Method

This report describes the findings of a representative survey of Qatari teens, 13 to 20 years old. It was conducted from April 22 through May 17, 2017 with 1,117 Qatari nationals attending grades eight to ten at 36 independent (public) and international (private) schools. The survey was administered in Arabic by the Social and Economic Research Institute (SESRI) staff at Qatar University and used the platform BLAISE. The data were gathered on laptop computers with a research assistant interacting with the participants, solely for the purpose of providing instructions and answering questions. Parental and teen consent were obtained beforehand.

Sample design  

In this study, the target population includes all Qatari students in preparatory and secondary schools, from grade 8 to grade 12. The sampling frame was developed by SESRI based on a comprehensive list of all public and international schools in Qatar provided by the Supreme Council of Education.

Almost 30,000 Qatari students were eligible for the survey. The schools in our sample were selected from the 114 schools with a reasonable number of Qatari nationals. Random systematic stratified sampling was performed to select a representative sample of 43 schools. Seven of those schools refused to participate, which left us with a final sample of 36.

Based on the information about the distribution of genders and grades, schools were divided into relatively homogeneous subpopulations (i.e., strata). Inside each stratum, students were randomly selected following a two-stage sampling process: In the first stage, the school was selected with a probability proportionate to its size. In the second stage, we randomly selected classes in that school. Ideally, all 1,698 students in those classes were supposed to participate in the survey.  However, 374 were absent during the time of our survey, and 106 refused to fill in the questionnaire. So, 1,218 students participated.

The sample

In total, 1,218 high school students at Qatari schools were interviewed. 1,117 of them were Qatari nationals--they are the basis of our analyses. Eighty-five percent of those respondents attend an independent (i.e., public/government-owned) high school. The other 15 percent were students from international (private) schools. The Qatari nationals consisted of 49 percent females and 51 percent males. Each of the grades (8 to 12) was represented by at least 15 percent of these respondents (Table 1). Respondents’ ages ranged from 13 to 20 years (Table 2), with a median age between 15 and 16 years.

 Survey Design

 Our questionnaire was prepared in several steps. First of all, whenever appropriate, we used questions from the 2015 U.S. study Teens, Health, and Technology (see above) to make intercultural comparisons possible. The questionnaire covers the following categories of questions:

Sources of health information and use of health-monitoring tools:

  • Use of communication channels of all kinds for health information

  • Trust in these sources/satisfaction with them

  • Criteria for selecting one’s sources of health information

  • Reasons for/purposes of going online for health information

  • Ways of finding health information online

  • Encountering specific health campaigns

  • Encountering problematic content online

  • The role of health classes at school

  • Willingness to post health-related questions online

  • Use and usefulness of health-related apps, games and health trackers

  • Perceived changes in one’s health behavior due to health information sources and tools

  Background information about our respondents:

  • Health issues that Qatari teens find personally important

  • Perceptions of which health issues one’s friends find important

  • Satisfaction with one’s body

  • Physical activity and reasons for being active

  • Eating behaviors

  • Traumatic experiences

  • Perceived self-efficacy, i.e., confidence in one’s ability to change behavior.

A first draft of the questionnaire was developed and translated into Arabic. Cognitive interviews followed in order to test the validity of the survey questions. The students that participated in these interviews were selected from across the target age/grade range from grade eight to grade twelve. Fourteen females and ten males were interviewed. The interviews were conducted face-to-face and took about 40 to 60 minutes to complete. Open-ended questions ensured that:

  • the final survey included any health issues and information tools relevant to our target group that we had not initially included,

  • our questions and response options were understood properly,

  • our respondents did not refuse to respond to questions because they were regarded as inappropriate or irrelevant.

Based on the findings of our 24 cognitive interviews, we constructed a revised version of the questionnaire. This version was programmed and tested in a pilot study under real-life conditions through a computer-aided self-administered interview. Thirty-four pretest interviews were completed from two classes at two schools from the two corner grades of our target group—grades eight and 12. The respondents were selected to represent the gender and age group distribution of the target population. Based on the results, the questionnaire was modified, but also shortened--to take about 30 minutes to fill in (see the appendix of this report for the final questionnaire).

Survey administration

The survey was programmed in BLAISE and administered using laptop computers. The data collection took place between April 22 and May 1, 2017. At the schools that agreed to participate in the study, opt-out consent forms were sent to the parents of eligible student participants. Signed student assent was collected prior to administering the survey. Trained data collectors provided students with laptops and instructions on how to complete the survey.

Weighting the data    

For the analysis of our data, the sample design is taken into account to ensure that statistical estimates are unbiased and efficient. Particularly, a weighting variable was created for the selection probability of each respondent and for non-responses. The weights were also calibrated to align our results with population estimates. The “raking” method was used for this calibration to adjust the weights in the sample--so that the proportions of the adjusted weights for certain characteristics (e.g., proportion of students by grades) agree with the corresponding proportions of the population.