Rapid Household Socio-Economic Impact Assessment Survey

The data collection for the household survey was conducted via CATI (Computer Assisted Telephone Interviewing) method. A system which randomly generates mobile numbers was used as a first step of selection of the respondent. The system has “built-in” algorithms which detect segments of mobile numbers which are in use based on the information given by the operator when the randomly generated number comes out to be non-existent. Throughout years of operation of the system, information has been constantly added to the database about the geographical locations of numbers that have been called for surveys in the past. Hence there is a possibility to predict the “location” of the number that will be called after being generated. This tool ensures the contacts are generated randomly.

Once the number is generated, a screening questionnaire was used to determine whether the respondents are suitable for the sampling requirements:

- Geographical location (municipality)

- Urban/Rural

- Age

- Gender

- Education level

- Employment status

Once the quota was checked and verified, then the main interview was conducted. The interviews for the Household Survey were conducted from 1 May to 9 May 2020. The screening helps the monitoring of data collection phase as it ensures each interview conducted fits in the planned distribution of interviews.

Sample distribution

Sample distribution of interviews is based on official data from Kosovo Agency of Statistics^[1]. The percentages in the table which show the proportion of population (18+ years old) are obtained by Census 2011 data. Also, settlement type, gender and age distribution was based on Census 2011 data, as shown in the tables below.

Based on these data, the interviews per each region were projected:

Table 1. Sample Distribution by Region

Table 2. Sample Distribution by settlement type

Table 3. Sample Distribution by Gender

Table 4. Sample Distribution by Age

The data collection process was closely monitored in order to reach a sufficient distribution regarding the employment status as the unemployed are perceived a particularly vulnerable group. The below tables show the calculations for the number of interviews.

Table 5. Sample by Employment

The margin of error for a total sample of 1,412 interviews in Kosovo is calculated with ±2.6%.

[1] Kosovo Agency of Statistics. Population breakdown by age, gender, settlement type, employment status. Source: https://bit.ly/36ICPCw