Understanding gardening and dietary habits among youth garden program participants using the Theory of Planned Behavior

ARTICLE IN PRESS

Appetite 49 (2007) 122–130 www.elsevier.com/locate/appet

Research report

Understanding gardening and dietary habits among youth garden program participants using the Theory of Planned Behavior Lauren Lautenschlager, Chery Smith Department of Food Science and Nutrition, University of Minnesota, FScN 225, 1334 Eckles Ave, St. Paul, MN 55108, USA Received 6 March 2006; received in revised form 5 December 2006; accepted 2 January 2007

Abstract Sedentary lifestyles, along with diets low in fruits, vegetables, and complex carbohydrates, and high in fat and total energy are increasing among youth. These unhealthy behaviors contribute to an increase in childhood overweight, which is associated with type 2 diabetes, hypertension, and heart disease. Healthful dietary behaviors, such as eating a balanced and varied diet may be addressed in garden-based programs for youth. Therefore, this project assessed the influence of a garden program, with a newly developed nutrition curriculum, on youth’s eating and gardening behavior using the Theory of Planned Behavior. The model included the constructs of attitude, subjective norm, and perceived behavioral control (PBC). Youth (age 8–15 years) involved in a garden program in Minneapolis/ St. Paul, Minnesota completed a pre- (n ¼ 96) and a post-survey (n ¼ 66) assessing the theory’s constructs with regard to eating and gardening behaviors. Fruit and vegetable consumption were assessed using survey questions and a 24-h recall. In addition to finding gender differences regarding associations between intention and behavior and the constructs correlated with behavior, results indicated that attitude was most predictive of intention at both pre- and post-survey for both boys and girls with behavior associated to PBC in girls, but not for boys. A high level of intention for boys pre-survey marginally predicted some behavioral change post-survey, but girls with high levels of intention at pre-survey did not show positive behavioral changes at post-survey. Additionally, the garden program positively impacted youth fruit and vegetable consumption, as determined from a mean computed from the responses to the fruit and vegetable behavior survey questions and the 24-h recall food group data. Because youth in the garden program consumed more fruit and vegetables at post-survey compared to pre-survey, we conclude that garden programs may be a viable way to assist youth in making healthy lifestyle changes. r 2007 Elsevier Ltd. All rights reserved. Keywords: Gardening; Dietary habits; Fruits and vegetables; Ethnic food; Theory of Planned Behavior; Youth

Introduction The recent rise in childhood overweight (Troiano & Flegal, 1998; Troiano, Flegal, Kuczmarski, Campbell, & Johnson, 1995) has been related to unhealthy behaviors among youth, including low dietary intake of fruits, vegetables, and complex carbohydrates (ADA, 2003; Birch & Fisher, 1997; Mun˜oz, Krebs-Smith, Ballard-Barbash, & Cleveland, 1997), high intakes of fat and total energy (Birch & Fisher, 1997; Mun˜oz et al., 1997), and increased sedentary lifestyles (USDHHS, 2000). The association between an imbalance of food intake and energy expendiCorresponding author.

E-mail address: [email protected] (C. Smith). 0195-6663/$ - see front matter r 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.appet.2007.01.002

ture is well established (Nicklas & Johnson, 2004); therefore, promoting healthful diets and exercise among youth could mitigate childhood overweight and potential long-term consequences such as type 2 diabetes, hypertension, and heart disease (Ludwig & Ebbeling, 2001; Smith & Rinderknecht, 2003; Young, Dean, Flett, & Wood-Steiman, 2000). Healthful behaviors, such as balanced and varied diets, can be addressed in garden-based programs for youth by incorporating gardening, nutrition education, and cooking into their curriculum (BCBG, 2005; NOFA, 2005; YFMP, 2005). These programs also have the potential to instill positive health behaviors (Morris & Zidenberg-Cherr, 2002). Morris and Zidenberg-Cherr (2002) found that a garden-based program, in conjunction

ARTICLE IN PRESS L. Lautenschlager, C. Smith / Appetite 49 (2007) 122–130

with nutrition education, increased youth’s preferences for certain vegetables. However, little research has examined how these programs impact youth’s intentions to change eating and gardening behaviors. In order to change such behavior among youth, educators need a better understanding of factors associated with intentions for making behavioral change. The Theory of Planned Behavior (TPB) has been used to study decision-making behavior in a variety of healthrelated topics (Ajzen, 1991; Armitage & Conner, 2001; Bissonnette & Contento, 2001). TPB, a modified version of the Theory of Reasoned Action (TRA), is an expectancyvalue model that provides a framework for the determination and analysis of behavioral, normative, and control beliefs that impact health behaviors (Ajzen, 1991). Constructs of the TPB include attitude, subjective norm, and perceived behavioral control (PBC). A person’s attitude towards a specific behavior is the result of their beliefs about the consequences of that behavior. Subjective norm measures an individual’s belief about what important others want them to do and the motivation to comply with those recommendations. PBC is the individual’s beliefs about the amount of control they have to successfully complete the behavior. Intention can be predicted from the three main psychosocial factors (attitude, subject norm, and PBC) related to that behavior (Ajzen, 1991). In addition, PBC is assumed to influence behavior directly as well (Ajzen, 1988). The predictive power of the TPB has been established in many health behavioral studies (Bissonnette & Contento, 2001; Fila & Smith, 2006; Kassem & Lee, 2004; Robinson & Smith, 2002, 2003). The purpose of the present study was to evaluate whether a garden project could change eating or gardening behavior among urban youth using the TPB model. We were interested in answering the following research questions: (1) How do the constructs of the study model explain intention and behavior during the pre–post gardening project, and (2) Does a high level of pre-survey intention result in behavioral change at post-survey? Methods Study population and design A multi-ethnic, low income sample of youth living in Minneapolis/St. Paul, MN, was recruited and enrolled by Youth Farm and Market Project (YFMP) staff. The YFMP is a multi-cultural gardening enterprise that educates youth about environmental responsibility, empowerment, and cultural expression through active involvement in the planting and harvesting of gardens at three inner-city locations, while fighting against racism and poverty (YFMP, 2005). While the YFMP was established in 1994, the original program did not include nutrition education and cooking as part of the curriculum. The present program design includes a mixture of youth who

