This article is accompanied by the sample form and data file saved in this folder; for help deploying it to your server, check out our support article on deploying form definitions and server datasets. The workflow is broadly based on common practices of household listing, which you can learn from these resources.
Household listing is a key activity in most household surveys, in which all structures within the relevant geographic area of the study are listed. The geographic area is normally divided into clusters, more commonly known as the census enumeration areas (EA).
Each form instance collects information for a single structure (house or building), including location information, and whether it is residential, non-residential, or both. Residential structures can have one or more dwelling units, and for each dwelling unit, all households and household members must be listed (see image below). This data is useful for determining the sample size and collecting relevant details for later studies or stages.
In this image, each floor of the residential structure is a dwelling unit, and a dwelling unit can contain one-or-more households.
In this use case, we combine a number of features available in SurveyCTO to demonstrate household listing. The key components of such an undertaking are:
- Selecting the enumeration area
- Listing the households
- Collecting key household information
These are the files that you will need to deploy on your server to test this workflow:
||- Basic Household Listing|
|Pre-load file||- locations.csv|
Features that enable this workflow
This use case uses the following features of SurveyCTO:
Pre-loading using pulldata()
Pre-loading using search()
Understanding the workflow
Context of workflow
In this workflow, an enumerator has been asked to list all the structures in a given enumeration area, and all households in each structure. This is usually the first step in preparing for a household survey. The enumerator will collect details about the structures and existing dwellings, including physical features that can be used to locate households in a later stage. Enumerators might also collect specific information useful in determining eligibility during the sampling process, if the sample is based on some criteria. Below we describe how we can achieve each of the key components in SurveyCTO.
1. Selecting the right enumeration area
Each enumeration area is normally located within the context of a larger area (e.g. town or village), which are identified prior to data collection. In most countries, information about these areas is readily available from the national statistics offices. In some cases, information about enumeration areas may be incomplete, outdated, or missing. A statistician may be required to confirm the sourced enumeration areas, or to define new enumeration areas. The supervisors are then responsible for sharing the enumeration areas with the enumerators and providing details which can help enumerators recognize the areas, such as base maps when available.
Enumerators will then select their assigned enumeration area in the questionnaire. To minimize errors related to human data entry, you can start off by pre-loading all existing area levels, and use them to drill down to the location of the said enumeration area. We use the concepts of searching and selecting from pre-loaded data to drill down to the lowest level possible, and then enter the enumeration area identifier. Each enumeration area should have a unique identifier, so enumerators can enter or pick that unique identifier to select the enumeration area they have been assigned. For help creating unique identifiers, check out our support article on enforcing unique study IDs.
2. Listing the households
The form is filled in once for each structure, but there are instances where more than one household occupies the same structure (see image above). In such cases, the households are collected using a repeat group. The fields in the repeat group are asked once for each household in the structure, so you can identify the individual households within a structure if that structure is shared.
3. Collecting key household information
3.1 Characterization of the dwelling
It is important that a household recorded during household listing can be located at a later stage in the survey. Thus, collecting details that can help identify the structure in which the household is located becomes an essential part of the household listing process. The GPS location of the structure is recorded, and a photo of the structure is also taken. For cases where more than one household shares the same structure, photos of entrances to each household’s entrance to can be taken. There should also be a set of questions to gather a detailed description of the dwelling. This can be any distinguishing feature such as a particular roof color, presence of a large mango tree, and so on. Further information on what data to collect and considerations to make can be found in our use case on locating households in the fieldwork.
3.2 Characterization of the household members
It is also important to be able to identify the household members in each listed household. Identifying details about each household member is collected, including information about the household head. Information useful in the sampling process should also be collected. For example, if your survey will target children under 5, you may want to include questions that ask for the age of each household member, so your sampling can take this into account. This information will help survey managers determine which households should be part of the main survey.
The goal of listing is usually to get an accurate list of households for sampling. As such, the listing tool should be limited to information that supports this process. It should not grow into a full blown questionnaire.
Alternative designs and improvements
This use case has been purposefully kept as simple as possible, but in practice, your survey might have slightly different needs. Here are just a few ideas on how to adapt this use case.
Locating an enumeration area
In some cases, it can be difficult to locate the enumeration area, especially if enumerators are not familiar with the area. Even if enumerators know the area, finding out the boundaries of the enumeration area can be difficult. To aid this, enumerators are normally provided with a map of the larger area for them to understand their context, as well as maps of the cluster itself. These maps can be added into SurveyCTO Collect so the enumerators can use them for reference, or you can include images of a base map or sketch map in the form as reference points. Further, if GPS coordinates for the cluster exist, these can be used to help the enumerator ensure they are working within the right boundaries. Depending on the available data, you can:
- Use the GPS of an identifiable place such as a school, clinic, or four way stop within the enumeration area as a starting point for the enumerator. The GPS coordinates of this location can be embedded in an HTML link that launches Google Maps to help them navigate to the identifiable place in the enumeration area.
- If boundaries of the enumeration area are available, you can use a function like the distance-between() function to check that the structure is within the bounds of these dimensions. For more strict requirements, you can explore geofencing to ensure the enumerator is within the boundaries enumeration area.
Keeping track of the listed structures
It's very important that enumerators keep track of which structures have been listed and which haven’t. In some cases, enumerators may leave physical markers such as stickers on doors/buildings to show that a household has been listed, but this is not always possible. You can help your enumerators by including some way of tracking the listed structures with SurveyCTO. Two possible methods you can use would be:
- Estimate the number of structures and pre-load a list of structures with numbering. You can estimate the number of structures by using tools like satellite views of the area. A good satellite view can allow one to count the number of structures in an area. You can also get these estimates from community leaders within the area such as village headmen, councilors, and so on. Once you have these numbers, you can create a list with all existing structures (e.g. structure 1, structure 2, etc.) for the enumerator to help set expectations for that particular area. The enumerator can select structures from the pre-loaded list saved in a server dataset, and update the list with form data to inform whether the structure was already selected or not. Using this method, the list can be filtered to only show structures that haven’t been visited yet. Alternatively, you can use server dataset publishing to count the number of structures already listed, and display this number against the total number of structures to be listed, which will give you a progress rate.
- Use the display iframe field plug-in to show the structures that have been listed so far. Using this plug-in, you will be able to display the progress made so far. You can use it to simply display a list of the households listed so far, or if you have an estimate, you can add graphical information that can show the progress the enumerator has made to encourage them.
The content of the household listing tools vary depending on the context, but the broad strokes outlined above cut across most implementations. Some examples of context specific application of these include:
- UNICEF’s Multiple Indicator Cluster Surveys (MICS) MICS Manual for Mapping and Household Listing (see Sampling section).
- The Demographic Health Survey (DHS) program household listing manual.
- Micronutrient Survey Manual and Toolkit.
Do you have thoughts on this guide? We'd love to hear them! Feel free to fill out this feedback form.