The Core Welfare Indicator Questionnaires (CWIQ) survey has been developed over the past 5 years and mainly deployed in African countries as an attempt to improve the timeliness and reliability of poverty monitoring.
The survey was designed to collect information needed to identify and classify target groups and provide basic welfare indicators. It was in addition meant to collect information to measure access, utilisation and satisfaction with social services. The survey was developed by a group of donors and institutions including the World Bank, DFID, the ILO, UNICEF and UNDP. It adapted the technique of optical reading which permits fast processing of the data and thus timely release of the results.
The St. Lucia CWIQ survey objectives were as follows:
• To introduce the survey to the Caribbean Region as an improved and affordable poverty monitoring tool.
• To test the suitability of the poverty predictors for the Caribbean.
• To provide welfare indicators for monitoring poverty and welfare programmes in St. Lucia
Kind of Data
Sample survey data [ssd]
Unit of Analysis
Unit of Analysis
Producers and sponsors
Authoring entity/Primary investigators
Department of Statistics
Goverment of St. Lucia
United Nations Development Programme
Department for International Development
International Labour Organization
United Nations Children's Fund
The CWIQ sample is based on a two stage systematic random sample, the first stage being the Enumeration District (ED), and the second the household. A full list of EDs is available from the 2001 Population Census arranged by 10 Administrative Districts with household numbers and details of the percentage of households belonging to professional / office based workers as well as the percentage engaged in agriculture. St Lucia comprises around 47,275 households and the average ED consists of 118 households.
The CWIQ sample of EDs was drawn from a list of all 401 EDs arranged (i) by District and then (ii) by percent of office employees. This ensured a broad geographical coverage and a balance of urban and rural households. The sample of EDs was drawn using probability proportional to size (PPS) by accumulating household populations and selecting a fixed sampling interval with a random start point.
The objective of the sample is to obtain a nationally representative result with minimum cost and in the available time frame of 10-14 days. The original intention was to sample of 1750 households in 220 EDs, working at 8 interviews per day. However, it was recognised that this work rate was not achievable given household accessibility. A work rate of 3-4 per day per enumerator was considered feasible. To enable rapid completion, the number of EDs was set at 109 and the number of households per cluster increased to 12, giving a total sample of 1308 (around 3% of the national population of households). This is expected to be large enough to generate statistically reliable results for the main reporting disaggregations (poverty quintile, urban-rural, age categories, etc.). Only for District tabulations, where in the smaller Districts1 the sample drawn is quite small, is there a possibility of rather high margins of error due to small samples.
The resulting sample contains 62 urban and 47 rural EDs.
Dates of Data Collection (YYYY/MM/DD)
Mode of data collection
Several control measures were organised to ensure the quality of the information collected. First, the supervisors were instructed to perform on spot interviews and to complete a review of all filled questionnaire. In the supervisors' manual, a special section with detailed instructions on verification and editing of questionnaire was added to serve as a guide for quality control. Second, the co-ordinators were also instructed to verify 10% of the filled questionnaires although in practice they verified most of them. In addition, the DOS senior staff carried out further checks on the completed questionnaires before they were sent for data processing. A full meeting with all co-ordinators, monitors and data processing team was held at the end of the first week to review the survey progress. Necessary instructions and recommendations made during the meeting were conveyed to the field staff for implementation.
Department of Statistics
The data processing system for CWIQ surveys is tightly integrated with all survey development and implementation activities. This is necessary to achieve the objective of producing reliable survey results as quickly as possible. The specific objectives of the CWIQ data processing system for the St. Lucia survey are to:
- Produce a survey database that is complete, consistent and has all known anomalies documented.
- Produce survey results (data summaries, cross-tabulations, core indicators and sampling errors) within a week of the end of field work.
- Facilitate the preparation of the survey report within three weeks of the end of field work.
- Create a survey archive with all data, documentation and reports for dissemination to data users as soon as the survey report is available.
The CWIQ data processing system has four principal functions: (1) data entry, (2) data validation and correction, (3) preparation of survey results, and (4) creation of the survey archive. Data entry consists of converting the information in the survey questions to a machine readable form for processing in the subsequent phases.
The CWIQ data processing system is designed to process the generic CWIQ questionnaire. Like the generic questionnaire, the generic system is never used for a real survey, but is modified to fit the specific needs of each survey. As the generic questionnaire is adapted to local requirements, the data processing system is adapted for changes to the questionnaire and survey outputs.
Data entry for CWIQ surveys is done by optical scanning of the questionnaires using TELEform, a form design and image processing system. As the questionnaire is modified, the processing specifications are also adapted. These specifications are the basis for the computer programmes that comprise the data processing system.
For the St. Lucia survey the questionnaire was modified in three iterations. Likewise, the data entry and validation parts of the system were modified for each version of the questionnaire. The first modifications were tested with questionnaires completed during the field pre-test of the questionnaire. As a result of these changes, the questionnaires modified for interviewer training.
During scanning, the scanner creates an image of each page of the questionnaire. The scanning software subsequently evaluates the scanned images and questionnaires with possible errors are subject to verification by the data entry operator. Typical errors include unidentified pages that can not be evaluated; questionnaires with missing or mismatched pages, unrecognisable hand printed characters or bubbles which are not completely shaded. The time required for image evaluation and subsequent verification depends on how well the questionnaire was filled in. The interviewers completed hand printing and shading exercises during training to einsure that questionnaires were properly filled in.
After all potential errors for the EA were verified by the data entry operator, the data from the questionnaires was copied to transfer to the questionnaire database maintained on the network file server. The output of the scanner on the Windows Server was checked for certain errors and the questionnaire data was not transferred to the database until all such errors were corrected. The questionnaire data for the EA was then validated to insure that all questions had valid responses and that the responses were logically consistent. Any errors detected during validation were printed on a validation listing along with the questionnaire data. This list was compared to the questionnaires to determine the reason for the error and how to correct it. Corrections were recorded directly on the validation list which was use to enter corrections in the database. The validation was repeated after the corrections were entered until all errors were eliminated.
This process continued as the fieldwork progressed and was finished two days after fieldwork ended. When the questionnaires from all enumeration areas had been validated, the entire database was validated to insure that no errors had been overlooked.
The database updated with the analysis variables was used to prepare data summaries of the questionnaires and to produce cross-tabulations defined in the CWIQ tabulation plan. The summaries and the Tables were available two days after the final data validation. Sampling errors for the core welfare indicators were computed as the final step of the data processing system.
Use of the dataset must be acknowledged using a citation which would include:
- the Identification of the Primary Investigator
- the title of the survey (including country, acronym and year of implementation)
- the survey reference number
- the source and date of download
Department of Statistics, Goverment of St. Lucia. St. Lucia Core Welfare Indicator Questionnaire 2004. Ref. LCA_2004_CWIQ_v01_M. Dataset downloaded from www.measuredhs.com on [date]
Disclaimer and copyrights
The user of the data acknowledges that the original collector of the data, the authorized distributor of the data, and the relevant funding agency bear no responsibility for use of the data or for interpretations or inferences based upon such uses.