With rigorous economic research and practical policy solutions, we focus on the issues and institutions that are critical to global development. Explore our core themes and topics to learn more about our work.
In timely and incisive analysis, our experts parse the latest development news and devise practical solutions to new and emerging challenges. Our events convene the top thinkers and doers in global development.
CGD’s work in technology and development focuses on the macroeconomic implications of technology change as well as technological applications for specific development challenges.
Technological advances are a driving force for development. But policy choices determine who benefits. CGD focuses on three key questions around innovation, growth, and inequality: How can governments use existing technologies to deliver services more effectively to citizens? How can international institutions help create and spread new technologies to tackle shared problems like climate change and pandemics? And how can policymakers ensure advances in artificial intelligence, automation, and communications bring shared benefits and not greater global inequality?
Today, we published this year’s Commitment to Development Index (CDI), which ranks 27 of the world’s richest countries in how well their policies help to spread global prosperity to the developing world.
We will be presenting the Index and our recommendations at the high-level period of the UN General Assembly (UNGA) later this month. As political leaders prepare to meet for UNGA, here are some key takeaways from our research that should help guide their policies and discussions.
1. Leadership on global development isn’t only for the richest!
The CDI analyzes the policies of 27 of the world’s richest countries in seven key areas: aid, finance, technology, environment, trade, security, and migration. The indicators adjust for size and economic prosperity—and the results demonstrate that country wealth does not determine the results. The wealthiest countries—represented by the G7—rank anywhere between fourth and twenty-sixth. Income per person averages half that of the United States in Visegrád countries (Czech Republic, Hungary, Poland, and the Slovak Republic), but all four now rank higher in their commitment to development. Portugal, who ranks sixth, performs well in most components despite being less prosperous than many of the CDI countries. Smart policy design is not a matter of prosperity only. Therefore, our first key message to all the leaders of the 27 CDI-countries:
Domestic economic challenges needn’t prevent leadership on smart policies to increase global prosperity.
2. Development is about much more than aid
The CDI draws attention to the fact that global development is about so much more than the amount or quality of foreign development assistance provided. Policymaking in various policy fields directly affect the lives of poor people around the globe.
For example, the design of our policies on technology or finance affect the prospect for people living in poorer countries. Both research and development policies and investment policies are mainly pursued for domestic goals. However, they have a lasting effect on developing countries. Smart intellectual property rights can enable knowledge sharing and technology transfer. Also, bilateral investment agreements with developing countries recognising specific public policy goals such as labour rights, environmental standards, or human rights can have an important effect on the prospect for development.
The commitment to implementing balanced and sustainable policies domestically also sends a strong signal about their importance globally and irrespective of countries borders. Money spent on foreign development assistance does not have the same lasting effect if countries don’t recognise the international impact of their actions in other policy areas. Therefore:
In our integrated world, your policies and decisions as a leader of a rich country have an important bearing on the lives of people in developing nations.
3. Even the bottom-ranked country has smart policies we all can learn from
Like the Sustainable Development Goals (SDGs), the CDI recognizes development has many angles. But while the SDGs cover all nations and their outcomes, the CDI concentrates on the richest countries and emphasizes how policies can make a huge difference to development globally. The fact that we limit our evaluation to high income countries means that policy recommendations are more tailored and relevant. Even the best-ranked countries have weaknesses where they can learn from their peers. Overall leader Denmark performs weaker on migration and could learn from the migration policy designs from countries as varied as Luxembourg, New Zealand, or neighbouring Sweden. Similarly, bottom-ranked South Korea could advise all other 26 CDI countries on how to build long-lasting support for research and development. Accordingly:
Use the CDI as a tool to learn from others and to inspire change through your own best-practice policies.
4. Some overall progress on the Environment component but stronger commitments are needed
Tragically, Hurricane Harvey has reminded the United States how vulnerable we all are when natural disasters hit. Further, people in South Asia were left suffering after massive flooding and devastation affected millions, while earlier this year we saw how the unprecedented drought in Africa affected the lives of millions facing malnutrition. These tragic events, sadly far from unique, remind us that we all need to do more to combat climate change.
This year’s CDI points out that progress has been made—CDI countries report progress in curbing new greenhouse gas emissions and the amount of Ozone-depleting substances has been cut significantly. However, many environmental challenges remain. We need to see an even bigger commitment to development from the CDI countries in the future, such as a complete support for the Paris Agreement and the willingness to tackle issues such as overfishing and deforestation. Thus, our final recommendation:
While progress has been made, many global challenges remain. We ask this generation of world leaders to strengthen and deepen their commitment to development.
