Note: This page summarizes the rationale behind a GiveWell grant to Evidence Action. Evidence Action staff reviewed this page prior to publication.
In a nutshell
In August 2024, GiveWell recommended a $198,993 grant to Evidence Action to conduct baseline coverage surveys of iron and folic acid (IFA) supplementation in two Indian states, Uttar Pradesh and Bihar. We recommended this grant because it will help us decide whether to make a larger grant to fund technical assistance (TA) to increase IFA coverage in these two states.
Published: April 2025
The intervention
IFA supplementation consists of providing oral supplements containing a combination of iron and folic acid, in either syrup or tablet form, to reduce anemia. See our intervention report on iron supplementation for more details.
Evidence Action provides TA to governments to implement IFA programs. See our page describing an August 2022 grant to Evidence Action's IFA program for more details.
The organization
Since March 2018, we've directed approximately $13.7 million to support Evidence Action's TA program for IFA supplementation in five Indian states: Haryana, Jharkhand, Madhya Pradesh, Rajasthan, and Uttarakhand.1 We are now considering expansion in new states, including Uttar Pradesh and Bihar. In January 2024, we also made a ~$100,000 grant to support baseline coverage surveys in two additional states, Karnataka and Tripura.
The grant
The grant will enable Evidence Action to run baseline surveys estimating IFA coverage in Bihar ($74,170) and 50 of the 75 districts in Uttar Pradesh ($124,823).2 Data collection was done by an independent survey firm, and the resulting survey data will be analyzed in collaboration with IDinsight, an independent evaluator.3 This grant will provide more information about existing levels of IFA coverage in schools in these states.4 If baseline survey results show that existing IFA coverage is low, we may consider an additional grant to expand Evidence Action’s IFA program in both states.
The case for the grant
We recommended this grant because:
- We think these surveys will help us decide whether to make a larger grant to fund TA to increase IFA coverage in two additional Indian states. Assuming existing IFA coverage levels are low, we think IFA supplementation in India is a promising program. Our current best guess is that it is at least as cost-effective as the program we funded in Rajasthan, Madhya Pradesh, Jharkhand, Haryana, and Uttarakhand in 2022, which we modeled as roughly 12 times as cost-effective as unconditional cash transfers.5 We estimate that the funding gap over five years would be around $8.4 million (~$3.7 million in Bihar and ~$4.7 million in Uttar Pradesh).6
- If we decide to make a larger grant, these surveys will enable us to assess the effect of Evidence Action’s program on increasing IFA coverage. If we decide to fund Evidence Action’s IFA program in both states, we will likely include an endline evaluation of coverage in that grant. Having survey results on baseline coverage will enable us to estimate the impact of the program on increasing coverage levels after program implementation. We think this will improve our assessments of the general value of providing TA to increase IFA uptake.
Risks and reservations
Our main reservations about this grant are:
- Even if the surveys show low coverage, we might decide not to fund further expansion due to other uncertainties about the program’s impact. We have a number of uncertainties about the impact of iron supplementation programs in general, and Evidence Action’s program in particular. For example, we're uncertain about the effect the program will have on IFA coverage (see our page about the 2022 grant to support Evidence Action's IFA program for details). We think we will learn more about this when we receive results from endline surveys we funded in the five states we currently support.7 It’s possible we should have waited for these results before funding baseline surveys in Uttar Pradesh and Bihar.
- If we decide to make a larger grant, and fund endline surveys to assess the effect of Evidence Action’s program on increasing IFA coverage, we will still not know what coverage would have been without Evidence Action. A comparison of post-intervention (endline) data to pre-intervention (baseline) data does not necessarily conclusively indicate the effect of a program, because there is no guarantee that any measured difference is due to the intervention. This is a major limitation of using baseline and endline data to estimate Evidence Action’s effect on coverage.
- It’s possible that we should have funded IDinsight to conduct the baseline survey instead. We think there is inherent risk of bias in Evidence Action conducting baseline and endline surveys of its own programs. We could have funded IDinsight to conduct these evaluations instead, but this would have been more costly,8 and we think the risk of bias is ultimately low since Evidence Action will hire an independent firm for data collection. However, we are still considering funding IDinsight to perform an independent analysis of baseline data to minimize this risk.
