How representative is the TGI sample?

The validity of TGI data rests on the very precise way in which the sample of respondents is selected, rather than on the size of the resulting database. Each year approx. 2,000 adults in Northern Ireland and 3,000 in the Republic of Ireland return a TGI questionnaire. Although a small percentage of the total population, carefully controlled sample design ensures that results are representative to a high level of detail.

Sample design

TGI selects respondents through a carefully controlled design that closely approximates the true random sample of sampling theory.

In the case of the Northern Ireland survey, respondents are recruited to TGI by way of face-to-face interview.

As a first step, recruitment is achieved by allocating interviewers to a set of defined sampling areas (or PSUs). These will be annually selected based on census characteristics to ensure coverage of the full range of different area types in the country. Interviews within each sampling area will then be conducted according to a carefully controlled quota system that ensures a representative final sample on a number of key socio-demographic criteria.

In the case of the Republic of Ireland survey, recruitment is attained by way of both face-to-face interview and through a CATI RDD methodology. As is the case in a F-F recruitment system, CATI RDD makes use of a geographically-representative set of ‘root numbers’ from which initial contacts are derived. A set of secondary quota controls then ensures that the final recruited sample is balanced in terms of both demography and regionality.

In the case of this dual methodology, quotas on each methodology can be adjusted independently to ensure that the optimal contact method is used for different sub-groups. In this dual scheme, F-F recruitment will usually focus on younger demographic groups – who may not be as receptive to telephone contact. Taken together, however, both methodologies form a strong complement that allows the ROI TGI to accurately represent all demographic groups.

Weighting

The weighting process adjusts demographic cells to known population levels. Correction factors are applied to adjust key quantities. These are calculated for each individual separately based on the following demographics:

Sex within Age within Region, Sex within Social Class within Region; and for Female Housewives on Working Full-Time and Presence of Own Children.