These findings show that we and our governments can do so much more to spread prosperity to poorer countries. The CDI serves as a useful tool to identify which national policies still have potential to be designed in a more development friendly way. We hope world leaders use the opportunity of UNGA to discuss ways to make further progress in all policy fields, inspiring each other to achieve more on global development.
While the right to privacy decision has not specifically pronounced a judgement on the constitutionality of Aadhaar (it will be decided by a separate bench of the Supreme Court), the judgement stated that “(t)echnological change has given rise to concerns which were not present seven decades ago and the rapid growth of technology may render obsolescent many notions of the present." However, the Supreme Court advised the government to create a "robust regime of data protection ensuring a careful and sensitive balance between individual interests and legitimate concerns of the State…(including) preventing the dissipation of social welfare benefits.”
Aadhaar has already demonstrated the potential of digital ID to transform systems of governance and increase efficiency of private transactions. By addressing the genuine concerns of individual privacy and data protection, it will be ready to move ahead and lead by example as it has already done on the technological side. The right to privacy judgement is an opportunity to make the world’s largest experiment in digital governance a bigger success than it already is. In this blog post I will explain how India can learn from other countries to do just that.
Changing how Indians interact with the state
Preventing leakage of government expenditure and ensuring inclusive growth is the original justification for Aadhaar which is now the largest biometric database in the world with over 90 percent of 1.25 billion Indians enrolled in the program. Until now, Aadhaar has had a relatively smooth ride, much like cruising on a relatively empty highway. Following a remarkably rapid rollout, Aadhaar has been widely accepted as a proof of identity, used as an instrument to expand banking and mobile phone services and restructure the delivery of basic services and subsidies, including the widely acclaimed cooking gas subsidy reform. Aadhaar’s authentication capabilities are driving digital applications in public food distribution systems (PDS), social pensions, mobile payments etc., impacting the lives of millions of people and billions of dollars in public resources. Aadhaar is changing the way ordinary people interact with the state but is also asking them to trust it to keep private information safe in an increasingly interconnected world.
Three attributes of ID systems
Resolving the issue of individual privacy and data security is therefore of utmost importance. India is not alone in facing this challenge. From the beginning of this century, there has been a rapid spread in identification programs both in high income countries (HICs) as well as in the developing world (see chart).
Identification Programs in High-Income (HIC) and Other Countries
Source: World Bank ID4D dataset; Jan 2016 edition
From the earliest times, identification systems have relied on some subset of the many different attributes that together distinguish each of us as a distinct person. They may be bundled into three main factors and combined in different ways to form the basis for identification:
Something you are: a biometric—physical appearance, or a distinctive voice or smell or behavioral pattern. This is surely the oldest form of recognition, as well as the most recent. In modern digital times, it can encompass digital biometrics such as fingerprint, iris or finger-vein-patterns, voice patterns, DNA, dynamic signature or, on a computer, keystroke rhythms or mouse movement patterns.
Something you have: a birth certificate, an ID card or a mobile ID, token or other physical credential. Something you have could also be a person—a well-respected and credible individual ready to vouch for you. The best identifier for an infant is still probably its mother.
Something you know: a personal identification number (PIN) or password, or the ability to provide personal information that is unlikely to be known by others.
There are examples of modern identification systems that combine all three identification factors: have, know and are. Estonia’s pioneering digital ID system is the best example. At the age of 15 fingerprints are taken on registration for the digital ID (who you are: to prevent multiple identities). A card is issued (what you have); it includes digital certificates to enable the holder to both authenticate his identity and to sign documents electronically. For these purposes, the holder authenticates himself against the card by providing two user-selected and private PINs (what you know), one to authenticate himself and the other to sign. Multi-factor authentication offers extra security relative to the use of only one factor. The main difference with Aadhaar is the option of the private PINs which gives the individual control over the authentication process and reduces the chance of impersonation without his or her knowledge, ensuring greater degree of privacy over the centrally warehoused information.
