The sample design for all nine nations in the Islamic survey is generally identical to that used by Gallup and virtually all other research companies for surveys based on personal-in-home interviews. It is designed to measure the views of the entire national population of each country included in this project, urban and rural, citizen and non- citizen, Muslim or non-Muslim, and is based on a multistage probability sample.
In accordance with the standardized procedures of this sample design, the first stage of the sampling process involved the selection of the primary sampling units (PSUs), that is, the main locations in which the interviews were to be conducted in each of the nine nations.
The selection process began by arraying the population data into strata for those factors for which national census or other official statistics were available, such as the distribution of the urban and rural population.
After the national population data for the nine countries were stratified into groups for the urban and rural populations, the individual PSUs were selected from each urban and rural stratum. Where available, the national population data were also stratified by such other variables as educational attainment and household income.
Because of the inherent difficulties of interviewing in rural areas, the rural population of each country was systematically undersampled relative to the urban population by assigning fewer PSU to rural areas. To subsequently bring the various national samples into line, undersampled rural areas were weighted up to accurately reflect their correct proportions in the total national population. v
Following the selection of the individual PSUs in a given country, interviewers were instructed to conduct from five to 10 interviews in each PSU. This produced national sample sizes varying from about 750 in the case of Jordan and Saudi Arabia to over 2,000 in Pakistan, resulting in a total sample size for the Islamic project of approximately 10,000 individuals.
Here is a list of the total sample sizes for each of the countries in the survey:TOTAL SAMPLE 10,004
Saudi Arabia 754
The next stage of the sampling process dealt with the selection of the households in which the interviewing was to be conducted. To avoid the possible bias of interviewing only the more accessible households, interviewers were provided with a specified starting point from which to commence interviewing. They were then instructed to follow a specified direction or travel pattern from this starting point, conducting an interview at each subsequent household (or every second or third household) until the interviewing assignment was completed.
At each household so selected, interviewers were instructed to select one person to be interviewed using the so-called "Kish Grid." This procedure requires the interviewer to pre-list, that is, to record the age and gender of every member of the household before commencing the interview. The individual to be interviewed is then identified by a mark that has been randomly printed in various positions on the grid.
The Kish Grid system ensures that the household member to be interviewed is selected entirely at random and has an equal chance of being interviewed. It thus avoids the possible bias that can be caused by interviewers interviewing only the most accessible household members.
All 10,000 interviews on which the Gallup Poll of the Islamic world was based were conducted in-person, in the home. The interview, which consisted of approximately 120 questions, required about one hour to administer.
In certain countries, e.g. Saudi Arabia, female respondents were interviewed by female interviewers only.
Handout cards or exhibits were read to those respondents who were unable to read.
Numerical scales were used exclusively to avoid the difficulties of translating verbal scales into a number of different languages.
There were no screens for citizenship or for religion. Each country's sample was designed to represent the entire 18+ adult population living in that country at the time of the interview.
The results of this project are for the most part discussed on a country-by-country basis without specific reference to an overall total. In certain graphic representations the "Total" bar is based on an unweighted sample of all interviews conducted and should be used for illustrative purposes only. Since the nine countries included in this sample are not intended to represent all predominantly Muslim countries across the world, since the residents in these nine countries do not represent all Muslims, and since there are significant variations among the nine countries in the responses to the questions included, we -- as noted above -- encourage readers using these data to examine them primarily on a country-by-country basis.