3 Theoretical and conceptual framework
3.6 Constructing questionnaires based on the Theory of Planned Behaviour
3.5.2 Subjective norm
Subjective norms refer to the social pressure to perform or to not perform a particular behaviour121. This includes beliefs about social pressure and motivation to comply121. In this context, for example, pupils who consume school meals often do so due to encouragement from the school, whereas pupils who often skip school lunch often might not be receiving sufficient encouragement to consume school lunches.
3.5.3 Perceived behavioural control
Perceived behavioural control refers to the individual’s perception of the ease or difficulty of performing or not performing the behaviour of interest121. Perceived behavioural control is made up of both internal and external factors121. It can be demonstrated with regard to consuming more school meals, which pupils may see as easy on an assessment of price, the availability of preferable food, and the facilities available. For example, available meals may be affordable, contain tasty food and pupils may have time available to consume them. However, they also may perceive a loss of behavioural control because the quality of school meals may be poor, such as if they were poorly packaged or if they were associated with unpleasant smells e.g. from fish dishes.
3.5.4 Intention and behaviour
Intention and behaviour refers to whether or not a person intends, and is able, to carry out a behaviour121. For example, the SFCs' intention to adopt a school food policy reflects how strong they intend to provide healthy school meals for school children. Another example is whether children would like to eat more fruit and vegetables in the future, and whether schools provide healthy nutritional menus for pupils.
3.6 Constructing questionnaires based on the Theory of Planned Behaviour 3.6.1 Adapted theory of planned behaviour
The iPOPY project provided the first draft of the questionnaire and recommended that the questionnaire development should be based on social research theory such as the TPB. By these recommendations, the questionnaires should be able to look at not only eating behaviours, but also the environmental factors that influence the eating habits of children. The present project thereby assumes that social factors play an important role in the creation of healthy eating environments, and so this study draws on an adapted theory of planned behaviour. The adapted theory of planned behaviour is derived from the basic beliefs and structure of the TPB model.
According to the TPB, behaviour does not stem from a direct path from attitude. The present studies were interested in investigating the direct effect of relationships between attitude and behaviour. In the adapted theory of planned behaviour (see Fig. 2), attitude, subjective norm, perceived behavioural control, intention and action/behaviour are further decomposed into
37
smaller constructs. This provides a more comprehensive explanation of the adoption of behaviour163-166 by using the hypotheses of the project that are listed in Table 1. These involve important determinants of SFCs and pupil’s action/behaviour. In the present studies, both quantitative and qualitative methods used the adapted theory of planned behaviour to construct the questionnaires and interview guidelines.
3.6.2 Attitudinal measures
For directly measuring SFCs’ and pupils’ attitudinal issues, the current study used the common rating methods of agree/disagree scales and verbal rating scales167,168. Previous reviews indicate that verbal rating scales improve the validity and reliability of results and are easily administrated to informants167,168. Verbal rating scales might not be appropriate for use in telephone surveys due to difficulties in remembering each scale by respondents167,168. In the web survey model of current project, however, this was not a barrier.
Example of a verbal scale from the questionnaire in study III Do you think you are healthy?
1. I am very healthy 2. I am healthy
3. I am almost healthy 4. I am not so healthy 5. I am unhealthy
Agree/disagree scales are commonly used in research due to the benefits of repeatable, easily understood questions168,169. However, agree/disagree scales are also prone to compliance bias because respondents are more likely to accept agreement than disagreement168,169. In order to avoid this type of bias, as well as to encourage respondents to read each question and think about each option thoroughly rather than simply repeating agree/disagree, the current project structured the positive and negative options with a Likert-type scale such as from “Strongly agree” to “Strongly disagree” 168,169.
Example of an agree/disagree scale from the questionnaire in study III I think our school meals are healthy.
1. Strongly agree 2. Agree
3. Somewhat agree 4. Somewhat disagree 5. Disagree
6. Strongly disagree
38
Table 1. Hypothesis setting based on the adapted theory of planned behaviour constructs.
Number Hypothesis statement
Adapted theory of planned behaviour
construct H1 Organic school food procurement policy is positively associated
with attitude towards promotion of organic food for pupils. Attitude H2
Organic school food procurement policy is positively associated with attitude towards promotion of healthy eating habits for pupils.