Control over data sharing
Countries have taken several approaches to mitigating the risks associated with integrated identification systems—something Aadhaar’s critics have repeatedly pointed out. One approach is to provide for multiple identifiers while still ensuring that identity itself is unique. This corresponds to the idea that a single, unique individual should be able to be identified in different ways for different purposes. In Austria’s distinctive system, cryptography enables users to access multiple government services using a single e-ID while ensuring that records cannot be matched across the different databases using a common number. Each e-ID card includes a unique personal identifier, the Source PIN, which is stored only on the card and cannot be held by any government body or private entity. When a card-holder wants to access a service using the card, a unique sector-specific identifier, called the sector-specific PIN or SS-PIN, is generated from the Source PIN using a one-way cryptographic function. The SS-PIN is different for each service and also for different private sector applications. It is not possible to derive one SS-PIN from another, for example, to generate the tax SS-PIN from the social security SS-PIN. Neither can an SS-PIN be used to calculate the Source PIN.
This unique system enables card holders to use the same credential to authenticate themselves for all purposes, but prevents the consolidation of their records using a common number although, of course, it does not prevent the use of names or other biographic data to compare records. The generation of SS-PINs without the card is only possible for the source-PIN Register Authority and is only permissible when personal data are to be processed or transmitted in conformity with the Data Protection Act of 2000.
These are just two examples of several possible country experiences that India can learn from. The Supreme Court judgement should serve as a pause button to take stock of what has been achieved and how it can go further while protecting ‘informational privacy’ as the Court has ruled.
With the US Congress considering cuts to foreign assistance and aid budgets in other donor countries coming under increased pressure, evidence about what works in global development is more important than ever. Evidence should inform decisions on where to allocate scarce resources—but to do so, evaluations must be of good quality. The evaluation community has made tremendous progress on quality over the past decade. Several funders have implemented new evaluation policies and most are conducting more evaluations than ever before. But less is known about how well aid agencies are evaluating programs.
To fill in the gap, we—together with our colleagues Julia Raifman Goldberg, Felix Lam, and Alex Radunsky—set out to assess the quality of global health evaluations (both performance and impact evaluations). We looked specifically at publicly available evaluations of large-scale health programs from five major funders: USAID, the Global Fund, PEPFAR, DFID, and IDA at the World Bank. We describe our findings in a new CGD Working Paper and accompanying brief. Check out the brief recap of our findings below.
What types of evaluations are aid agencies conducting?
We identified a total of 299 evaluations of global health programs published between 2009 and 2014. One feature stood out to us: performance evaluations made up an overwhelming majority (91 percent), with impact evaluations accounting for less than 10 percent. This is comparable to the share found across USAID evaluations in all sectors by an earlier study. And among impact evaluations, those using experimental methods, known as randomized controlled trials or RCTs, constituted a minority (we only found five RCTs). When looking at evaluations commissioned or conducted by major funders, the often-made criticism that RCTs are displacing other forms of evaluation doesn’t hold up.
How well are aid agencies evaluating global health programs?
We randomly sampled 37 evaluations and applied a standardized assessment approach with two reviewers rating each evaluation. To answer questions about evaluation quality, we used three criteria from the evaluation literature: relevance, validity, and reliability. We considered evaluations as relevant if the evaluation addressed questions related to the means or ends of an intervention, and used appropriate data to answer those questions. Evaluations were considered valid if analyses were methodologically sound and conclusions were derived logically and consistently from the findings. Evaluations were considered reliable if the method and analysis would be likely to yield similar conclusions if the evaluation were repeated in the same or similar context.
We constructed four aggregate scores (on a three-point scale) to correspond with these criteria. Overall, we found that most evaluations did not meet social science standards in terms of relevance, validity, and reliability; only a relatively small share of evaluations received a high score.
Looking across different types of evaluations, we found that impact evaluations generally scored better than performance evaluations on measures of validity and reliability.
What can aid agencies do better going forward?
Building on our analysis, we developed 10 recommendations for aid agency staff overseeing and managing evaluations to improve quality.
Classify the evaluation purpose by including this information in the title and abstract, as well as coding/tagging categories on the agency website.
Discuss evaluator independence by acknowledging the evaluators’ institutional affiliation and any financial conflicts of interest.
Disclose costs and duration of programs and evaluations.
Plan and design the evaluation before program implementation begins; we found that early planning was associated with higher evaluation quality.
State the evaluation question(s) clearly to ensure the right kinds of data are collected and an appropriate methodology is used.
Explain the theoretical framework underlying the evaluation.
Explain sampling and data collection methods so subsequent researchers could apply them in another context and readers can judge the likelihood of bias.
