Fieldwork and sampling methodology

The success of any survey requires an effective survey tool – neutral and objective questions comprehensible to respondents from any demographic background – and a carefully designed sample. Of the two, the sample is the most important: choosing the wrong sample can produce incorrect or inaccurate results, even if the questions themselves are neutral and precise. Broadly speaking, there are two kinds of sample:

  1. Probability samples based on random selection from the population.
  2. Non-probability samples based on targeted selection of individuals from the population.

It goes without saying that the second type of sample is not representative of the target population because it is not random and focuses on particular groups over others, and its ability to cover variations within the population is limited. To obtain results representative of the target population, it is necessary to take a random or probability sample in which any given individual from the population has the same chances of appearing in the sample.

The AOI's methodology is naturally based on random sampling, since its aim is to create a representative picture of opinion in Arab countries. Because of the complex and heterogenous societies that it surveys, a regular, proportional, random multistage stratified cluster sampling technique is used in all AOI surveys. This guarantees that surveys will be representative of the constituent units of each Arab society and of the effects of this differentiation on attitudes towards various key issues.

This sort of sample involves using more than one sampling style in order to make it both as broad and as representative as possible. In order to produce a confidence level of no less than 96% and a maximum margin of error of ±3%, no sample consists of less than 1,200 respondents.