123

have gardened with the program the previous year and those who have not; but none of the youth had been exposed to the revised curriculum. The 10-week program (3 days per week) was evaluated with a self-administered survey, using the TPB as the theoretical framework. The nutrition curriculum was taught by a nutrition educator and each week a new topic was introduced (e.g. the food cycle, nutrients) and followed with an activity (e.g. role playing) to foster participatory learning. Similarly, each week a new gardening lesson was introduced that focused on planting, cultivating, and harvesting techniques and the food system. The cooking curriculum emphasized ethnic foods and various kitchen skills. All youth were exposed to the same curriculums; however, attendance was voluntary, therefore the amount of program exposure varied between youth. Ninety-six boys (n ¼ 42) and girls (n ¼ 54), aged 8–15 years, completed the pre-survey with 66 completing the post-survey (boys n ¼ 25, girls n ¼ 41) (Table 1). Youth self-reported their age, grade, gender, and ethnicity and signed assent forms. Approximately one-third were African-American, one-third were White, one-sixth were Hispanic/other Hispanic, and one-eighth were Hmong (Table 1). In addition, at pre-survey, girls consumed more fruit and vegetables than boys; however, boys increased consumption of fruit (p ¼ 0:029) and vegetables (p ¼ 0:007) from pre- to post-survey, with no significant results found in girls (Table 1). Only ethnicity was significantly different between those completing both surveys and those completing only the pre-survey, with more Hmong youth dropping out of the study (7/12). It is probable that because many Hmong families grow crops, their children may have been needed to assist in their fields and thus had to drop out of the YFMP. The project was approved by the University of Minnesota’s Institutional Review Board for Human Subjects. Description of 24-h recall Twenty-four-hour recalls were collected by trained researchers and analyzed using the ESHA Food Processor (v. 7.5, 2000) for nutrients and food groups but only the fruit and vegetable groups were used for these analyses. Youth were asked to describe what they ate the previous day and three-dimensional food models were used to improve estimation of portion sizes. Serving sizes for food groups were consistent with the food guide pyramid; for example, one medium apple was equal to one serving of fruit according to ESHA. Description of survey The survey was developed using the TPB procedures defined by Ajzen (1991) (Fig. 1). Formative assessment for survey development included information obtained through six focus groups with inner-city youth (Lautenschlager & Smith, 2007) and from a review of the literature.

ARTICLE IN PRESS L. Lautenschlager, C. Smith / Appetite 49 (2007) 122–130

124

Table 1 Sample characteristics of Youth Farm and Market Project (YFMP) participants Characteristic

Pre N (%)

Post N (%)

Total Sex Boys Girls

96

66

42 (43.8%) 54 (56.3%)

25 (37.9%) 41 (62.1%)

Age 8–10 11–12 13–14

33 (34.3%) 39 (40.7%) 24 (25.0%)

26 (39.4%) 25 (37.9%) 15 (22.8%)

Ethnicity African American White Hispanic/other hispanic Hmong American Indian Other

34 32 16 12 1 1

24 28 7 5 1 1

(35.4%) (33.3%) (16.7%) (12.5%) (1.0%) (1.0%)

Pre meana (SD)

Post meana (SD)

p valueb

2.01 2.64 2.05 2.85

3.05 2.58 3.43 2.91

0.029** 0.253 0.007** 0.682

(6.4%) (42.4%) (10.6%) (7.6%) (1.5%) (1.5%)

Fruit and vegetable consumptionc Boys’ fruit serving (per day) Girls’ fruit serving (per day) Boys’ vegetable serving (per day) Girls’ vegetable serving (per day)

(1.73) (1.61) (1.34) (2.44)

(3.05) (2.10) (2.52) (1.89)

**Indicates significance at po0.05. a The mean values are indicative of the average number of fruit and vegetable servings consumed per day. b Denotes Wilcoxon Signed Ranks test performed. c Fruit and vegetable consumption was calculated as the mean of the survey behavior questions and 24-h dietary recall data.

Attitude

Subjective Norm

Perceived Behavioral Control Intention

Behavior

Fig. 1. Study model of gardening and dietary habits based on the Theory of Planned Behavior model.