Improve data collection methods by using purposeful or random sampling, where possible, that provide more confidence in findings.
Triangulate findings using varied sources of qualitative and quantitative data.
Be transparent on data and ethics by publishing data in useable formats, and taking appropriate measures to protect privacy and assure confidentiality.
This set of recommendations draws on the high-quality evaluations we found in our sample. These examples showed that it is possible to conduct good quality evaluations for a range of methodologies and purposes. In many cases, quality improvement is possible within existing budgets by planning early or using better data collection approaches. Taking steps to improve quality can help ensure evaluations promote learning about what works and hold funders and implementers accountable—with an eye on increasing value for money and maximizing development impact.
How has the H-1B visa program—the US program that enables US employers to hire foreign nationals with specialty skills—impacted the US and Indian economies? Amidst the ongoing debates in both the United States and India about the H-1B visa program, The IT Boom and Other Unintended Consequences of Chasing the American Dream, prepared with my colleague Nicolas Morales, demonstrates the positive impacts of the H-1B visa program in both the United States and India. We find that the program provides benefits to US and Indian workers and consumers, and that it is a contributing factor to the expanding hi-tech sectors in both countries.
Relevant to the ongoing US debate about immigration policy, we offer this evidence-based report examining the relationship among migration, the IT sector, and the economy. We found that:
US workers are on average, better off by about $431 million (or $1,345 per additional migrant) in 2010 because of the H-1B program. The study incorporates crucial mechanisms like innovation by businesses, trade with other countries, and the choices made by students and workers to become computer scientists.
While there are some negative impacts for a subset of US workers (earnings for US born computer scientists are lower by 1.5 percent), the overall gains outweigh the losses as the combined incomes of the US and India both rise under the H-1B program by about $17.3 billion or 0.36 percent. And total IT output from both nations rose steadily under the H-1B regime by about 0.45 percent in 2010.
Better technology, as a by-product of this immigration of tech workers, increased the overall productivity of other sectors as well, and consumers of computer-related goods enjoyed better software and lower prices. The study found a 1 percent decrease in price for US IT products and an 7.4 percent fall in Indian IT products.
Similarly, in India, critics assert that the H-1B program is causing a “brain drain” from the country, i.e., highly skilled Indian workers are lost to the United States under this visa program. But to the contrary, our study found that the US H-1B visa program leads to a brain gain for India:
The prospect of migrating to the US and earning a higher wage induced many students and workers to switch to Computer Science and Engineering fields. Those who could not join the US workforce—due to a cap on the number of H-1B visas—remained in India, enabling the growth of the Indian IT sector.
Those who migrated to the US acquired skills, technical know-how, and established networks with US companies. As their visas expired, some returned with this acquired human capital and technology and contributed to the growing tech-workforce in India. Together, the “brain gain” to India under the H-1B program outweighs any “brain drain.”The increase in IT sector productivity, because of the additional knowledge and skilled workers spurred by the H-1B visa program, allowed India to eventually surpass the US in software exports. Over time, some IT production begins to be outsourced from the US to India.
The bottom line? Each country’s economy and workforce as a whole are better off because of the H-1B visa program. Furthermore, instead of causing “brain drain” from India to the United States, the H-1B visa program has led to a “brain gain,” spurring Indian workers to acquire education for high-skilled tech jobs.
With the majority of all H-1B visas going to Indians, we study how US immigration policy coupled with the internet boom affected both the US and Indian economies, and in particular both countries’ IT sectors. The H-1B scheme led to a tech boom in both countries, inducing substantial gains in firm productivity and consumer welfare in both the United States and India. We find that the US-born workers gained $431 million in 2010 as a result of the H-1B scheme.
Cutting through the layers of hype surrounding blockchain technology is tough work. Underlying the buildup in excitement, however, is a remarkable tool that could, if designed and used appropriately, help improve processes related to several long-standing development challenges. In our new paper “Blockchain and Economic Development: Hype vs. Reality,” we examine the technology’s potential role in addressing four of those challenges:
making aid disbursement more secure and transparent;
facilitating faster and cheaper international payments;
providing a secure digital infrastructure for verifying identity; and
securing property rights.
We argue that, while blockchain-based solutions have the potential to increase efficiency and improve outcomes dramatically in some use cases and more marginally (if at all) in others, key constraints must be resolved before blockchain technology can meet its full potential in this space. Overcoming these constraints will require increased dialogue between the development and technology communities and a stronger commitment to collecting and sharing data about what’s working and what isn’t in pilot projects that use the technology.
