (taste, mouth feel), snack food purchases, price, parental-, peer-, and media influence, preparation skills, home accessibility, school availability.

Cullen et al. 2000 [48]

US

FJV intake

School-based (6 schools): 180 schoolchildren and 40 parents.

A: no info. B: 9-12. C: 4-6. D: mixed. E: mixed: 3 Afro-American schools, 1 Euro-American school, 2 Mexican-American schools.

School year and ethnically homogeneous FG discussions: 6 African-American schoolchild FGs, 6 Euro-American schoolchild FGs, 5 Mexican-American schoolchild FGs, and 8 Parent FGs.

Social cognitive theory: reciprocal determinism.

Data-based analysis. Systematic coding of transcripts and comparisons of results by ethnicity. Data-based variable names assigned to text passages.

Home availability/accessibility (variety), peer-, parental-, and media influence, sensory attributes (taste), food categorisation.

Cullen et al. 2007 [26]

US

School food

School-based (6 schools): schoolchildren, school staff and district school food administrators (no. of participants not provided).

A: no info. B: 11-14. C: middle school. D (school-level): at least 50% of schoolchild population received free or reduced price meals. E (school-level): at least 50% of schoolchild population was African- American and Hispanic.

11 FGs with schoolchildren/school staff. Interviews with 7 district school food administrators.

No info.

No clear description of analytical procedures.

School V availability (variety, freshness).

Evans et al. 2006 [40]

US

Healthful eating

48 adolescents from two middle schools and one recreation & parks centre.

A: mixed. B: 10-14. C: 6 & 7. D: low income. E: Mainly Black (81%).

3 male and 2 female FGs.

Social cognitive theory

Systematic analysis based on pre-specified coding scheme (categorisation of data according to gender, location, and motivational theme) and standardised procedures.

Nutritional knowledge/outcome expectations (misconceptions), home availability (access to competitive food choices), peer pressure/symbolic value of food, school availability (appearance, appeal, freshness), FV availability at restaurants. Gender differences.

Fitzgerald et al. 2009 [41]

Australia

Eating behaviour (and physical activity)

School-based (1 school): 37 schoolchildren.

A: mixed. B: no info. C: kindergarten & year 1-6. D: low SEP community. E: no info.

3 FGs: kindergarten + year 1-2, year 3-4, year 5-6.

The socio-ecological approach is cited in the introduction.

Open coding/thematic analysis of transcripts.

Outcome expectancies, adult-, peer- and media influence, symbolic value of food, sensory attributes (taste), convenience, access to unhealthy food in local area, time limitations.

Gellar et al. 2007 [31]

US

Healthy eating

140 youth from diabetes camp.

A: mixed. B: 7-16 (mean age: 11.8). C: no info. D: mixed. E: mixed (71% white, 18% Black, 6% Hispanic).

12 female and 6 male FGs (almost similar age).

No info.

Content analysis: Systematic coding of transcripts using a pre-specified coding system.

Preferences, sensory attributes (taste), knowledge, outcome expectancies, school FV availability (appeal, methods of preparation/form, competitive unhealthy food choices), convenience, home availability of unhealthy food, peer- and parental influence.

Goh et al. 2009 [53]

US

Healthy eating (and physical activity)

School-based (2 schools): 119 schoolchildren, 63 parents, and 28 key stakeholders.

A: mixed. B: mean age: 12. C: 7 & 8. D: no info. E: mixed (58% Latino).

6 male and 8 female schoolchild FGs, 8 parent FGs, interviews with 28 key stakeholders.

No info.

Systematic content analysis.

School availability (accessibility, appearance, methods of preparation, visibility, braces-friendly FV, unhealthy food), knowledge, parental influence.

Hill et al. 1998 [32]

New Zealand

FV con-sumption

Community-based: 20 teenagers and their parent.

A: mixed. B: 13-16. C: no info. D: mixed. E: Pakeha (European ancestry).

20 interviews: Separate interviews with teenager and parent responsible for food preparation.

No info.

Cross-household analysis. The analysis focus on interaction between teenager and parent-shopper in each household. No clear description of analytical procedures.

Situational norms, convenience, FV preparation skills, FV availability at home, school, and in local area (appeal, quality, parental facilitation, price, variety), peer-, parental- and media influence, preferences, outcome expectancies, knowledge. Age and gender differences.

Husby et al. 2008 [51]

Denmark

Meals & snack consumption

Children with a healthy diet (N = 9) and a less healthy diet (N = 8) were recruited through a dietary survey among their parents.

A: mixed. B: 10-11. C: no info. D: mixed. E: no info.

17 photo-elicited, semi-structured individual interviews.