Attitude
H3 Organic school food procurement policy is positively associated
with pupils’ attitude about organic food and health. Attitude H4 Organic school food procurement policy is positively associated
with pupils’ attitude towards school meals. Attitude
H5
Organic school food procurement policy is positively associated with pupils’ received encouragement from school regarding eat healthily and consume school meals.
Subjective norm H6 Organic school food procurement policy is positively associated
with pupils’ confidence on school food service.
Perceived behavioural control H7 Organic school food procurement policy is positively associated
with pupils’ willingness on school food service.
Perceived behavioural control H8 Organic school food procurement policy is positively associated
with adoption of a school food policy. Intention
H9 Organic school food procurement policy is positively associated
with adoption of a health promoting school policy. Intention H10 Organic school food procurement policy is positively associated
with pupils’ intention of healthier eating habits. Intention H11 Organic school food procurement policy is positively associated
with provision of healthy school food and drink practices. Action
H12 Organic school food procurement policy is negatively associated
with provision of unhealthy school food and drink practices. Action
H13 Organic school food procurement policy is positively associated
with encouragement of healthy eating for pupils. Action
H14 Attitude items towards are positively associated with action items. Attitude
39
Figure 2. Adapted theory of planned behaviour for studying factors influencing school food environment and pupils’ eating habits.
Action
Encouragement from school to eat healthily and
consume school meals Intention
40
Table 2. Example questions based on the adapted theory of planned behaviour constructs.
Adapted theory of planned behaviour
construct Example questions
Attitudes Beliefs about a behaviour
I think school has a responsibility to promote healthy eating habits via the school food service.
Evaluation about performance of behaviour
It is important for me to eat healthy food.
Subjective norms Social pressure Does your school encourage you to eat healthy food?
Motivation to comply
Have you sometimes not eaten this food because it is unhealthy?
Perceived
behavioural control
Internal factors Do you like to eat school meals?
External factors Do you have enough time to eat lunch?
Intention I would like to eat more organic food in the future.
Action Does your school serve nutritionally calculated
meals according to official guideline?
For indirect measurements of attitudinal issues, this study designed the questions to be specific170,171. Asking specific questions can improve the accuracy of collected data and avoid vagueness from respondents172. However, the disadvantages of question specificity are an increased cost of surveys and questionnaire length, which add to respondent burden and can cause a reduction in response rate139,173. Table 2 lists the example questions of attitudinal measures.
3.6.3 Behavioural measures
This project was interested in measuring actions undertaken by schools with respect to building a healthy school environment, as well as actions related to pupils’ school food consumption. The units of measurement for school action were chosen to measure the specific activities carried out by schools to improve school health, particularly with respect to school meals and how these might impact pupils’ eating habits. In the process of developing the units of measurement, the project considered the respondents’ perspectives carefully174. Accordingly, we defined different key terms in the questionnaire, such as what a Food and Nutrition Policy (FNP) contains, in a way the school staff would find easy to understand and provided any information thought necessary141,175. The present study did not address questions that required plenty of cognitive effort from respondents, e.g., ‘How many grams of salt did you consume yesterday?’149 .
41
Example question of measuring school action from the questionnaire in studies I & II Does your school recommend nutritional menus for pupils in canteen?
1. Yes 2. No
3. Don’t know
The units of measurement for pupils focused more on the frequency of their action towards defined food categories (e.g. fresh fruit, processed vegetables, fizzy drinks without sugar) so that pupils’ general eating habits could be measured more specifically. The questionnaire included banded frequency questions for the consumption of food categories168. In order to avoid the differences in respondent’s opinions as to what composes various time intervals such as “often”
or “sometimes”, the present study chose to employ specific quantifiers (e.g. never, everyday, 1-2 times per day)168. For banded frequency questions, the longer the reference period the more suitable it is for accurate data collection168. On the other hand, it is often difficult for respondents to recall food consumption if the period is for more than a number of days168. In addition to this, the current study aimed to investigate frequency of action within school time only. Consequently, we determined the reference period for this part of the study as the children’s school period from start point to research date.
Example question of measuring frequency action from the questionnaire in study III How often do you buy school meals?
1. Never
2. Less than one day per week 3. One day per week
4. 2-4 days per week 5. 1 time per day 6. 2 times per day
7. More than 2 times per day