The case of aid distribution
Making aid disbursement more efficient and transparent is one of the areas in which blockchain technology shows promise. Consider the Start Network, which brings together 42 national and international aid agencies—including the International Rescue Committee, Oxfam, and World Vision—with the goal of improving their ability to collectively respond to humanitarian crises, including by enabling its members to agree upon projects and disburse funds within 72 hours of a crisis. Given the group’s belief that the humanitarian system must radically change to accelerate crisis response, it’s not surprising that it is now exploring how blockchain technology might help to provide aid more efficiently and effectively. In July, the Network announced a partnership with blockchain start-up Disberse that will use the company’s platform to speed up the distribution of funds and better trace how funding is spent.
This is just the latest example of how aid organizations are putting blockchain technology to the test. The UN’s World Food Programme (WFP) also recently conducted a successful pilot project in Jordan, where it used a blockchain to manage cash-based transfers to 10,000 Syrian refugees living in the Azraq camp in Jordan. The organization hopes to expand the pilot to cover all 500,000 WFP beneficiaries in Jordan by the end of the year.
From the perspective of individual donors, conducting aid payments on a blockchain can provide three advantages: speed, transparency, and the ability to bypass traditional financial intermediaries. Further, if multiple donors share project information on a single distributed ledger, they can improve coordination both between themselves and with recipient governments.
In the case of aid distribution, the biggest challenge is the inherent nature of the development agencies and large non-profits who may ultimately use the technology. These organizations tend to be risk-averse and slow to innovate—a sensible stance since they act as stewards of donor resources and provide services that can mean the difference between life and death for beneficiaries. For that reason, the development and tech communities will need to work together to address concerns about data security, governance, and operational resiliency that relying on a blockchain raises before wider adoption is likely.
These challenges all appear to be solvable, but the ability of technologists to prove that the solutions they offer provide a significant advantage over existing approaches may be hampered by an absence of quality data. In part, this lack of data simply reflects the newness of the technology. But there is also a concerning trend in which start-ups announce pilot project “successes” without backing up their claims with metrics. This reticence is understandable given the stiff competition for funding and market share, but it undermines the broader effort to design effective solutions. The government agencies and international institutions that partner with start-ups on pilot projects can solve this problem very simply by requiring their partners to collect and publish relevant metrics, and working with them to make that a reality.
These organizations should also work with technologists to develop a set of principles and (eventually) standards for using the blockchain-based solutions in the context of development. The Principles for Digital Development, which have been endorsed by over 100 organizations working in international development, provide a useful model for this effort. While it may be counterproductive to set standards now, given the rapid pace of innovation, it is important to have conversations with an eye towards what these standards might look like in the future to prevent different organizations from developing systems that are ultimately incompatible. Recently formed projects like New America’s The Blockchain Trust Accelerator and Consensys’ Blockchain for Social Good are already bringing actors from both communities together for this discussion and these efforts should get a further boost from the World Bank’s recently announced Blockchain Lab.
Despite the over-hyping of blockchain technology’s potential in certain use cases, we believe that several applications show real promise. A little coordination and a lot of quality data would go a long way towards realizing that promise and improving development outcomes.
When Pratham used simple “report cards” to provide information about learning outcomes to villages in India, the intervention largely failed. There was no improvement in attendance of children or teachers, no improvement in learning outcomes; and parents, teachers, and village education committees did not become more engaged with the schools (Banerjee et al., 2010). However, when Pratham-trained youth volunteers offered basic reading classes outside of regular school, reading skills of children who attended improved substantially after one year.
Why did information provision fail to improve learning outcomes? Principals and teachers in these schools were both formally and informally accountable only via top-down reporting to the bureaucracies they worked for. Providing information to parents did not change their ability to hold the government schools accountable for results and hence could do little to improve outcomes. This is an example of incoherence between “client power” (parents to schools) and “management” (ministry to schools) in the RISE 4x4 framework of diagnostic for systems of basic education. The information via the community effort intended to strengthen the “client power” accountability relationship but it was incoherent with the information that the education bureaucracy in the “management” relationship collected and used for decision-making.
RISE Conference 2017: Information provision can improve learning outcomes under the right conditions
The role of information on learning to help improve education systems was a significant theme at the recent RISE Conference hosted at CGD.