The survey was developed based on common themes identified from focus groups, which included gardening, dietary habits, social influences, nutrition knowledge, and cooking. The survey was reviewed by experts in the fields of public health, nutrition, and agriculture for content and face validity and breadth of coverage and pilot tested with 25 program-eligible youth. Survey scales of the pilot test were examined for internal consistency using Cronbach’s alpha coefficients, which evaluate how well a set of variables measure a single unidimensional latent construct (UCLA, 2005). A high inter-item correlation gives evidence that the items are measuring the same underlying construct (UCLA, 2005). The pilot scales (0.50–0.90) indicated generally good to excellent validity. The final survey was

slightly modified and the Cronbach’s alpha coefficient ranged from 0.67 to 0.92, which is considered in the substantial (0.61–0.80) to almost perfect (0.81–1.0) range (Landis & Koch, 1977). The survey consisted of 177 questions and took on average 20–30 min to complete. A five-point Likert response scale was used for all constructs, except behavior, with youth choosing from ‘‘strongly agree’’, ‘‘agree’’, ‘‘don’t know’’, ‘‘disagree’’, or ‘‘strongly disagree’’. Positively scaled questions were coded from +2 to 2 and negatively scaled questions were coded from 2 to +2. The attitude and PBC constructs were assessed by taking direct measures of each, as set forth by Montano and Kasprzyk (2002). Scale means were used to illustrate youth’s attitudes, subjective norms, PBC, and

ARTICLE IN PRESS L. Lautenschlager, C. Smith / Appetite 49 (2007) 122–130

behavioral intentions related to dietary and gardening behaviors. Fruit and vegetable consumption behavior was determined from a mean computed from the responses to the fruit and vegetable behavior survey questions measured from 0 to 5 servings in the previous 24-h and the 24-h recall food group data. The responses to the survey questions ‘‘How many pieces of fruit did you eat yesterday’’ and ‘‘How many vegetables did you eat yesterday’’ were added, respectively, to the fruit and vegetable data from the 24-h recall and divided by two. For example, if a youth reported eating two pieces of fruit on the survey and three pieces during the 24-h recall (as determined through ESHA), the youth would have a final fruit consumption of 2.5 servings. The attitude scale questions were developed from salient beliefs raised during the focus groups. Fifty-four questions asked the importance of: (1) fruit and vegetable consumption; (2) gardening; (3) cooking; (4) attending YFMP; (5) junk and fast food consumption; (6) ethnic and unfamiliar food consumption; (7) selling produce; (8) sensory properties (e.g. smell, taste, and appearance) of food; (9) vegetarianism; and (10) food advertisements (e.g. ‘‘It is important to eat the foods advertised on TV’’). All questions were treated as a direct measure of attitude which is consistent with other TPB research (Bissonnette & Contento, 2001; Liou & Contento, 2001). Attitude questions were identical for the pre- and post-surveys. The Cronabach’s alpha coefficient for the attitude scale was 0.87. Subjective norm measured who youth felt influenced their decision to consume ethnic or unfamiliar food, garden, cook, eat foods that are good for them, or eat junk or fast food. Seventeen questions asked if family, YFMP, school teachers, friends, or TV told youth to partake in the aforementioned behaviors (e.g. ‘‘My friends tell me to eat fast food’’). The important ‘‘others’’ were identified during focus groups. An additional 17 questions asked if the youth complied to the referent (e.g. ‘‘I eat fast food when my friends tell me to’’). The subjective norm value was calculated as a summation of the product of normative behavior and the motivation to comply for each question. Subjective norm questions were identical for the pre- and post-surveys. The Cronbach’s alpha coefficient for the subjective norm scale was 0.86. PBC measured external factors which may affect youth’s behavior, and in our study 17 questions determined youth’s behavioral control over junk and fast food consumption, gardening, cooking, healthy food availability, fruit and vegetable consumption, ethnic food consumption, and inclusion of herbs in meals (e.g. ‘‘It would be easy to include herbs in my meals because I like the taste’’). Questions were identical for both pre- and post-survey. The Cronbach’s alpha coefficient for the PBC scale was 0.78. Intention was measured at pre-survey with 20 questions asking about youth’s plans for taking care of a garden, cooking, consuming fruits and vegetables, learning about nutrition, consuming ethnic or unfamiliar foods, selling produce, and consuming junk and fast food in the next

