Method

Method of Sampling: ICT Usage in Households and by Individuals Survey is applied to sample households selected by a 2 stage stratified clustered sample. In first stage; in all areas in the frame is divides into clusters (blocks) of approximately 100 households each. These were selected as probability proportional to size (PPS) .In the second stage; the sample addresses are determined from selected clusters (blocks) by using systematic selection method.

Statistical Regions Level 1 are used as strata.

Estimation level: The sample size of the study is calculated to give estimation by Turkey in general and Classification of Statistical Regions Level 1. No substitution is used in the study because the non-response rates are taken into consideration when calculating the sample size.

Weighting: Weighting is a method used to obtain parameters from the data set resulting from sampling so as to represent the universe. These parameters include design weights, adjustments for non-response, external distribution checks and ultimate multiplying factor. Reaching accurate estimates depends on a step-by-step and careful process of weighting.

Data Sources: Data were collected from the households which were selected by defined sampling method. Statistical unit used is “household”. Demographic information (age, sex) is asked to all members of the household. But, questions on educational status, labour force status and ICT Usage are asked for persons aged 16-74.

Data Collection Method: Computer aided personal interview

The System of Data Collection: Questionnaires are prepared based on model questionnaire which prepared and recommended by EUROSTAT. Model questionnaires are adapted to the conditions in Turkey and also regulated in accordance with national institutions’ needs.

Reference and application period: The survey carried out in April of each year. Reference period of survey are defined according to the compilation period.

Field application and non-response rate: Non-response rate is calculated by the proportioning of real non-response reasons such as refusing, not at home to the households number after deducting of sample errors or units out of coverage from total sample.

Top of page