Double Counting in International Development
February 7, 2022
Double counting can pose a threat to transparent and true reporting. Organisations who use data to report their work must ensure that the effects of double counting are mitigated so that results are reliable and valid. This article will explain double counting, looking at how it arises in international development projects and how it can be prevented.
What is Double Counting?
Double counting causes an inflation of project results. This happens when an organisation or project counts a recipient or event more than once during a reporting period. When all the data is combined, this causes the overall project results to be unintentionally inflated. This inflation of project results is known as over-reporting. Over-reporting makes it difficult to understand how effective the project actually is. This in turn can affect decision making around the project as coverage gaps can be masked, making it hard to see what areas are worth scaling up and which aspects need to be improved.
Financial Double Counting During the Pandemic
Double counting not only occurs in international development projects, it also poses an issue in other kinds of reporting, such as financial reporting. At the beginning of the Pandemic, the topic of double counting in such an instance was brought into focus as the OECD’s Development Assistance Committee agreed to allow donor countries to count debt relief as official development assistance. Such an agreement came under scrutiny as it meant that donors could double count aid money, once when they provided a loan and again if they provided debt relief. This double counting would make it possible for donors to artificially inflate their aid statistics so that they hit their aid targets, all while recipient countries could in fact be receiving less money than before. This discrepancy between donor reporting and the amount of assistance recipient countries actually received is an example of double counting.
How Does Double Counting Occur in International Development Projects?
Within the contexts of international development projects, MEASURE Evaluation identifies three different types of Double Counting in their research:
- Within-Partner Double Counting of Individuals
- Between-Partner Double Counting of Individuals; and
- Double Counting of Sites
Within-Partner Double Counting of individuals occurs when a partner or project at one site provides the same service multiple times to the same individual. In their reporting, this individual is counted for each time they receive the service as a separate recipient, thus inflating the number of beneficiaries. For example, an organisation is running three different workshops on quality teaching approaches. Teachers from all over the country are invited to participate. Each teacher can choose to attend 1,2 or all 3 of the workshops depending on which topics they are interested in. In their reporting, the organisation wants to keep track of the number of teachers who received training. However, since the same individual may take multiple training workshops within that reporting period, the partner may over-report the number of individuals trained if this is not taken into consideration.
Similarly, Between-Partner Double Counting of Individuals, involves two or more partners providing the same service to the same individual either at the same site or at different sites within the reporting period. Both partners then include that individual in their count of the number of beneficiaries. For example, partners aim to count the number of beneficiaries of malaria treatment across two projects active in the same region. In the reporting, if the team is not aware that an individual might be receiving treatment from project A as well as project B, resulting in that individual beneficiary being counted twice in the reporting period.
Lastly, Double Counting of Sites occurs when different partners provide different supplies and/or services to the same organization within one reporting period and each partner counts that same organization as one of its service points. For example, partner A provides training on handwashing and hygiene to providers at site Z. Partner B provides PPE to this same site. When reporting the number of service outlets carrying out testing, both partner A and partner B count and report site Z.
Ways to Mitigate Double Counting
With these cases of double counting in mind, we can now look at how these situations can be better accounted for in reporting.
1. Separating Partners by Project or Site.
To avoid Double Counting of Sites and some of the effects of Within Partner Double Counting of Individuals, it can be helpful to separate out partners by site. For example, Partner A provides clinical care to COVID patients in Region 1 while Partner B provides care to COVID patients in Region 2. Similarly, partners can also be separated by project. For example, Partner A provides clinical care to COVID patients in Region X whilst Partner B provides medical equipment in the same region. By separating partners by site or project, project coverage can be maximized and reduce duplication efforts. However, it should be noted that such practice will not fully eliminate Between-Partner Double Counting of Individuals. This is because some beneficiaries may move between sites and receive services from more than one partner.
2. Estimating the Degree of Overlap in the Project.
Often, partners will overlap both geographically and in their projects. To mitigate all three different types of double counting, partners can adjust their reporting after data has been collected. For example, if an organisation has a KPI measuring the total number of beneficiaries across two of their projects in one region, they will be aware that some beneficiaries are receiving services from each project during the reporting period. The organisation can estimate this overlap as a percentage. This estimated overlap can then be included in the calculation of the adjusted total number of beneficiaries.
Just as this estimation can help with over reporting, it can also be used to tackle under reporting instances. For example, an indicator measuring the total population in two different states uses census data from each state in its calculation. However, historical analysis of this census data shows that for state 1 the census reports are always 15% over reported whereas in state 2 they are 5% underreported. With this in mind, using this percentage weighting can help organisations to correct for either double counting or underreporting.
3. Adjusting Client Visits to Number of Clients Reached.
To prevent Within-Partner Double Counting of Individuals, partners can move away from counting the number of unique individuals they have served and instead track the number of clients reached instead. To do this, the partner can distribute surveys asking people how many visits to the facility they have made during the reporting period. After collecting responses, adjustments can be made to data based on these empirical findings.
4. Unique identifiers.
To avoid Within-Partner Double Counting of Individuals, organisations can assign project beneficiaries unique identification numbers through either paper-based or computer based monitoring systems. If such unique identifiers are shared between partners, this can also mitigate Between-Partner Double Counting of Individuals. It’s important to ensure privacy and confidentiality concerns are taken into account when sharing data in such a way.
Unique identifiers can also be used within the indicator workflow itself to help differentiate between sites and services which different partners are involved in. This can help prevent double counting in sites as if all partners utilize the same unique identifiers in their indicator workflow, data can be easily aggregated and adjusted.
Double-counting is a common problem in reporting. While it cannot always be remedied, being aware of double counting and the way in which it can present itself in projects is key to mitigating its effects. This ensures transparent and powerful reporting for organizations.
If you want to learn more about how you can adjust results for double counting in TolaData check out our guide in the Help Center.
We hope our article was helpful. If you have any feedback or suggestions on how we can improve it, please feel free to write to us in the comment section below.