125

month (e.g. ‘‘I plan to eat fast food in the next month’’). The post-survey asked questions about their behavior in the last 1–2 months in relation to pre-survey questions as a measure of whether youth followed through with their intentions (e.g. ‘‘I ate fast food in the last month’’). Additionally, five questions measuring future intention were unchanged from the pre-survey. The Cronbach’s alpha coefficient for the intention scale was 0.92. Twelve questions measured behavior by asking about frequency of working in the garden and consumption of ethnic food (e.g. Chinese), fruits, and vegetables. For questions relating to gardening and fruit and vegetable consumption behaviors, youth responded from 0 to 5 times yesterday or last week, depending on the appropriate time frame for the particular behavior. Questions regarding ethnic food consumption behavior were coded as never, once this week, two to three times this week, or daily. Fruit and vegetable consumption was also assessed using the results of the 24-h recall; food group servings and total servings were analyzed for each participant using ESHA. The pre- and post-surveys contained identical behavior questions. The Cronbach’s alpha coefficient for the behavior scale was 0.67. Description of survey administration Surveys were self-administered to youth at the three YFMP sites. They were instructed to mark only one answer per question and research assistants helped any youth who had difficulties reading and/or understanding questions. All surveys were rechecked for multiple and missed responses and were corrected prior to the youth departing. Data analysis After analyzing the 24-h recalls for the food group data, the fruit and vegetable data were then entered into Statistical Package for Statistical Sciences (SPSS, v. 12.0, Chicago, IL, 2003) for further analysis. Descriptive statistics were used to determine means and standard deviations of constructs and to summarize the demographic data. To evaluate differences between gender and ethnicity, Mann–Whitney U analyses were conducted on the means because the construct data were non-normally distributed. Gender differences were found regarding associations between intention and behavior and the constructs correlated with behavior. However, no ethnic differences were observed; therefore, further statistical analyses were stratified by gender only. Spearman rho correlations were conducted to examine associations between the psychosocial variables (constructs). In order to examine constructs (attitude, subjective norm, and PBC) most predictive of intention at pre-survey and behavior at post-survey, stepwise regression analyses were conducted. A stepwise regression analysis was also done to determine if pre-survey intention predicted post-survey behavior. Preand post-survey intention questions were also evaluated for

ARTICLE IN PRESS L. Lautenschlager, C. Smith / Appetite 49 (2007) 122–130

126

Table 2 Boys’ and girls’ stepwise regression analyses of gardening and eating behavior using the Theory of Planned Behavior (TPB) model Boys Variable Pre-survey Dependent variable: intentiona Attitude Post-survey Dependent variable: behaviorb Perceived behavioral control

Girls

B

Standard error

R2

p-value

B

Standard error

R2

p-value

1.525

0.306

0.519

0.000

1.421

0.251

0.452

0.000

ND

ND

ND

ND

0.303

0.117

0.147

0.014

ND indicates not determinable. a Candidate variablesc: attitude, subjective norm, and Perceived Behavioral Control. b Candidate variables: attitude, subjective norm, and Perceived Behavioral Control. c Candidate variables are all of the variables included in the stepwise regression.

differences using a paired sample t-test, because variables in the intention construct were normally distributed. The pre-survey intention questions were parallel to the postsurvey questions; however, the post-survey questions were worded to measure past behavior. For example, the presurvey question ‘‘I plan to learn more about planting a garden’’ corresponded with the post-survey question: ‘‘In the last 2 months, I learned more about planting a garden’’. Changes in fruit and vegetable consumption from pre- to post-survey were analyzed using the Wilcoxon Signed Rank Test. The significance for data analysis was set at po0.05.

Results Boys Correlations were conducted on the model constructs. No association was found between intention and behavior. Additionally, none of the constructs correlated directly to behavior; therefore, analysis was conducted only on predictors of intention for boys. Examination of the presurvey TPB model constructs found that subjective norm (r ¼ 0:83; po0.01) and attitude (r ¼ 0:72; po0.01) were significantly correlated with intention to change. For the post-survey, attitude (r ¼ 0:83; po0.01), subjective norm (r ¼ 0:74; po0.01), and PBC (r ¼ 0:72; po0.01) were all significantly correlated with intention. Stepwise regression was conducted with intention as the dependent variable and found that attitude was most predictive at pre-survey and accounted for 52% of the variance, respectively (Table 2). Results from the stepwise regression analysis at post-survey found that none of the constructs predicted behavior (Table 2). Stepwise regression was also conducted on pre-survey intention and post-survey behavior and results indicated that pre-survey intention did not predict post-survey behavior (data not shown).

Girls Unlike the boy’s pre- and post-survey data, a significant association was found between intention (r ¼ 0:33; po0.05) and behavior for the girl’s at pre-survey. Constructs associated with pre-survey intention included attitude (r ¼ 0:67; po0.01) and subjective norm (r ¼ 0:54; po0.01). In contrast to the pre-survey data, no association was found between intention and behavior for the postsurvey, but PBC (r ¼ 0:38; po0.05) was found to be associated with behavior. Stepwise regression analysis showed that attitude alone predicted 45% of the variance in intention for the pre-survey (Table 2). For behavior at post-survey, results indicated that only PBC was slightly predictive of behavior and predicted 15% of the variance (Table 2). Stepwise regression was also conducted on presurvey intention and post-survey behavior and results indicated that pre-survey intention did not predict postsurvey behavior (data not shown). All youth In order to determine if youth with high levels of intention to change behavior at pre-survey time exhibited a high degree of post-survey behavioral change, specific questions within the pre-survey intention construct were matched with post-survey questions and paired sample t-tests were performed on the differences. Results indicated no significant differences for boys between pre- and postsurveys; however, three trends were evident (Table 3). Boys who intended to learn more about planting or weeding a garden on the pre-survey reported on the post-survey that they did follow through and learned about planting (0.440,1 p ¼ 0:078) and weeding a garden (0.520,1 p ¼ 0:091). However, boys who indicated on the pre-survey that they planned to help their family garden did not follow 1 Survey scale: +2—strongly agree; +1—agree; 0—don’t know; 1— disagree; 2—strongly disagree.