- It’s possible that we should have funded baseline evaluations in all 75 districts in Uttar Pradesh. Baseline evaluations will be conducted in the 50 districts where Evidence Action would work if we make an additional grant to support its program. The remaining 25 districts are currently receiving TA support for IFA implementation from UNICEF and Nutrition International.9 We opted not to support baseline surveys in these districts because they are already receiving TA support, but it’s possible we could have learned something useful from these evaluations.
- It's possible that the timing of these surveys will be unusually poor. Evidence Action plans to run the survey in December 2024.10 There may be school closures due to winter weather or holidays in Bihar and Uttar Pradesh in December, which could disrupt the survey and mean we won’t get representative baseline data.
Plans for follow up
- We plan to review endline results from the coverage surveys once they are published by IDinsight, likely in June or July of 2025. Based on these results, we will make a decision on further funding in Bihar and Uttar Pradesh by December 2025.
Internal forecasts
For this grant, we are recording the following forecasts:
| Confidence | Prediction | By time | Resolution |
|---|---|---|---|
| 65% | We recommend a >$3.5 million grant to support Evidence Action's IFA program in either Bihar or Uttar Pradesh. | December 2025 | - |
Our process
- We had several email conversations with Evidence Action.
- We completed a quick, back-of-the-envelope cost-effectiveness analysis of an IFA program in both Uttar Pradesh and Bihar.
Sources
- 1
- In March 2018, we recommended funding ($320,000) to the Evidence Action Accelerator to scope a technical assistance (TA) program for IFA supplementation in India.
- Based on the results of this scoping work, we recommended a grant (approximately $3.4 million) in December 2018 to Evidence Action to scale a pilot of the IFA TA program and a grant in March 2019 ($800,000) to evaluate the program.
- In August 2022, we recommended funding (approximately $9.2 million) to support the IFA TA program across five states in India.
- $320,000 + ~$3,400,000 + $800,000 + ~$9,200,000 = ~$13,720,000.
- 2
- “On the budget front, we have revised our estimates to include 1) hiring a consultant for support and 2) expanding the survey slightly to include 300 schools + 100 ASHAs per state (which lets us keep our margin of error under 5%). That means the total survey costs $198,993, split between $74,170 for Bihar and $124,823 for UP.” Ruben Conner, Senior Manager, Evidence Action, email to GiveWell, July 25 2024 (unpublished).
- “In UP, we are planning to provide TA support in 50 of the 75 districts, as the remaining 25 districts are supported by UNICEF.” Chau Hoang, MLE Strategy Manager, Evidence Action, email to GiveWell, July 22 2024 (unpublished).
- 3
- "Data collection was done by an independent local firm and data analysis will be done by IDinsight. Evidence Action helps provide coordination and management for both activities, but does not influence data or analysis decisions." Evidence Action, comment on a draft of this page, January 31 2025 (unpublished).
- 4
"The survey includes 500 schools, 150 ASHAs/AWWs, and 600 households in each state." Evidence Action, comment on a draft of this page, January 31 2025 (unpublished).
- 5
- 6
Email from Evidence Action, June 27, 2024 (unpublished).
- 7
See additional details on these surveys in the "Plans for follow-up" section of this grant page.
- 8
Our best guess is that the total cost for baseline and eventual endline surveys through Evidence Action would be around $600,000 ($200,000 for baseline, $200,000 for endline, and $200,000 for IDinsight oversight). This estimate is based on unpublished emails with Evidence Action staff.
IDinsight's budget for the baseline survey alone is around $600,000, so we'd guess that total survey costs for baseline and endline would be around $1.2 million (twice as high as costs through Evidence Action). IDinsight, DataDelta Proposal - IFA supplementation coverage survey in Uttar Pradesh and Bihar, 2024 (unpublished). - 9
"In UP [Uttar Pradesh], we are planning to provide TA support in 50 of the 75 districts, as the remaining 25 districts are supported by UNICEF and NI [Nutrition International]." Chau Hoang, MLE Strategy Manager, Evidence Action, email to GiveWell, July 22 2024 (unpublished).
- 10
"Timing: We have confirmed with the team in India that running a December survey is feasible. This sets up the following approximate timeline: … Dec 2nd - Survey begins; Dec 30th - Data collection completed; Jan 31st - Data is cleaned and passed to IDinsight." Chau Hoang, MLE Strategy Manager, Evidence Action, email to GiveWell, July 22 2024 (unpublished).