Meals are examined as social events. Meals involve the establishment and re-establishment of the family unit.

Template analysis (pre-specified themes) and comparative analysis.

Peer influence (food swapping), snack, outcome expectancies, FV preparation skills, parental facilitation, food rules. Gender differences.

Keim et al. 2001 [33]

US

FV intake

Community-based: 27 Caucasian and 30 Mexican-American healthy, low income children from public school, migrant worker summer schools and community centres.

A: mixed. B: 8-11. C: 3. D: low income. E: Caucasian- and Mexican-American.

FG discussions: 4 FGs of Caucasian children and 6 FGs of Mexican-American children.

Social cognitive theory

Transcripts analysed and coded within the context of Social cognitive theory.

Parental facilitation, FV preparation skills, FV shopping, price, home availability/accessibility (visibility, convenience, variety, unhealthy food), parental- and peer influence, preferences, sensory attributes (taste, mouth feel, appearance, quality, freshness, methods of preparation, familiarity), outcome expectancies, knowledge. Ethnic differences.

Khunti et al. 2008 [50]

UK

Healthy lifestyle

School-based: Pupils (no. not provided but can be estimated to maximum 144) and school staff.

A: mixed. B: 11-15. C: 7-10. D: schools located in a very deprived area. E: mixed: In the overall sample 77% of the pupils were of South Asian origin.

Action research approach. Baseline: 18 schoolchild- and 5 staff FGs. Follow-up: 8 schoolchild- and 5 staff FGs. Observational visits at all schools.

No info.

Open coding (in line with the 1st analytical step of grounded theory) of data. A process of progressive focussing is used to develop a thematic framework.

Peer influence (image), cost & risk of wasting money, hunger satisfaction.

Kim et al. 2007 [42]

US

Dietary practices/FV intake (and physical activity)

Community-based: Low-income Hmong American parents (N = 44) and youth (N = 40). Key informants (N = 5) in Hmong communities.

A: mixed. B: 11-14. C: no info. D: low-income. E: Hmong Americans.

8 FGs with adults and youths and 5 individual interviews with key informants.

No info.

No clear description of analytical procedures. The transcripts were coded and organised.

Outcome expectancies, knowledge, preferences, parental influence, sensory attributes (smell, freshness), time/occasions for eating FV, school availability.

Kirby et al. 1995 [34]

US

FV intake

School-based (6 schools from 3 regions): 398 schoolchildren, 108 parents, 43 teachers, 29 school food service workers.

A: no info. B: no info. C: 4-5. D (region level): mixed. E: mixed. 2 schools of predominantly white. Caucasian ethnic composition (1 high and 1 middle SEP) and 4 schools of non-white (African-, Asian-, Hispanic American, other or multi-ethnic) composition (2 Low and 2 very low SEP).

FG discussions: 15 schoolchild- (school year homogeneous), 11 parent-, 6 teacher- and 6 Food service worker-FGs.

Social learning theory: reciprocal determinism.

Systematic, theory-guided coding of transcripts. The assigned variable names were developed based on the discussion guide and theoretical framework.

Home availability/accessibility (variety, parental facilitation), preparation skills, price, preferences (variety liked), sensory attributes (taste, mouth feel), food categorisation, knowledge, convenience, methods of preparation, outcome expectancies, time/occasions for eating FV (restaurants), peer influence, availability in local area/restaurants. SEP differences.

Kubik et al. 2005 [35]

US

Dietary practice (and physical activity)

School-based (7 Alternative High Schools): 70 schoolchildren.

A: mixed. B: no info. C: 9-12. D: (school-level): mixed composition: 46% of schoolchildren qualified for free reduced lunch program. E: mixed composition: 36% of schoolchildren were of non-Caucasian origin (American-Indian, African-American, Hispanic, Asian).

7 schoolchild FGs.

Ecological theory and social learning theory.

Systematic 3-step analytical process as described by Miles & Huberman (1994).

Convenience, home and school FV availability/accessibility, access to unhealthy competitive food in school and local area, price, quality, preferences, cooking skills.

Lauten-schlager & Smith 2007 [36]

US

Dietary behaviour (values, beliefs and gardening & cooking behaviours)

Community-based: 40 inner-city youth. Two subgroups: involved in Youth Farm Garden Program (N = 26) and not involved (N = 14).

A: mixed. B: 9-15. C: no info. D: no info. E: mixed: white (15%), African-American (30%), Hispanic (17%), Asian (27%), Somali (7%), other or multiracial (14%).

6 FGs: 3 with garden program participants and 3 with youth not involved in garden program.

Theory of planned behaviour

Application of systematic, content analysis procedures by Miles & Huberman (1994).