ARTICLE IN PRESS L. Lautenschlager, C. Smith / Appetite 49 (2007) 122–130

127

Table 3 Boys’ and girls’ pre-survey intentions and corresponding post-survey behavior differences Boys (n ¼ 25) Variable pairs

Mean

Standard deviation

I plan to learn more about planting a garden In the last 2 months I learned more about planting a garden Difference between pre-survey intention and post-survey behavior means I plan to learn more about planting fruits/vegetables/herbs In the last 2 months I learned more about planting fruits/vegetables/herbs Difference between pre-survey intention and post-survey behavior means I plan to learn more about nutrition In the last 2 months I learned more about nutrition Difference between pre-survey intention and post-survey behavior means I plan to learn more about how to cook In the last 2 months I learned more about how to cook Difference between pre-survey intention and post-survey behavior means I plan to learn more about ethnic foods In the last 2 months I learned more about ethnic foods Difference between pre-survey intention and post-survey behavior means I plan to learn more about selling vegetables/fruits/flowers In the last 2 months I learned more about selling vegetables/fruits/flowers Difference between pre-survey intention and post-survey behavior means I plan to learn more about watering a garden In the last 2 months I learned more about watering a garden Difference between pre-survey intention and post-survey behavior means I plan to learn more about weeding a garden In the last 2 months I learned more about weeding a garden Difference between pre-survey intention and post-survey behavior means In the next month, I plan to help my family cook at least once In the last month I helped my family cook at least once Difference between pre-survey intention and post-survey behavior means In the next month, I plan to help my family garden In the last month, I helped my family garden Difference between pre-survey intention and post-survey behavior means In the next month, I plan to try new foods In the last month, I tried new foods Difference between pre-survey intention and post-survey behavior means For the next month, I plan to eat two pieces of fruit everyday In the last month I ate two pieces of fruit everyday Difference between pre-survey intention and post-survey behavior means For the next month, I plan to eat three servings of vegetables everyday In the last month I ate three servings of vegetables everyday Difference between pre-survey intention and post-survey behavior means I plan to eat fast food in the next month I ate fast food in the last month Difference between pre-survey intention and post-survey behavior means I plan to eat foods like pop, chips, and/or sweets in the next month I ate foods like pop, chips, and/or sweets in the last month Difference between pre-survey intention and post-survey behavior means

0.40b 0.84b 0.44 0.60b 0.72b 0.12 0.40b 0.84b 0.44 0.80b 0.84b 0.04 0.36b 0.44b 0.08 0.40b 0.64b 0.24 0.28b 0.60b 0.32 0.36b 0.88b 0.52 0.80b 0.72b 0.08 0.20b 0.20b 0.40 0.60b 0.72b 0.12 0.64b 0.32b 0.32 0.32b 0.08b 0.24 0.60c 0.84c 0.24 0.44c 0.76c 0.32

1.26 1.11

Girls (n ¼ 41) p-valuea (2-tailed)

0.078* 1.19 1.17 0.559 1.26 1.11 0.141 1.12 1.03 0.901 1.11 1.08 0.759 1.19 1.00 0.387 1.24 1.16 0.235 1.25 0.83 0.091* 0.91 1.14 0.603 1.19 1.29 0.096* 0.82 1.02 0.622 1.08 1.07 0.148 1.07 0.86 0.341 1.12 0.99 0.387 1.33 1.05 0.349

Mean

Standard deviation

0.95b 1.00b 0.05 1.00b 1.10b 0.10 0.95b 1.10b 0.15 1.17b 0.73b 0.44 0.81b 0.42b 0.39 0.81b 0.44b 0.37 0.56b 0.78b 0.22 0.39b 0.73b 0.34 1.10b 0.93b 0.17 0.49b 0.39b 0.88 1.00b 0.73b 0.27 0.68b 0.46b 0.22 0.71b 0.20b 0.51 0.07c 0.61c 0.54 0.42c 0.85c 0.43

0.84 0.84

p-valuea (2-tailed)

0.736 0.67 0.80 0.486 0.84 0.86 0.421 0.86 1.07 0.012** 0.93 1.00 0.048** 0.84 1.03 0.007** 0.95 0.94 0.262 1.07 0.87 0.085* 0.94 0.91 0.369 1.03 1.07 0.000** 0.87 0.95 0.182 1.04 0.95 0.291 1.01 0.90 0.003** 1.03 1.05 0.005** 1.12 0.96 0.006**

*Indicates trend at p40.05o0.10. ** Indicates significance at po0.05. a Denotes paired sample t-test performed. b Intention and behavior scales: 2—strongly disagree; 1—disagree, 0—don’t know; 1—agree; 2—strongly agree. c Intention and behavior scales: 2—strongly agree; 1—agree; 0—don’t know; 1—disagree; 2—strongly disagree.

through with this behavior (2.20,1 p ¼ 0:096). Alternatively, for girls significant differences were seen with seven variables (Table 3). Specifically, girls who indicated that they intended to learn more about ethnic foods, cooking skills, or selling vegetables/fruit/flowers on the pre-survey

reported on the post-survey that they did not learn more about how to cook (0.438,1 p ¼ 0:012), eat ethnic foods (0.390,1 p ¼ 0:048), or sell produce (0.366,1 p ¼ 0:007). Girls did not follow through with their presurvey intentions to help their family garden, eat three

