Questionnaire survey
Guided by an NHS Grampian steering group, a questionnaire was designed for the quantitative survey based on the Theory of Planned Behaviour (TPB) [
16] and Social Cognitive Theory (SCT) [
17], both commonly used for health behaviour change. The questionnaire included demographic factors including self reported height and weight; three PA behaviours (active exercise, hours of TV watching and time spent on computer/games console), attitudes, subjective norm, perceived behavioural control (PBC), intention towards PA and barriers and facilitators for achieving recommended levels of PA.
For active PA behaviour, participants were asked on average the number of days per week they would normally be moderately physically active (that is exercise sustained for many minutes, without exhaustion or extreme fatigue that increases the breathing and heart rate, such that the pulse can be felt with increased warmth and possible sweating) as recommended by National guidelines. Further the guidelines suggest that adults should achieve this a minimum of 30 minutes a day on at least five days or more a week for general health benefit [
18]. This is the definition used in this study denoted here as being ‘adequately physically active’. Physical activity might include sports, recreational activity and general active living. Those achieving the recommended levels of PA only up to 4 days a week were grouped as being ‘inadequate exercisers’ while those managing this on 5–6 days per week were grouped as ‘adequate exercisers’. Two questions addressed sedentary behaviour. These were the number of hours spent each day watching TV and similarly on computer/games consoles. For each, the response originally had five options but these were compressed into three categories ‘Less than half an hour’, ‘1-4 hours’ and ‘>4 hours’.
Attitudes toward PA were assessed using four concepts – difficult/easy, relaxing/stressful, not enjoyable/enjoyable and unhealthy/healthy. These were assessed by a 5-point scale 1 (disagree) up to 5 (agree) but later regrouped into ‘positive’, ‘neutral and, ‘negative’. A question on PBC asked about the confidence young people had about being moderately physically active. This was coded from 1 (Not very confident) up to 5 (Very confident). The question about young peoples’ intention about being physically active was another 5-point scale, and remained as such, 1 (disagree) up to 5 (agree).
In addressing the facilitators, participants were asked if they would consider doing more exercise for any of 11 reasons given in the question each with a ‘yes/no’ option. Three of the statements related to ‘health’ (improve health, lose weight or maintain healthy weight, and feel fit), one was to improve appearance, three statements referred to relaxing (have fun, socialise, to relax or feel better), one was about competing (to win), two were related to the subjective norm (to please family/friends or to impress) and the last one was ‘others’. Apart from the subjective norm statements, the rest were grouped into four categories: health, appearance, relaxing/socialising and winning.
Similarly, for barriers, the original question had 19 statements, (‘yes/no’ response options), where each statement represented a reason preventing them from taking more exercise. After inspection, these statements were regrouped into 12 barrier classes: PA with the opposite sex; competition; a lack of privacy, information, company, facilities, time and money; having a disability; feeling that they do enough exercise already; bad weather; and finally a poor choice of activities.
These compressed facilitator and barrier classes required revised coding. Classes that combined 3 statements were coded: ‘Strong (facilitator or barrier)’ if all three statements were ‘yes’; ‘Mostly yes’ if two were ‘yes’; ‘Mostly no’ if two were ‘no’ and ‘Not a (facilitator or barrier)’ if all three were ‘no’. Similarly when 2 statements were combined, the coding was revised to: ‘Strong (facilitator or barrier)’ if both statements were ‘Yes’; ‘Not a (facilitator or barrier)’ if they said ‘no’ to both and ‘Mixed’ if they ticked ‘yes’ to one and ‘no’ to the other.
Recruitment of the sample was only possible through an institutional or global approach, since direct access to young people was not permitted. Consequently, the questionnaire was sent electronically via institutes to all university/college students in the Grampian area in 2007–08. They were asked to complete the questionnaire if they were between 18–25 years of age (those not in this range were filtered out). To capture young people not in education, employment or training (NEET), hard copies were sent to co-ordinators of the NEET groups in the Grampian area to be completed by participants at their groups meetings. To capture those at work and young people who may not attend the NEET group sessions, a postal hard copy of the questionnaire was sent to a 2% random sample of 18–25 year olds (n = 1800) in the community using the Community Health Index (CHI), a computer based population index used by NHS Scotland.
Focus groups
Using the website for the university, young people between the ages of 18–25 years were invited to take part in focus group discussions using a ‘pop up’ advert. An institutional e-mail with an information letter was also sent to all the students. All the NEET groups and other youth groups/clubs in Grampian area were again approached through the group co-coordinators and given an information letter. Recruitment was also conducted through local media (radio). Seven focus groups resulted with a total of 26 participants from the same population as the quantitative survey. Focus groups gather participants’ attitudes, feelings, beliefs, experiences and reactions in a collective way, not feasible using other methods, for example, observation, one-to-one interviewing or questionnaire surveys [
19]. A topic guide, based on issues identified from the survey and grounded in TPB and SCT, facilitated discussion and participants were encouraged also to discuss issues of concern to them, ensuring an inductive approach. Question addressed in the focus group discussions related to actual physical activity behaviour, the importance and perceived relevance of regular exercise at this stage in life and in the future, positive and negative outcome expectations of regular exercise, perceived and actual barriers to undertaking regular exercise, self-efficacy and exercising control over undertaking regular exercise, and finally factors that might facilitate and encourage regular exercising. A purposive sampling method was used based on the previous survey results (age, level of education, employment status) with the intention of obtaining a balance in terms of socio-economic groups. The focus group data collection was terminated after obtaining saturated data from a wide range of relevant groups. A written informed consent was obtained from the participants at the time of the focus groups ensuring anonymity and confidentiality. Ethical approval was obtained from NHS Grampian for the quantitative study and from university ethics committee for the qualitative study.
Multistage modelling
After identifying significant variables from the univariate analyses detailed above, a strategic stepwise methodology was developed for modelling executed in three stages. Initially, behavioural intention was modelled against demographics/risk factors and each TPB construct (attitudes, subjective norm (SN) and perceived behavioural control (PBC)). From this, only significant variables were considered in a Combined Intention Model. Secondly, each PA behaviour was separately modelled with 1) demographics 2) constructs of TPB 3) intention and 4) barriers and facilitators. Finally a combined model was developed using Forward Stepwise Logistic Regression to predict each of the physical activity behaviour, based on only those significant variables from the previous stages. The final model(s) above (one for each PA behaviour) provided the most important associates for PA behaviour.
‘Framework analysis’ was used to analyse focus group data in a systematic way [
20]. Framework analysis uses a thematic framework to classify and organise data according to
apriori themes and concepts and also emergent categories from the data. It allows transparent data management and comparison of data between groups. As each group was analysed, themes were added and amended until an agreed framework of themes was developed. Data was therefore explored within a common framework that was both grounded in the theory and informed by participants’ views and experiences.
After analysing the quantitative and qualitative data separately using their respective appropriate analytical approaches, a ‘
side-by-side comparison’, method was used. This enabled the comparisons and synthesis of the results from both quantitative and qualitative components [
21].