Tag Archives: evaluation

Integrating big data into program evaluation: An invitation to participate in a short survey

As we all know, big data and data science are becoming increasingly important in all aspects of our lives. There is a similar rapid growth in the applications of big data in the design and implementation of development programs. Examples range from the use of satellite images and remote sensors in emergency relief and the identification of poverty hotspots, through the use of mobile phones to track migration and to estimate changes in income (by tracking airtime purchases), social media analysis to track sentiments and predict increases in ethnic tension, and using smart phones on Internet of Things (IOT) to monitor health through biometric indicators.

Despite the rapidly increasing role of big data in development programs, there is speculation that evaluators have been slower to adopt big data than have colleagues working in other areas of development programs. Some of the evidence for the slow take-up of big data by evaluators is summarized in “The future of development evaluation in the age of big data”.  However, there is currently very limited empirical evidence to test these concerns.

To try to fill this gap, my colleagues Rick Davies and Linda Raftree and I would like to invite those of you who are interested in big data and/or the future of evaluation to complete the attached survey. This survey, which takes about 10 minutes to complete asks evaluators to report on the data collection and data analysis techniques that you use in the evaluations you design, manage or analyze; while at the same time asking data scientists how familiar they are with evaluation tools and techniques.

The survey was originally designed to obtain feedback from participants in the MERL Tech conferences on “Exploring the Role of Technology in Monitoring, Evaluation, Research and Learning in Development” that are held annually in London and Washington, DC, but we would now like to broaden the focus to include a wider range of evaluators and data scientists.

One of the ways in which the findings will be used is to help build bridges between evaluators and data scientists by designing integrated training programs for both professions that introduce the tools and techniques of both conventional evaluation practice and data science, and show how they can be combined to strengthen both evaluations and data science research. “Building bridges between evaluators and big data analysts” summarizes some of the elements of a strategy to bring the two fields closer together.

The findings of the survey will be shared through this and other sites, and we hope this will stimulate a follow-up discussion. Thank you for your cooperation and we hope that the survey and the follow-up discussions will provide you with new ways of thinking about the present and potential role of big data and data science in program evaluation.

Here’s the link to the survey – please take a few minute to fill it out!

You can also join me, Kerry Bruce and Pete York on September 5th for a full day workshop on Big Data and Evaluation in Washington DC.

September 5th: MERL Tech DC pre-workshops

This year at MERL Tech DC, in addition to the regular conference on September 6th and 7th, we’re offering two full-day, in-depth workshops on September 5th. Join us for a deeper look into the possibilities and pitfalls of Blockchain for MERL and Big Data for Evaluation!

What can Blockchain offer MERL? with Shailee Adinolfi, Michael Cooper, and Val Gandhi, co-hosted by Chemonics International, 1717 H St. NW, Washington, DC 20016. 

Tired of the blockchain hype, but still curious on how it will impact MERL? Join us for a full day workshop with development practitioners who have implemented blockchain solutions with social impact goals in various countries. Gain knowledge of the technical promises and drawbacks of blockchain technology as it stands today and brainstorm how it may be able to solve for some of the challenges in MERL in the future. Learn about ethical design principles for blockchain and how to engage with blockchain service providers to ensure that your ideas and programs are realistic and avoid harm. See the agenda here.

Register now to claim a spot at the blockchain and MERL pre-workshop!

Big Data and Evaluation with Michael Bamberger, Kerry Bruce and Peter York, co-hosted by the Independent Evaluation Group at the World Bank – “I” Building, Room: I-1-200, 1850 I St NW, Washington, DC 20006

Join us for a one-day, in-depth workshop on big data and evaluation where you’ll get an introduction to Big Data for Evaluators. We’ll provide an overview of applications of big data in international development evaluation, discuss ways that evaluators are (or could be) using big data and big data analytics in their work. You’ll also learn about the various tools of data science and potential applications, as well as run through specific cases where evaluators have employed big data as one of their methods. We will also address the important question as to why many evaluators have been slower and more reluctant to incorporate big data into their work than have their colleagues in research, program planning, management and other areas such as emergency relief programs. Lastly, we’ll discuss the ethics of using big data in our work. See the agenda here!

Register now to claim a spot at the Big Data and Ealuation pre-workshop!

You can also register here for the main conference on September 6-7, 2018!

 

How MERL Tech Jozi helped me bridge my own data gap

Guest post from Praekelt.org. The original post appeared on August 15 here.

Our team had the opportunity to enjoy a range of talks at the first ever MERL Tech in Johannesburg. Here are some of their key learnings:

During “Designing the Next Generation of MERL Tech Software” by Mobenzi’s CEO Andi Friedman, we were challenged to apply design thinking techniques to critique both our own as well as our partners’ current projects. I have previously worked on an educational tool that is aimed to improve the quality of learning of students who are based in a disadvantaged community in the Eastern Cape, South Africa. I learned that language barriers are a serious concern when it comes to effectively implementing a new tool.