ARTICLE IN PRESS 128

L. Lautenschlager, C. Smith / Appetite 49 (2007) 122–130

servings of vegetables everyday, eat fast food,2 and/or eat foods like pop, chips, and/or sweets in the next month (0.878,1 p ¼ 0:000; 0.512,1 p ¼ 0:003; 0.537,3 p ¼ 0:005; 0.4393, p ¼ 0:006, respectively). Discussion To our knowledge, this is the first study to examine the influence of a garden program on gardening, cooking, and the consumption of fruits, vegetables, and ethnic foods using the TPB model. The major findings of this study were that: (1) the TPB model was successful in explaining the variance in intention and behavior in this group of youth; (2) attitude was most predictive of pre-survey intention for both boys and girls; (3) gender differences were noted, with post-survey behavior associated to PBC in girls, with no constructs associated to post-survey behavior in boys; (4) the garden program positively impacted boys fruit and vegetable consumption; (5) a high level of intention for boys pre-survey marginally predicted some behavior post-survey; and (6) high level of intention for girls pre-survey did not result in positive behavioral change at post-survey. The TPB model The variance in intention and behavior was sufficiently explained by the TPB model, indicating that the constructs were appropriate for this population, which consisted of boys and girls of varying ethnicities and a substantial age range (8–15 years). Similarly, Backman, Haddad, Lee, Johnston, and Hodgrin (2002) also surmise that the TPB model predicts healthy dietary behavior in adolescents of both genders from different ethnic groups. Intention Attitude was found to be a stronger predictor of presurvey intention for both boys and girls compared to subjective norm or PBC. This finding is consistent with the predictive patterns found by Backman et al. (2002) who used the TPB to investigate eating behavior in youth. An attitude toward a specific behavior reflects an individual’s beliefs about the consequences of that behavior. In terms of our study, we believe that the youth personally felt it was important to partake in the specified behavior (e.g. garden or cook), therefore reported that they intended to act. Bissonnette and Contento (2001) also expressed this belief and others have also found that attitude is an important factor of intention in youth (A˚strøm & Okullo, 2004; Bissonnette & Contento, 2001; Kassem & Lee, 2004). 2 Defined in survey as ‘‘McDonalds, Burger King, Taco Bell, KFC, burgers, fries, shakes, and pizza’’. 3 Survey scale: +2—strongly disagree; +1—disagree; 0—don’t know; 1—agree, 2—strongly agree.

Gender differences: behavior Gender differences also existed in the prediction of behavior. None of the constructs correlated directly to post-survey behavior for boys, but PBC was found to be associated with the girls’ post-survey behavior. The finding that PBC was identified as most predictive of behavior change in girls, suggests that in this population, the perception of control is related to reported behavior. However, PBC only marginally predicted post-survey behavior, a finding consistent with others (Bissonnette & Contento, 2001). Perhaps the manner in which the variable was constructed provides a reason for its low predictability (Bissonnette & Contento, 2001). The variable reflects both inner control factors and external perceived difficulty factors, and as Sparks, Guthrie, and Shepherd (1997) suggested, the variable may not be one scale, as it is traditionally used. Future research is needed to examine the effect of using a multi-scale PBC construct with youth in order to determine its possible effectiveness. Fruit and vegetable consumption only increased among boys; however, girls consumed more fruit than boys at the onset of the program and would be expected to have only a modest increase at best. Others have also reported that girls consume more fruits and vegetables than boys (Reynolds, Hinton, Shewchuk, & Hickey 1999; Le Bigot Macaux, 2001). Several explanations for the disparity in fruit and vegetable consumption between boys and girls have been reported, including greater preference or liking among girls (Le Bigot Macaux, 2001; Cooke & Wardle, 2005) and social desirability bias (Cooke & Wardle, 2005). A greater understanding of the gender differences related to fruit and vegetable intake is needed if nutrition-related youth programs are to be successful. To increase girls’ consumption of fruits and vegetables, programs might consider incorporation of taste tests to demonstrate different ways to prepare the food. Gender differences: pre-survey intention vs. post-survey behavior Differences between boys and girls in terms of pre-survey intention and post-survey behavior were evident, with girls less likely to follow through with a behavior change. It is unclear why girls, in contrast to the boys, did not follow through and make behavioral changes. We believe that the skill was learned, but girls decided that they disliked the activity, and therefore, chose not to make a behavioral change. Our results suggest that the girls did not develop the needed skills for behavior performance as strongly as the boys did from the YFMP program. The importance of gender differences in cognitive abilities has been suggested (Rocha, Rocha, & Menezes, 2005) and perhaps the YFMP curriculum did not take these dissimilarities into account when disseminating the material. Learning, primarily an active process, occurs most optimally when there is internal motivation on the part of the learner to engage and