Sensory attributes (flavour/taste, mouth feel/texture, appearance) convenience, preferences, method of preparation, outcome expectancies, knowledge, availability in the neighbourhood (seasonality, quality, quantity, supply), parental-, peer- and media influence.

Libman 2007 [43]

US

Food conscious-ness and V eating habits

Community-based: 10 schoolchildren from a children garden program, four mothers and one father.

A: mixed. B: 10-14. C: no info. D: no info. E: African-American, Puerto Rican, Dominican, and Guyanese.

Schoolchildren: 1 FG, 10 semi-structured seated interviews, 6 walking interviews (youth-led garden tours). 5 parent telephone interviews. Observations of program and material.

Developmental psychology by Lev Vygotsky.

Systematic coding for themes relevant to research questions (Miles & Huberman 1994).

Sensory attribute (taste, methods of preparation), cooking skills, food consciousness/knowledge, home FV availability (appearance, freshness, safety of organic FV).

McKinley et al. 2005 [44]

England and Northern Ireland

Healthy eating

School-based (11 schools):106 schoolchildren.

A: Mixed, B: 11-12. C: 1st year of post-primary school. D (school level): mixed SEP backgrounds. E: mixed ethnic backgrounds: White Europeans (76%), Asian (18%), Afro-Caribbean (6%).

11 FGs (2 discussion sessions per group). 4 of the FGs were gender-homogeneous as they were conducted at single-sex school.

No info.

Systematic coding of transcripts using the cut-and-paste technique described by Stewart & Shamdasini (1990).

Food categorisation, school availability (appearance, quality), sensory attributes (texture, mouth feel), convenience & time costs, cost & taste guarantee, cost & filling power, rebellion. Gender differences.

Molaison et al. 2005 [37]

US

FV intake

Community-based: 42 southern, low-income black American adolescents recruited from National Youth Sport Program.

A: Mixed. B: 10-13. C: no info. D: low income. E: Black Americans.

6 gender- and age-homogeneous FGs.

Social cognitive theory

Theory guided the analysis. Transcripts were coded by content analysis methods and codes/themes were assigned to the theoretical framework.

Sensory attribute (taste, method of preparation, form (canned vs. fresh)), allergies, preferences, variety (vegetable boredom), outcome expectancies, food preparation skills, home and neighbourhood availability, appropriate settings for FV, family- and peer influence, self-efficacy. Gender differences.

Monge-Rojas et al. 2005 [28]

Costa Rica

Healthful eating

School-based (3 schools): 108 schoolchildren.

A: mixed. B: 12-18. C: 7-11. D (school-level): mixed (2 public high schools and 1 private high school). E: Costa Rican.

12 gender- and age- homogeneous FGs (3 sessions per group).

Conceptual model for adolescent eating behaviours based on Social cognitive theory and ecological perspective proposed by Story et al. (2002).

The transcripts were reviewed systematically for emerging themes. Themes were identified according to the theoretical framework.

Knowledge, school availability of FV and unhealthy food, home availability, parental facilitation, peer influence/norms (gender roles, symbolic value of food), cost & satiety value, sensory attributes (taste, methods of preparation), convenience & time considerations, outcome expectations, parental- and media influence. Gender differences.

Neumark-Stzainer et al. 1999 [45]

US

Food choices and eating behaviours

School-based (2 schools): 141 schoolchildren.

A: mixed. B: 12-14 (mean age: 12.6) & 15-19 (mean age: 16.0). C: 7 & 10. D: no info. E: mixed composition: white (40%), Asian-American (25%), African-American (21%), multiracial (7%), Hispanic (6%), Native American (1%).

21 age- and gender-homogeneous FGs.

Social cognitive theory is included in the discussion.

Systematic analytical approach using the constant comparative method of grounded theory.

Sensory attributes (taste, appeal, appearance, methods of preparation), convenience & time considerations, hunger satisfaction & costs, availability at home, school and fast food restaurants (visibility, accessibility). Gender differences.

Nicklas et al. 1997 [27]

US

FV intake

School-based (4 schools): 55 high school schoolchildren.

A: mixed. B: no info. C: 9. D: no info. E: mixed. Participants drawn from a student cohort of mainly Caucasian background (79%). The rest are of Hispanic, African-American, Asian or Native American origin.

4 FGs (white male, white female, black male, black female) - unclear if FGs mix schoolchildren from different schools.

The intervention is based on the PRECEDE model of health education (6 levels of behaviour change).

No clear description of analytical procedures.