We mapped out a visual representation of solving a communication issue that one of the partners had for an educational programme implemented in rural Eastern Cape, which included drawing various shapes on paper. What we came up with was to replace the posters that had instructions in words with clear visuals that the students were familiar with. This was inspired by the idea that visuals resonate with people more than words.

-Perez Mnkile, Project Manager

Amy Green Presenting on Video Metrics

I really enjoyed the presentation on video metrics from Girl Effect’s Amy Green. She spoke to us about video engagement on Hara Huru Dara, a vlog series featuring social media influencers. What I found really interesting is how hard it is to measure impact or engagement. Different platforms (YouTube vs Facebook) have different definitions for various measurements (e.g. views) and also use a range of algorithms to reach these measurements. Her talk really helped me understand just how hard MERL can be in a digital age! As our projects expand into new technologies, I’ll definitely be more aware of how complicated seemingly simple metrics (for example, views on a video) may be.

-Jessica Manim, Project Manager

Get it right by getting it wrong: embracing failure as a tool for learning and improvement was a theme visible throughout the two day MERL Tech conference and one session highlighting this theme was conducted by Annie Martin a Research Associate at Akros, who explored challenges in Offline Data Capture.

She referenced a project that took place in Zambia to track participants of an HIV prevention program, highlighting some of the technical challenges the project faced along the way. The project involved equipping field workers with an Android tablet and an Application developed for capturing offline data and synching data, when data connectivity was available. A number of bugs due to insufficient system user testing along with server hosting issues resulted in field workers often not successfully being able to send data or create user IDs.

The lesson, which I believe we strive to include in our developmental processes, is to focus on iterative piloting, testing and learning before deployment. This doesn’t necessarily mean that a bug-free system or service is guaranteed but it does encourage us to focus our attention on the end-users and stakeholders needs, expectations and requirements.

-Neville Tietz, Service Designer

Slide from Panel on WhatsApp and engagement

Sometimes, we don’t fully understand the problems that we are trying to solve. Siziwe Ngcwabe from the African Evidence Network gave the opening talk on evidence-based work. It showed me the importance of fully understanding the problem we are solving and identifying the markers of success or failure before we start rolling out solutions. Once we have established all this, we can then create effective solutions. Rachel Sibande from DIAL, gave a talk on how their organisation is now using data from mobile network providers to anticipate how a disease outbreak will spread, based on the movement patterns of the network’s subscribers. Using this data they can advise ministries to run campaigns in certain areas and increase medical supplies in another. The talk by Siziwe showed me the importance of fully understanding the problem you are trying to solve and how to effectively measure progress. Rachel’s talk really showed me how easy it is to create an effective solution, once you fully understand the problem.

-Katlego Maakane, Project Manager

I really enjoyed the panel discussion on Datafication Discrimination with William Bird, Director of Media Monitoring Africa, Richard Gevers, Director of Open Data Durban, Koketso Moeti, Executive Director of amandla.mobi that was moderated by Siphokazi Mthathi, Executive Director of Oxfam South Africa. The impact that the mass collection of data can have on communities can potentially be used to further discriminate against them, especially when they are not aware on what their data will be used for. For example, information around sexuality can be used to target individuals during a time when there is rapid reversing of anti-discrimination laws in many countries.

I also thought it was interesting how projection models for population movement and the planning of new areas for residential development and public infrastructure in cities in South Africa are flawed, since the development of these models are outsourced by government to the private sector and different government departments often use different forecasts. Essentially the various government departments are all planning cities with different projections further preventing the poorest people from accessing quality services and infrastructure.

For me this session really highlighted the responsibility we have when collecting data in our projects from vulnerable individuals and that we have to ensure that we interrogate what we intend to use this data for. As part of our process, we must investigate how the data could potentially be exploited. We need to empower people to take control of the information they share and be able to make decisions in their best interest.

-Benjamin Vermeulen, Project Manager

Evaluating for Trust in Blockchain Applications

by Mike Cooper

This is the fourth in a series of blogs aimed at discussing and soliciting feedback on how the blockchain can benefit MEL practitioners in their work.  The series includes: What does Blockchain Offer to MERL,  Blockchain as an M&E Tool, How Can MERL Inform Maturation of the Blockchain, this post, and future posts on integrating  blockchain into MEL practices. The series leads into a MERL Tech Pre-Workshop on September 5th, 2018 in Washington D.C.  that will go into depth on possibilities and examples of MEL blockchain applications. Register here!