ARTICLE IN PRESS L. Lautenschlager, C. Smith / Appetite 49 (2007) 122–130

assimilate information (deCharms, 1976; Thomas, 1980). Individual differences in children’s motivational orientation have been documented to influence performance (Boggiano, Barret, & Main, 1985; Harter & Connell, 1984) and we believe this is evident in our findings. Motivators can be intrinsic (e.g. self-determination) or extrinsic (e.g. rewards or grades) in nature (Grolnick & Ryan, 1987) and differ in the degree to which the learner’s internal motivation or autonomy is encouraged (Deci & Ryan, 1985; Ryan, Connell, & Deci, 1985). The YFMP project appears to emphasize intrinsic control, as few external motivators are used during the education curriculum. Intrinsic controls enable more interest in the activity and more willingness to do tasks of similar nature in the future (Grolnick & Ryan, 1987). Boys may have been more successful in behavior change because they respond best to intrinsic motivators compared to girls. Perhaps girls are motivated to learn and perform through extrinsic controls, a component that is not present in the YFMP program. However, to date, there is a paucity of literature concerning learning and motivational factor differences between genders; in order to develop and properly disseminate youth education curriculums, this area needs further exploration. Conclusions In summary, this study found that a garden program modestly changed eating and gardening behavior among an urban sample of youth. Increased fruit and vegetable consumption in boys were positive behavioral changes observed from those participating in this gardening project. In addition, both boys and girls with intentions to learn about gardening were found to have made this behavior change at post-survey. These findings suggest that a school or community garden for youth may be a viable way to assist youth in making healthy lifestyle changes. This study also found that the TPB model is useful in providing insight into whether a garden program is successful in changing eating or gardening behavior among youth. Our results show that multiple factors, including intention, attitude, subjective norm, and PBC influence gardening and dietary behavior in youth. The synergistic use of these factors in the development and implementation of future gardening programs may assist in the promotion of gardening and healthy eating among youth. However, in order for garden programs to play a role in improving gardening and dietary habits, an understanding of youth attitudes, subjective norms, and PBC is needed. Limitations Although this study found that multiple factors influence gardening and dietary behavior among youth, several limitations are evident. First, because our sample of Hmong youth already gardened with their families (Lautenschlager & Smith, 2007), it is possible that this

129

group exerted a bias on the sample. However, gardening experience is something that would be difficult to control for because all youth bring different agriculture skill sets to the program. Second, this study did not include a control group because of program constraints. The small sample size made it difficult to split the group and YFMP staff felt it would be unethical to not expose a sub-group of YFMP to the new curriculum. In addition, because of time and cost limitations, we were unable to use a delayed program control study design that would allow a sub-group of YFMP to receive the curriculum later. Acknowledegments We would like to thank the YFMP staff, including Jeff Bauer, Jessica Flannigan, Deb Klein, and Gunnar Liden for their help in execution of the surveys. We gratefully acknowledge the youth for their willingness to complete the surveys. We also thank Jamie Butterfass, Pa Lee, and Rickelle Richards for their assistance in survey administration and data entry. References Ajzen, I. (1988). Attitudes, personality and behavior. Milton Keynes: Open Univ. Press. Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Decision Processes, 50, 179–211. American Dietetic Association (ADA). (2003). Position of the American Dietetic Association, Society for Nutrition Education, and American School Food Service Association—Nutrition services: An essential component of comprehensive school health programs. Journal of the American Dietetic Association, 103, 505–514. Armitage, C. J., & Conner, M. (2001). Social cognitive determinants of blood donation. Journal of Applied Social Psychology, 31, 1431–1457. A˚strøm, A. N., & Okullo, I. (2004). Temporal stability of the theory of planned behavior: A prospective analysis of sugar consumption among Ugandan adolescents. Community Dentistry and Oral Epidemiology, 32, 426–434. Backman, D. R., Haddad, E. H., Lee, J. W., Johnston, P. K., & Hodgrin, G. E. (2002). Psychosocial predictors of healthful dietary behavior in adolescents. Journal of Nutrition Education and Behavior, 34, 184–193. Berkeley Community Gardening Collaborative (BCBG). (2005). Accessed November 7, 2005 from /http://www.ecologycenter.org/bcbg/S. Birch, L. L., & Fisher, J. O. (1997). Food intake regulation in children. Fat and sugar substitutes and intake. Annals of the New York Academy of Sciences, 23, 194–220. Bissonnette, M. M., & Contento, I. R. (2001). Adolescents’ perspectives and food choice behaviors in terms of the environmental impacts of food production practices: Application of a psychosocial model. Journal of Nutrition Education, 33, 72–82. Boggiano, A. K., Barret, M., & Main, D.S. (1985). Mastery motivation in children: the role of an extrinsic vs. intrinsic orientation. Paper presented at the meeting of the American Educational Research Association, Chicago. Cooke, L. J., & Wardle, J. (2005). Age and gender differences in children’s food preferences. British Journal of Nutrition, 93, 741–746. deCharms, R. (1976). Enhancing motivation: Change in the classroom. New York: Irvington. Deci, E. L., & Ryan, R. M. (1985). Intrinsic motivation and selfdetermination in human behavior. New York: Plenum Press. Fila, S., & Smith, C. (2006). Applying the theory of planned behavior to healthy eating behaviors in urban Native American youth.

