World Values Survey 2005, Wave 5

The Mexico 2005 survey used a multi-stage sampling procedure. Interviewers selected an adult using a random selection method. However, we also employed control quotas according to sex and age this practice was more common in rural areas, where the male population is more difficult to find at home during interviewing hours. Interviewers made sure that respondents were at least 18 years old, that they lived in the selected household. Interviews were all conducted in-home.

Remarks about sampling:
The first stage was the selection of polling points based on the list of electoral sections defined by the Federal Elections Institute. The sections were previously stratified as urban (70 percent), and rural and mixed (30 percent). Each section is relatively homogeneous in size, with about 1,092 registered voters in 63,810 sections that cover all the countrys adult population. Respondents included, of course, also adults nonregistered as voters. We selected 130 electoral sections in a systematically random fashion in each stratum, based on the list arranged proportionally to size of population. In the second stage we selected the household with a systematic random selection, based on a standard strategy of walking around the housing districts selected in the sample. In the third stage, interviewers selected an adult respondent in each household. We used control quotas based on sex and age in districts where random selection of interviewers was disproportionately leaning towards a specific group. Each polling point represents 12 interviews, and quota control established that 6 were male respondents and 6 women respondents, to ensure an appropriate distribution, especially in areas where some specific group is difficult to reach during the hours of interviewing (i.e. rural towns and communities). The Mexican countryside presents problems, for example, to reach male populations during the day in their households. In terms of age, the following quotas were employed where needed: 4 out of 12 were 18 to 29 years old; 5 out of 12 were 30 to 49 years old, and 3 out of 12 were 50 years old or older. We substituted four of the originally selected addresses; three in rural areas and one in an urban area. In the rural cases, the interviewers were not able to get to them because of the absence of roads and transportation. In the urban case, the polling point was substituted because the neighborhood represented serious safety problems at the time of the survey. All the polling points were substituted with addresses with the same socioeconomic level, in the same region, state and electoral district. Substitution of households and respondents were also employed, in the cases where either one of them was registered as a no contact or a refusal and remained under those categories after call backs or returns. Interviewers kept record of non response items (no contact, refusals, suspension) at every time.

The sample size for Mexico is N=1560 and includes the national population aged 18 years and over for both sexes.