80% of private households participated in labor survey – Labor Force Survey 2014: Main Findings – Part II

KUWAIT: The Central Statistical Bureau’s Labor Market Information System released its most recent report, the Labor Force Survey 2014: Main Findings (October – December, 2014), just before the Eid Al- Fitr holiday, which provides details, analysis and other statistical information concerning Kuwait’s labor market. Part two explains the methodology used in preparing the report.

The Sample The sample design for the Kuwait Labor Force Survey (KLFS) includes both private households and collective households. The 2011 census was used as the sample frame. The sample was designed with additional households to take into account potential non-response. Due to lack of information about the population of collective households in terms of economic activity, occupation and nationality, all members of sampled collective households have been included in the sample. The number of collective households members amounted, in some cases, hundreds or even thousands. Data have been collection from 10,161 members of collective households with an average of about 73 persons per collective household. It is planned that data from the current survey will be used to enhance the sample design within the collective households in the future.

The questionnaire was developed to collect information from both private and collective households, and the same questionnaire was used for all households. The questionnaire is different from the one used in 2008, and includes revisions to provide information needed to calculate the standard labor force indicators using standard international definitions. The questionnaire was pilot tested multiple times. After each pilot test, an evaluation of the results was performed and modifications were made. Modifications included re-writing questions that were misunderstood by the respondents, changing the order of questions, adding new questions and incorporating concepts that were needed for the Kuwait Occupational Projection System and the Labor Market Information System. After the paper questionnaire was completed, it was converted to the Computer Assisted Personal Interviewing (CAPI) format for collecting information in the private households. This required the creation of variables that were needed for the automatic verifications that can be done with CAPI.

Use of Computer Assisted
Personal Interviewing (CAPI) The questionnaire was converted into a CAPI application for administration in private households. The survey used a computerized system of data collection where field staff directly captured information using tablets during the interview. The tablets were loaded with a data entry application with in-built range and consistency checks, and validation routines to ensure good quality data. Supervisors ran checks on the data while still in the field and sent the data electronically to the Central Statistics Bureau (CSB) Headquarters for additional verification. Every team was facilitated with Internet connectivity while in the field.

Data Collection in Collective Households
The information from collective households was collected using paper questionnaires. In each of the collective households that were included in the sample, all individuals residing in the household were interviewed. This information will be used in designing future surveys to determine if it is possible to interview only a sample of respondents in the larger collective households.

Training of Field Staff
Fifteen days of training were provided to the enumerators and supervisors in how to administer the questionnaire in both CAPI and paper forms. Labor force concepts were discussed and each question was reviewed to ensure that all enumerators understood how to collect the data and what each question provided. Enumerator manuals that were provided for use during the training were also used during the actual data collection so that enumerators could resolve any problems encountered in the field. The training was done separately for the private households and the collective households. The training for the private households included training in how to administer the questionnaires as well as how to use the CAPI application on the tablets.

Field Work
Mobile field teams were used to collect the data from the different sampled areas in each governorate for the private households. Each team included one supervisor and three or four enumerators. The field teams visited each private household up to three times to collect data from all members of the household. Two mobile teams were used to collect the data for the collective households. These households were visited as many times as needed to collect information from all of the residents because collective households could contain hundreds of individuals. Data collection took three months from October 12th 2014 through January 6th 2015.

Data Quality
As an aid to ensuring good quality data, extensive monitoring of the field work was conducted. Monitoring and evaluation guidelines and formats for fieldwork were developed as well. There were three levels of monitoring and supervision. The first level was performed by the Supervisors who provided two types of supervision. Supervisors were responsible for making sure that all households were visited and accompanied enumerators during the interviews to ensure that the questionnaires were administered correctly.

The second type of supervision was a review of the information collected in the questionnaires. Whether the data were collected using CAPI or paper and pencil, the Supervisors reviewed each question to make sure that the data collected was within the guidelines. The third level of monitoring and supervision was from CSB headquarters. After questionnaires were reviewed by the Supervisors, they were sent to headquarters where the staff reviewed them again. Questionnaires that did not meet the guidelines were returned to the supervisors who returned them to the enumerators who returned to the households to fix the problems. In addition, CSB headquarters staff and the World Bank team went to the field to observe interviews and to provide suggestions for improvements in the data collection process. Coding for the occupation and industry was done by coders specifically trained in coding this information.

Two teams of coders were used one for both private and one for collective households. The data from the private households were checked for outliers and out of range values. A list of Kuwaiti employees with monthly salaries found to be less than 400 Kuwaiti Dinar (KD) were prepared with phone numbers and CSB staff called them and double checked their salaries and corrections were made accordingly. In some cases, these apparently low salaries were found to be correct; in one case, for example, the man was retired and then worked in a supermarket “Jamiya” with 390 KD monthly salary. Frequency distributions and cross-tabulations were used to check for out-of-range values, inconsistencies and missing data and corrections were made accordingly. Missing data in main demographic variables, like sex and age, were imputed using data available on other relevant variables (eg. Date of birth, civil number, and names), missing data in industry were imputed, when possible, using data on occupation.

Finally, main labor force indicators were produced and compared to external sources such as 2008 labor force survey , 2011 population census, and 2013 income and expenditure survey. The data from the collective households were key entered and then an independent team reviewed all the data entered by comparing with the original questionnaires. This procedure corrected for possible data entry errors. Additional checks were run to verify information including birthdate, age, education level, date of arrival in Kuwait, the questions regarding whether or not the respondent was working, and to ensure that all sections were completed as required. Because collective households consist of only non-Kuwaitis, all have to respond positively to the questions regarding whether or not the respondent was working. When errors were found, the original questionnaires were reviewed and corrections were made. In two cases, the birth date was recorded incorrectly in the questionnaire so the information from the Civil ID was used to make the corrections. The data set of collective household information was originally stored separately from the data set of private household information. The variable names were different from those used for the private households. The variables in the collective household data set were renamed to match those in the private households and the two data sets were combined into a single data set. Variables were created, “derived variables”, to assist in the analyses of the data by CSB. The derived variables are combinations of the information as collected in the questionnaire that can be used to determine the labor force indicators needed for the analyses. These variables included such items age group variables; dummies for working, not working and searching for jobs; dummies for working age or not (15 years of age and older); and others.

Response Rates
The response rate reached about 80 percent for private households and 74 percent for collective households.

Sampling Errors
Sampling errors were calculated for the main indicators. Calculations included standard errors, relative errors, 95 percent confidence intervals, and design effects. Standard errors for the main indicators ranged between 0.003 and 0.015

This article was published on 22/07/2015