ARTICLE IN PRESS 130

L. Lautenschlager, C. Smith / Appetite 49 (2007) 122–130

International Journal of Behavioral Nutrition and Physical Activity, 3, 11 (Epub ahead of print). Grolnick, W. S., & Ryan, R. M. (1987). Autonomy in children’s learning: An experimental and individual difference investigation. Journal of Personality and Social Psychology, 52, 890–898. Harter, S., & Connell, J. P. (1984). A model of the relationship among children’s academic achievement and their self-perception of competence, control, and motivational orientation. In J. Nicholls (Ed.), The development of achievement motivation (pp. 219–250). Greenwich, CT: JAI Press. Kassem, N. O., & Lee, J. W. (2004). Understanding soft drink consumption among male adolescents using the Theory of Planned Behavior. Journal of Behavioral Medicine, 27, 273–296. Landis, J. R., & Koch, G. G. (1977). The measurement of observer agreement for categorical data. Biometrics, 33, 159–174. Lautenschlager, L., & Smith, C. (2007). Beliefs, knowledge, and values held by inner-city youth about gardening, nutrition, and cooking. Agriculture and Human Values, in press. Le Bigot Macaux, A. (2001). Eat to live or live to eat? Do parents, children agree? Public Health Nutrition, 4, 141–146. Liou, D., & Contento, I. R. (2001). Usefulness of psychosocial theory variables in explaining fat related dietary behavior in Chinese Americans: Association with degree of acculturation. Journal of Nutrition Education., 33, 322–331. Ludwig, D. S., & Ebbeling, C. B. (2001). Type 2 diabetes mellitus in children: primary care and public health considerations. Journal of the American Medical Association, 286, 1427–1430. Montano, D. E., & Kasprzyk, D. (2002). The theory of reasoned action and the theory of planned behavior. In K. Glanz, B. K. Rimer, & F. M. Lewis (Eds.), Theory, research, and practice in health behavior and health education. San Francisco, CA: Josey Bass. Mun˜oz, K. A., Krebs-Smith, S. M., Ballard-Barbash, R., & Cleveland, L. E. (1997). Food intakes of US children and adolescents compared with recommendations. Pediatrics, 100, 323–329. Morris, J. L., & Zidenberg-Cherr, S. (2002). Garden enhanced nutrition curriculum improves fourth-grade school children’s knowledge of nutrition and preferences for some vegetables. Journal of the American Dietetic Association, 102, 91–93. Nicklas, T., & Johnson, R. (2004). Position of the American Dietetic Association: Dietary guidelines for healthy children ages 2 to 11 years. Journal of the American Dietetic Association, 104, 660–677. Northeast Organic Farming Association (NOFA). (2005). NOFA Massachusetts Youth Urban Gardening Project. Accessed November 7, 2005 from /http://www.nofamass.org/programs/justice/index.phpS.

Reynolds, K. D., Hinton, A. W., Shewchuk, R. M., & Hickey, C. A. (1999). Social cognitive model of fruit and vegetable consumption in elementary school children. Journal of Nutrition Education, 31, 23–30. Robinson, R., & Smith, C. (2002). Psychosocial and demographic variables associated with consumer intention to purchase sustainably produced foods as defined by the Midwest Food Alliance. Journal of Nutrition Education and Behavior, 34, 316–325. Robinson, R., & Smith, C. (2003). Integrating issues of sustainably produced foods into nutrition practice: A survey of Minnesota Dietetic Association members. Journal of the American Dietetic Association, 103, 608–611. Rocha, F. T., Rocha, A. F., & Menezes, R. (2005). Brain mappings of the arithmetic processing in children and adults. Cognitive Brain Research, 22, 359–372. Ryan, R. M., Connell, J. P., & Deci, E. L. (1985). A motivational analysis of self-determination and self-regulation in education. In C. Ames, & R. E. Ames (Eds.), Research on motivation in education: The classroom milleu (pp. 13–51). New York: Academic Press. Smith, C., & Rinderknecht, K. (2003). Obesity correlates with increased blood pressure in Urban Native American youth. American Journal of Human Biology, 15, 78–90. Sparks, P., Guthrie, C. A., & Shepherd, R. (1997). The dimensional structure of the perceived behavioral control construct. Journal of Applied Social Psychology, 27, 418–438. Thomas, J. W. (1980). Agency and achievement: Self-management and self-regard. Review of Educational Research, 50, 213–240. Troiano, R., & Flegal, K. M. (1998). Overweight children and adolescents: Description, epidemiology, and demographics. Pediatrics, 101, 497–504. Troiano, R. P., Flegal, K. M., Kuczmarski, R. J., Campbell, S. M., & Johnson, C. L. (1995). Overweight prevalence and trends for children and adolescents. The National Health and Nutrition Examination Surveys, 1963 to 1991. Archives of Pediatric and Adolescent Medicine, 149, 1085–1091. UCLA Academic Technology Services. (2005). SPSS FAQ: What does Cronbach’s alpha mean? Accessed June 1, 2005 from /http:// www.ats.ucla.edu/stat/spss/faq/alpha.htmlS. US Dept of Health and Human Services (USDHHS). (2000). Healthy people 2010: Understanding and improving health (2nd ed). Young, T. K., Dean, H. J., Flett, B., & Wood-Steiman, P. (2000). Childhood obesity in a population at high risk for type 2 diabetes. Journal of Pediatrics, 136, 365–369. The Youth Farm and Market Project (YFMP). (2005). Accessed March 7, 2005 from /www.youthfarm.netS.