Outcome expectancies, sensory attributes (taste), inconsistency in taste, home & school FV availability (visibility, variety, presentation/appearance), cost, access to competitive unhealthy food in school and local area.

O'Dea 2003 [52]

Australia

Healthful food (and physical activity)

School-based (34 schools): 213 schoolchildren and 38 school principals.

A: mixed. B: 7-17. C: 2-11. D & E: a representative mix of SEP and ethnicity.

38 FGs.

Theory of planned behaviour and social learning theory.

Content analysis (Miles & Huberman), systematic approach.

Outcome expectancies, food categorisation, knowledge, home availability (unhealthy competitive food choices), convenience & time costs.

Ross 1995 [46]

Scotland

Food choices and preferences

School-based (one school): 46 schoolchildren.

A: mixed. B: 10-12 (mean age: 11). C: primary 7 year. D (area): School situated in catchment area encompassing all SEP groups. E: schoolchildren were predominantly white (only a few from ethnic minority backgrounds).

FG discussions: 2 male FGs, 3 female FGs and 2 mixed-gender FGs. Planned observations during lunch time were not feasible because of the fact that lunch occurred in several sites simultaneously and only one researcher being involved in the project.

No info.

Grounded theory approach

Sensory attributes (taste, texture), peer norms/influence (food swapping, socially acceptable food), affordability.

Steven-son et al. 2007 [47]

Ireland

Healthy eating

School-based (no info. on number of schools): 73 adolescents.

A: mixed. B: 12-15. C: second level schools. D: mixed. E: no info.

12 age- and gender-homogeneous FGs.

Socio-ecological approach.

Systematic coding of transcripts and deviant case analysis.

Sensory attributes (taste), parental influence.

Walker et al. 1973 [16]

US

FV intake

Schoolchildren and parents (primarily mothers). No. not provided, but can be estimated to maximum 220 participants.

A: mixed. B: 9-12 & 13-17. C: elementary and high school. D: middle and low income families. E: no info.

FG discussions (school year-gender-SEP homogeneous): 8 elementary schoolchild FGs (2 boy- & 2 girl-low income FGs and 2 boy- & 2 girl-middle income FGs), 8 high school student FGs, and 6 parent groups (3 low and 3 middle income FGs).

No information. Study conducted by social psychologists.

No clear description of analytical procedures.

Availability/exposure to FV at home and in local area (variety), price, parental style/attitude, preferences, sensory attributes (appearance, colour, texture, taste, odour, form, method of preparation), food prejudices.

Wind et al. 2005 [38]

The Netherlands and Belgium-Flanders

FV intake

School-based: 3 schools from the Netherlands, 60 schoolchildren. 32 schoolchildren from Belgium, no. of schools not provided.

A: mixed. B: 10-11. C: 5-6. D: no info. E: Netherlands: In two of the schools almost all children were from ethnic minority groups, in one school all except one child had both parents born in the Netherlands. Belgium-Flanders: 4 children had parents born in a foreign country.

FG discussion: Netherlands: 2 boy FGs, 1 girl FG, 5 mixed gender FGs. Belgium-Flanders: 1 boy FG, 1 girl FG, 2 mixed gender FGs.

Health belief model, theory of planned behaviour, social ecological models.

No clear description of analytical procedures. Determinants are analysed separately for fruit and vegetables. Determinants classified as personal, home- or school environmental factors.

Outcome expectancies, food categorisation, sensory attributes (taste, appearance, texture), preferences, knowledge/awareness, preparation skills, situational/social norms (time/settings for eating FV), convenience, home and school availability/accessibility (visibility, family rules, parental facilitation), unhealthy food shopping, peer-, parental and teacher influence. Ethnic and international differences.

Zeinstra et al. 2007 [39]

The Netherlands

FV preferences

School-based (1 school): Schoolchildren representing 3 different stages of cognitive development.

A: mixed. B: 4-5 (group A), 7-8 (group B) and 11-12 (group C). C: 1st, 4th and last school year of primary school. D: no info. E: no info.

4 + 4 duo interviews with group A and B and 4 FGs with group C.

Cognitive theory (Piaget).

Transcripts were coded systematically using a coding framework based on research aims, the interview guide and previous findings in the literature.

Preferences, sensory attributes (taste, texture, appearance, methods of preparation, familiarity, food categorisation), outcome expectancies, appropriate time and occasions for eating FV. Age differences.

  1. Abbreviations: FV = fruit and vegetables; FG = focus group; Info: information; FJV = Fruit, juice and vegetables; No. = number: Ref. ID = ID number of study in the reference list; SEP = socioeconomic position; US = the United States of America; V = vegetables; vs. = versus; UK = the United Kingdom.