4. Chapter Four: Main Study Methodology
4.1 Introduction
This chapter deals with the research methodology and the aim of the chapter is to discuss the practical approaches that were used in this study. So, this section introduces and covers study type, study design, sampling, pilot study, data collection instrument, data collection processing, the statistical analyses proposed, and the limitations of the study.
4.2 Study Setting
The study took place in Riyadh, Saudi Arabia. The population of Riyadh is estimated to be about five million. The city is the capital city of Saudi Arabia and it has a fast rate of growth. Gathering comprehensive data about health related behaviours of adolescents in Riyadh will attract more attention than data from any city and therefore will have major implication on the society.
4.3 Study Population
The specific population selected for sampling in this study survey is male adolescents attending public and private high schools in Riyadh. Due to political, religious and cultural reasons, female adolescent students are separated in schools from male students; therefore, female students cannot be reached by the researcher and will not be included in this study, although there is no doubt that their future role in community and family health is fundamental. The majority of the students are usually aged between 15, and 18 years, in their tenth, eleventh and twelfth grade of their school years in Riyadh City, Saudi Arabia. The reasons for selecting this age group are as follows:
Adolescents are highly exposed to risky behaviours that may influence their health directly or in the future.
The behaviour changes during adolescence often determine the lifestyle and health habits of adulthood, creating long-term health implications.
83
The majority of health risk behaviours, which contribute to the leading cause of mortality and morbidity, often can be prevented by adopting healthy behaviours and effective prevention.
Many high-risk behaviours among adult first began during adolescents and become well-established patterns of behaviour by young adulthood (Park et al., 2006).
4.4 Study Design
This study employed a cross-sectional design to investigate adolescent health- related behaviours. This type of research was appropriate to answer the research questions. The data collection took place between February and April 2008.
4.5 Study Tool
The method to collect the data was a self-completion questionnaire. The questionnaire approach is a very common technique used for collecting and one of the major approaches to collect primary data. Questionnaire is usually distributed by one or more of variety of methods which include posting, by hand, phone or internet. The survey questionnaire is useful in the process of collecting data from a large number of people and this is useful especially when resources and time scale are limited.
Although, questionnaires are widely used to measure health-related behaviours, it is important here to mention that there is potential for bias using self-reported questionnaire to assess behaviours such as physical activity. It has been reported that an accurate assessment of physical activity in epidemiological studies is critical to the evaluations not only of health outcomes of physical activity but also of potential confounding effects by this complex behaviour (Orsini et al., 2008). Physical activity dimensions include intensity, frequency, and duration, which together make up the total volume of activity. Another important dimension of physical activity is type or mode, e.g., walking and cycling (Corder et al., 2008). There are measures that the researcher uses to validate and measures physical activity. Accelerometers and machines/equipment records are used to validate physical activity (Orsini et al., 2008). There is evidence, however, that the validity of shorter physical activity questionnaires can be acceptable to classify persons according to physical activity levels (Mader et al., 2006). Self-report instruments and movement sensing are
84
currently the most commonly used methods for the assessment of physical activity in epidemiological research (Corder et al., 2008).
4.6 Questionnaire Items
The questionnaire aimed to collect comprehensive data from school students regarding behaviours, knowledge and attitude. The questionnaire items (105 items) were based on international validated questionnaires which have been used worldwide (See Appendix C for the items sources). The questionnaire include students’ personal characteristics (age, school grade level, type of school, home district, family structure, fathers’ and mothers’ level of education), self-reported health and school performance and school satisfaction, dietary behaviours and height and weight, oral health hygiene, physical activity behaviours, smoking behaviours, injuries and violence, bullying and abuse, factors associated to psychological health, and driving and safety behaviours (See Appendix A: 1 for English version and Appendix A: 2 for Arabic version).
4.7 Sampling Strategy
In this study I applied multi-stage, stratified sampling among 10th to 12th grade high school students in Riyadh Capital city, Saudi Arabia in 2008. Stratified sampling by school type, school district, and class level of high schools, to yield a representative sample of students to make sure that I have recruited representative sample size from the whole of Riyadh city. Schools were divided into two types: those in the private sector (about 44%) and those in the public sector (about 56%), but 72 % of the total high school students belong to the public schools, whereas 28% of them in private schools (see Table 4.3 below). Moreover, in order for this study to be relatively representative Riyadh was divided to five districts (north, south, east, west and central), so as to ensure a distribution of adolescents and schools from all areas of the city. Also, to avoid any potential bias by not including students from different areas. However, no previous study has investigated the difference between these districts and there is not enough information about any cultural or socio-demographic status differences. Two schools (one public and one private) were selected, using random numbers from each district, providing 10 schools in total from which to
85
sample classes and pupils. After selecting the schools from each district, they were contacted and informed about the study and procedure and they were asked to participate. From previous experience with the permission from the Ministry of Education, schools usually tend to participate.(See table 4.1).