Farzana Afridi kick-started the second day of the conference by presenting her findings from a study (with Bidisha Barooah and Rohini Somanathan) where information on performance was provided to households and/or schools in Ajmer, Rajasthan in India. The study had four different “treatments” (based on what kind of information was given) and the results are separated into “high” and “low” competition areas. While there was no improvement in test scores of public schools in any treatments, there was significant improvement in test scores of private school students when both parents and schools knew relative school quality—especially when there was already high competition and lots of available choices for parents.
The figure below shows that providing information led to large learning gains when information on relative school performance was provided to both parents and students, and when competition was high. These are of course exactly the conditions that simple theory of choice would suggest—people will change their child’s school to seek higher performance when provided information that makes the choice across schools salient and the choices are otherwise close substitutes.
Figure 1. Average treatment effects by school competition
Source: Afridi, F., Barooah, B., & Somanathan, R., (2017), as presented at RISE Conference 2017
Treatment 1: Parents receive information on intra-school performance
Treatment 2: Both parents and schools receive information on intra-school performance
Treatment 3: Schools receive information on intra- and inter-school performance and parents receive information on intra-school performance only
Treatment 4: Both parents and schools receive information on intra- and inter-school performance.
In contrast, none of the information “treatments” had a positive impact on learning in public schools—in fact, in low competition environments all four treatments appeared to lead to lower learning. This is consistent with early results that changing only one aspect of a relationship of accountability (information) without changing others (e.g., the motivation of the providers) is unlikely to succeed—particularly when there are other incoherent accountability relationships at work (e.g., teachers reporting to a bureaucracy).
In another conference session, Jishnu Das presented new results building on previous work with Asim Khwaja and Tahir Andrabi where providing information on performance improved test scores and decreased private school fees in Pakistan. What is particularly noteworthy is that the effects were sustained over a long period of time at relatively low cost.
Figure 2. 8-year effect of information in private schools
Source: Das, J., (2017), as presented at RISE Conference 2017
Looking back eight years later, the intervention: cost $2 per child, led to a $5.5 reduction in private school fees, and improved test scores for both private and public school students by 0.35 standard deviations and 0.12 standard deviation respectively. What we have here is a low-cost intervention with large, sustained results.
While all three studies mentioned study the impact of information provision on education outcomes, there are design differences across them. Banerjee et al. (2010) focused on public schools while Andrabi et al. (2016) and Afridi et al. (2017) focused on both public and private schools. Unlike Banerjee et al. (2010), Andrabi et al. (2016) and Afridi et al. (2017) provided information to both sides of the market, parents and schools. The results across the three studies varied as well. Banerjee et al. (2010) found no effect, the most recent results presented by Jishnu Das at the RISE Conference 2017 found that test scores increased by 0.35 standard deviations and 0.12 standard deviations for private and public school students respectively, and Afridi et al. (2017) found that test scores of private school students increased by 0.31 standard deviations with no impact on scores of public school students.
As always, design matters. Within Afridi et al.’s experiment, different treatment arms saw varying results. The improvements in test scores are highest when both parents and schools receive information about performance both within and across schools. The results seem to be mostly driven by high competition, as parents seem to move kids to better quality private schools. Giving report cards alone to schools didn’t make a big difference.
So what does all this mean? Does information work?
The answer is “it depends”—but the evidence appears to support common sense ways in which information provision “depends.” In a market-like environment where parents have lots of options, can vote with their money, and are receiving noisy signals of school performance, information matters. However, when parents are provided information but are constrained in their ability to exercise power over schools or teachers, information is likely to be ineffective.
Furthermore, design matters for construct validity. Different choices about the design elements of a specific intervention can have wide ranging consequences on the effectiveness of programs (Pritchett, L., 2017). Granular information about the design of experiments is key for determining their effectiveness. Statements like “information interventions don’t work” or “information interventions on average have small impact” based on a smattering of different experiments are useless, at best.
There are two main lessons
First, information improves outcomes when it improves accountability, and that requires designs that consider the realities of current accountability, who has power and why, and coherence within the existing system—in many existing systems teachers need not be accountable to parents at all.
Second, experiments are wonderful, but to understand and apply findings, knowing the experimental design and some theory is key.
This is one of a series of blog posts from “RISE"—the large-scale education systems research programme supported by the UK’s Department for International Development (DFID) and Australia’s Department of Foreign Affairs and Trade (DFAT). Experts from the Center for Global Development lead RISE’s research team.