Methodology

Sampling

Sampling for the Opinion Leaders’ Panel was carried out by the Central Statistical Office (CSO) who have provided the following details of how it was carried out: The methodology employed for this survey followed the survey design for the CSO Labour Force Survey. The CSO Labour Force Survey is a cluster survey in which households are selected in two stages.

At the first stage, Enumeration Districts (EDD), which are small geographic areas, are selected with probability proportional to ‘size’ (PPS). The measure of size used is the number of households. At the second stage, households are selected with probability inversely proportional to size (PPS).

This pps-pps design is intended to select approximately five households in each selected ED in the Island of Trinidad and approximately three households in each selected ED in the Island of Tobago. However, because of growth in EDD over time, one anticipates that larger number of households will be selected in order to maintain the fixed probability of 1 in 120 households being selected in the case of Trinidad and 1 in 75 households in the case of Tobago. Furthermore, it is to be noted that the sample is self-weighted within each Island Stratum, that is, each household has approximately the same chance of being selected within each Island Stratum.

The sample generated by the above procedure is approximately 2,800 households. The sample is stratified by and proportionately allocated to Cities, Boroughs and Regional Corporations, in the case of Trinidad, and by Parishes or combination of Parishes in the case of Tobago.

The sample of households drawn for the Survey of Public Service was generated from the sample of EDD, which were selected for the Labour Force Survey conducted during the first quarter of 2002. However, to avoid returning to households already visited during the Labour Force Survey (that is to say, to avoid respondents’ burden), households were selected using different random starts within each selected ED.

Response Rates

Interviews were carried out face-to-face, in home, between 15th July – 9th By August 2002. 2,747 successful interviews were completed. The overall response rate was sixty-nine percent (69%). A preliminary analysis of the data was done when the achieved sample was 2,500. This showed under representation of young adults aged 18-24.

The Young Persons Booster

At this stage the remaining sample was restricted to young persons in this age category. The method used to select this “booster” was essentially the same as for the main sample. The ED maps provided by the CSO were used to select households using the provided skipping interval. Within each household only eligible persons aged 18-24 were listed and the person selected was based on the Kish grid.

Weighting

Given that the young persons booster consisted of less than 300 persons, the impact on the overall sample was not significant enough to avoid weighting. Indeed on detailed examination of the demographic profile of the sample, it was found that the sample was also unrepresentative in terms of race. Weights were based on the data from the 1990 census since details from the 2000 census were unavailable at the time of the survey.  For purposes of analysis a compound weighting scheme was devised which combined the age and race population data. The weights applied were as follows:

Services | About Trinidad and Tobago | Help
Copyright © Government of the Republic of Trinidad and Tobago