Table 4-1. Number of schools in each district
District Number of public schools Number of Private schools Total North 1 1 2 South 1 1 2 East 1 1 2 West 1 1 2 Central 1 1 2 Total 5 5 10 4.8 Sample Size
Statistical advice was obtained before collecting data. Based on the pilot study, the numbers in Table 4.2 are the overall sample sizes required to estimate a prevalence of 20% to the stated precision or accuracy (i.e. precision of 2% would translate to 95% confidence values for the prevalence estimate of between 18% and 22%, precision of 3% to prevalence between 17% and 23% and accuracy of 5% would change this range to 15% to 25%). Table 4.2 below shows that when the level of accuracy increases, the sample size required is much greater.
Table 4-2. Sample size needed for different levels of precision
Precision Sample Size
2% 1,494
3% 464
5% 245
I decided on a sample size of over 1,494, which provides 2% precision, offered a reasonably high level of precision for prevalence estimates and provided reasonable power for subsequent multivariable analysis. The aim was to sample from schools in line with the population proportions and average class sizes (Table 4.3); taking approximately 72% of the sample of adolescents from public schools and 28%
86
from private schools. This means that the sample would be approximately 1080 students from public and 420 from private schools.
Table 4-3. Distribution of Public and Private Schools in districts of Riyadh, and numbers of classes’ students within them (Sampling Frame)
Type of Schools Schools N. % Classes N. % Students N. % Public School 95 56 1477 64 38604 72 Private School 75 44 834 36 15337 28 Total 170 100 2311 100 53941 100
(Numbers obtained from Ministry of Education in Riyadh)
An equal number of schools and class were sampled from each type of schools and districts as I do not have the detailed population information available about the number of students in each district, table 4.4 shows all the details.
Table 4-4. Structure of study sample by District, school type, classes and number of students
District Class level Type of School
Students No (%) Public School Classes Students No. No. Private School Classes Students No. No. North 10th grade 3 78 2-3 36 114 11th grade 3 78 2-3 36 114 12th grade 3 78 2-3 36 114 Total 9 234 6-9 108 342 (20) South 10th grade 3 78 2 36 114 11th grade 3 78 2 36 114 12th grade 3 78 2 36 114 Total 9 234 6-9 108 342 (20) East 10th grade 3 78 2 36 114 11th grade 3 78 2 36 114 12th grade 3 78 2 36 114 Total 9 234 6-9 108 342 (20) West 10th grade 3 78 2 36 114 11th grade 3 78 2 36 114 12th grade 3 78 2 36 114 Total 9 234 6-9 108 342 (20) Centre 10th grade 3 78 2 36 114 11th grade 3 78 2 36 114 12th grade 3 78 2 36 114 Total 9 234 6-9 108 342 (20) Total 45 (68) 30-45 (32) 1710 (100)
87
At the outset, although I knew that 56% of schools are public and 44% private. I also knew that class numbers, size and number of students in classes were different in the two types of school, students in public school classes range from 25-35 (in some school 50 students in one class) and around 15 in private schools. Given the different class sizes, randomly sampling nine classes from each public school (three from each grade, 10-12) and six to nine classes from each private school (2-3 from each grade, 10-12), resulted in approximately 1170 students from the five public schools and 540 from the five private schools. This provided approximately 68% public and 32% private pupils, 45 (60%) public school classes (of approximately 25 pupils each) and 30 (40%) private school classes (of approximately 15 pupils each) and five each of public and private schools (50% each). These proportions compared reasonably well with the proportions of pupils, classes and schools in the public and private sectors in Riyadh (see Table 4.4), although there was a slight over- representation of the private sector (these figures are only estimates and it was possible to add more classes and student if needed. This sampling strategy gave an expected sample size of 1710, allowing for about 12% non-response to the survey. However, based on the pilot study it was assumed that the response rate would be high since it was 99% and 96% in the first and second phase of the pilot study. In general, the sample size was large enough to enable acceptable precision of prevalence.
4.9 Survey Procedures
The survey procedures were designed to protect student privacy and allow for anonymous participation. I distributed and collected the questionnaire after completed by the students in classes and under my supervision. The aim and the objectives of the study were explained to the students and for high response participation and accurate results. I explained and ensured confidentiality and motivated students to participate. Students were assured that the survey was anonymous and that no-one at their school would see their questionnaire. Students were instructed not to talk to anyone or look onto anyone else work-everyone is entitled to their privacy, and to think about each question and answer it as it applies to you. They were also told that there was no right or wrong answer, and not to answers any question in a way which they think they should answer it to impress other. To insure the privacy, students were separated in
88
the class. Teachers were asked to leave the class and the researcher was present in the room at all times to answer any questions and to ensure that the instructions were followed. Anyone who did not participated in completing the questionnaire was asked to remain silent and sit quietly. Students recorded responses on the questionnaire booklet. When the students complete their questionnaire they handed it to the researcher. Then, their height and weight was measured in private and recorded on the back of each student questionnaire respectively.