Sampling Procedure
The Informal Surveys are conducted in selected urban centers, which are intended to coincide with the locations for the implementation of the main Enterprise Surveys. The overall number of interviews is pre-determined.
In The Democratic Republic of the Congo, the urban centers identified were Kinshasa, Kisangani, Lubumbashi, and Matadi. At the outset, the target sample in Kinshasa was 60 interviews, in Kisangani - 30 interviews, in Lubumbashi - 40 interviews, and in Matadi was 20 interviews. The sample will be confined to the major cities covered in the running in parallel enterprise survey of the formal economy. The target number of interviews will reflect, as far as practical, the individuals' population distribution but with no more than 60% sample from a single city and no city with fewer than 20 interviews in total.
Sampling in the Informal Surveys is conducted within clearly delineated sampling zones, which are geographically determined divisions within each urban center. Sampling zones are defined at the beginning of fieldwork, and are delineated according to the concentration and geographical dispersion of informal business activity. After the sampling sizes are defined for each location, every city is divided into several zones that may or may not correspond to the administrative districts.
The zones are selected using local knowledge (information about the districts/parts of the city where most of the informal businesses are located).
The number of zones depends on the sample size for this geographic location. For example, if in a particular city there are 40 planned informal interviews (20 manufacturing and 20 services) in the city should have 10 zones (4 interviews per zone including 2 manufacturing and 2 services firms).
In DRC, using Google maps or local city maps, the target areas within each city were identified. With input from the local agency applying local knowledge, the starting points were defined. The number of zones was determined by the target sample size for each city divided by the cluster size (4 interviews). This is the only element of the sample selection where there is local influence.
In Kinshasa, for a total of 60 interviews, 15 sampling zones were initially identified (60/4=15 zones). In Lubumbashi, a total of 40 interviews yielded 10 sampling zones (40/4 = 10 zones). In Matadi, a total of 20 interviews yielded 5 sampling zones (20/4 = 5 zones). In Kinsangani, only 6 sampling zones were drawn and some zones were over or under sampled. As described above, the criteria used in choosing these sample sectors was a combination of territorial dispersion and the presence of informal businesses.
In order to provide information on diverse aspects of the informal economy, the sample is designed to have equal proportions of services and manufacturing (50:50). These sectors are defined by responses provided by each informal business to a question on the business's main activity included in the screener portion of the questionnaire. Manufacturing activity in the informal sector includes business activity requiring inputs and/or intermediate goods. Thus, for example, the processing of coffee, sugar, oil, dried fruit, or other processed foods is considered manufacturing, while the simple selling of these goods falls under services. If an informal business conducts a mixture of these activities, the business is considered under the manufacturing stratum.
In order to ensure a degree of geographical dispersion within each sampling zone, two starting points were identified. Each starting point is marked on the map and is usually located at a major intersection or other prominent point at the edge of a zone. The road(s) that the interviewer should follow and the direction the interviewer should take is also marked on the map created by the local consultant.
Each sampling zone, including its two starting points, were marked using Google maps, with the GPS coordinates of the starting points being systematically recorded. Additionally, when obtaining a complete interview, the exact address of the informal business (or where the interview took place) was registered by the interviewer. Once in the office, this address was searched in Google maps, and its GPS coordinates were registered in a fieldwork report.
If no address was immediately available, using local knowledge, the GPS coordinates were determined using imaging via Google maps. In order to preserve confidentiality, the exact coordinates of businesses are not published.
Due to issues of non-response, in the process of fieldwork, the implementing contractor was unable to obtain the targeted four interviews in each of the originally delineated sectors.
As a result, replacement sectors were delineated, ex post. Additionally, the implementing contractor noted that in various interviews there were notable shortfalls in response rates to certain questions. For these reasons, additional interviews were authorized. These were distributed according to the discretion of the implementing contractor in DRC, with authorization from the World Bank.
In sum, there were 36 zones in DRC; Kinshasa (15 zones), Kisangani (6 zones), Lubumbashi (10 zones), and Matadi (5 zones).
Complete information regarding the sampling methodology can be found in "Description of The Democratic Republic of the Congo Informal Survey Implementation" in "Technical Documents" folder.