Measures
The survey was developed according to procedures defined by Ajzen and Fishbein. Formative assessment included a review of literature and six focus group discussions with 39 youth, ages 9 to 18 years, to determine common beliefs about eating healthy, advantages and barriers to eating healthy, and important people who may influence behavior. Focus groups were audio taped and transcribed verbatim and analyzed for common themes. The themes were used to develop a survey using the constructs of the TPB to investigate healthy eating behavior. These include the original constructs of attitudes, subjective norm, PBC, as well as two additional constructs, self-efficacy and barriers (Figure 1). Self-efficacy was included to assess whether it was more predictive of intention and/or behavior than PBC. The construct 'barriers' was included in the model because the staff of the program was particularly interested in identifying factors that prevent youth from eating healthfully. The survey included 90 questions pertaining to eating behavior and measured all constructs. It was evaluated by an expert panel of nutritionists accomplished in the areas of survey development and behavioral theory for face validity and breadth of coverage and then pilot tested in a group of urban Native American youth (n = 32) attending an alternative high school for ease of comprehension and readability. The final version of the survey contained slight revisions in wording and question ordering based on recommendations from professionals and findings from the pilot study. Scales for the final survey were assessed using the Cronbach alpha coefficient, an index of inter-item homogeneity (internal consistency.) The reliability levels for the: attitude scale was 0.66, subjective norm scale was 0.66, PBC scale was 0.80, intention scale was 0.84, behavior scale was 0.72, barriers scale was 0.89, and self-efficacy scale was 0.85. These scores indicate a substantial (0.61 – 0.80) to almost perfect (0.81–1.0) range of reliability.
Survey questions were assessed using a five-point Likert response scale, youth had to choose one of the following responses "strongly agree", "agree", "unsure", "disagree", or "strongly disagree". For positively scaled questions responses were coded from 2 to -2 (strongly agree to strongly disagree) and for negatively scaled questions responses were coded from -2 to 2 (strongly agree to strongly disagree). Attitudes to eating healthy were measured by using the responses to eighteen questions about the importance of eating healthy foods, fruits, vegetables, regular pop, junk food, and fast food, and perceptions of eating healthy, being under or overweight, and diabetes (e.g. "It is important to me to eat healthy foods everyday"). Subjective norm was measured by using the responses to eight questions asking if parents, friends, elders, community programs, or television told youth to eat healthy everyday (e.g."My parents tell me it is important to eat healthy everyday"). PBC measured external factors that may directly or indirectly affect healthy eating behavior. Responses to eight questions concerning youth's perceived control over eating healthy, eating junk food, drinking regular pop, eating fast food, eating in front of the TV, and getting diabetes, as well as having fruits and vegetables available (e.g. "I have control over whether or not I eat healthy") were used to determine perceived behavior control. Intention to eat healthy was measured by using the responses to eight questions regarding youth's plans for the next week to eat healthy, eat vegetables, eat fruit, not eat junk food, not eat fast food, not drink regular pop, eat healthy foods in front of the TV, and eat healthy foods to keep a healthy weight everyday (e.g. "For the next week I plan to eat healthy everyday"). Eating behavior was measured by using the responses to questions assessing dietary intake of vegetables, fruits, soft drinks, and fast food consumption, along with eleven additional behavior questions using the Likert-scale. Eating behaviors questions included general questions such as, "I mostly eat healthy foods," "I eat healthy to keep me from getting diabetes," and "I eat junk food when I watch TV," but also used specific foods such as "fruits" and "vegetables" because youth defined foods, especially fruits and vegetables, as being healthy. Both healthy and junk foods were defined on the survey using terminology that the youth used during the focus group discussions. The assessment and behavioral question scales were averaged, recoded, and the sum of the two were calculated to measure youth's behavior.
The two additional constructs, barriers and self-efficacy, were included in the expanded model. Barriers to eating healthy were measured by using the responses to fourteen questions concerning youth's perceptions about the taste of fruits, vegetables, regular pop, and junk food; the ease of eating healthy away from home, with friends, with family, in front of the TV and to keep a healthy weight; and availability of healthy foods (e.g. "It is hard for me to eat healthy foods at fast food restaurants"). Self-efficacy measured internal factors that may directly or indirectly affect healthy eating behavior. Responses to fifteen questions determining youth's ability to eat healthy foods and choose specific healthy foods (fruit, vegetables, chocolate milk, juice, white milk, low-fat milk, salads) over specific unhealthy foods (junk food, chips/cheetos, regular pop, chocolate milk, whole milk, hamburgers) (e.g. "I can eat healthy foods everyday") were used to assess self-efficacy to eat healthy.