Enabling trust in an efficient manner is the primary innovation that the blockchain delivers through the use of cryptology and consensus algorithms.  Trust is usually a painstaking relationship building effort that requires iterative interactions to build.  The blockchain alleviates the need for much of the resources required to build this trust, but that does not mean that stakeholders will automatically trust the blockchain application.  There will still need to be trust building mechanisms with any blockchain application and MEL practitioners are uniquely situated to inform how these trust relationships can mature.

Function of trust in the blockchain

Trust is expensive.  You pay fees to banks who provide confidence to sellers who take your debit card as payment and trust that they will receive funds for the transaction.  Agriculture buyers pay fees to third parties (who can certify that the produce is organic, etc.) to validate quality control on products coming through the value chain  Often sellers do not see the money from debit card transaction in their accounts automatically and agriculture actors perpetually face the pressures resulting from being paid for goods and/or services they provided weeks previously. The blockchain could alleviate much of these harmful effects by substituting trust in humans by trust in math.

We pay these third parties because they are trusted agents, and these trusted agents can be destructive rent seekers at times; creating profits that do not add value to the goods and services they work with. End users in these transactions are used to using standard payment services for utility bills, school fees, etc.  This history of iterative transactions has resulted in a level of trust in these processes. It may not be equitable but it is what many are used to and introducing an innovation like blockchain will require an understanding of how these processes are influencing stakeholders, their needs and how they might be nudged to trust something different like a blockchain application.  

How MEL can help understand and build trust

Just as microfinance introduced new methods of sending/receiving money and access to new financial services that required piloting different possible solutions to build this understanding, so will blockchain applications. This is an area where MEL can add value to achieving mass impact, by designing the methods to iteratively build this understanding and test solutions.  

MEL has done this before.  Any project that requires relationship building should be based on understanding the mindset and incentives for relevant actions (behavior) amongst stakeholders to inform the design of the “nudge” (the treatment) intended to shift behavior.

Many of the programs we work on as MEL practitioners involve various forms and levels of relationship building, which is essentially “trust”.  There have been many evaluations of relationship building whether it be in microfinance, agriculture value chains or policy reform.  In each case, “trust” must be defined as a behavior change outcome that is “nudged” based on the framing (mindset) of the stakeholder.  Meaning that each stakeholder, depending on their mindset and the required behavior to facilitate blockchain uptake, will require a customized nudge.  

The role of trust in project selection and design: What does that mean for MEL

Defining “trust” should begin during project selection/design.  Project selection and design criteria/due diligence are invaluable for MEL.  Many of the dimensions of evaluability assessments refer back to the work that is done in the project selection/design phrase (which is why some argue evaluability assessments are essentially project design tools).  When it comes to blockchain, the USAID Blockchain Primer provides some of the earliest thinking for how to select and design blockchain projects, hence it is a valuable resources for MEL practitioners who want to start thinking about how they will evaluate blockchain applications.  

What should we be thinking about?

Relationship building and trust are behaviors, hence blockchain theories of change should have outcomes stated as behavior changes by specific stakeholders (hence the value add of tools like stakeholder analysis and outcome mapping).  However, these Theories of Change (TOC) are only as good as what informs them, hence building a knowledge base of blockchain applications as well as previous lessons learned from evidence on relationship building/trust will be critical to developing a MEL Strategy for blockchain applications.  

If you’d like to discuss this and related aspects, join us on September 5th in Washington, DC, for a one-day workshop on “What can the blockchain offer MERL?”

Michael Cooper is a former Associate Director at Millennium Challenge Corporation and the U.S. State Dept in Policy and Evaluation.  He now heads Emergence, a firm that specializes in MEL and Blockchain services. He can be reached at emergence.cooper@gmail.com or through the Emergence website.

How can MERL inform maturation of the blockchain?

by Mike Cooper

This is the third in a series of blogs aimed at discussing and soliciting feedback on how the blockchain can benefit MEL practitioners in their work.  The series includes: What does Blockchain Offer to MERL,  Blockchain as an M&E Tool, this post, and future posts on evaluating for trust in Blockchain applications, and integrating  blockchain into MEL practices. The series leads into a MERL Tech Pre-Workshop on September 5th, 2018 in Washington D.C.  that will go into depth on possibilities and examples of MEL blockchain applications. Register here!

Technology solutions in development contexts can be runaway trains of optimistic thinking.  Remember the play pump, a low technology solution meant to provide communities with clean water as children play?  Or the Soccket, the soccer ball that was going to help kids learn to read at night? I am not disparaging these good intentions, but the need to learn the evidence from past failure is widely recognized. When it comes to the blockchain, possibly the biggest technological innovation on the social horizon, the learning captured in guidance like the Principles for Digital Development or Blockchain Ethical Design Frameworks, needs to not only be integrated into the design of blockchain applications but also into how MEL practitioners will need to assess this integration and test solutions.   Data driven feedback from MEL will help inform the maturation of human centered blockchain solutions that mitigate endless/pointless pilots which exhaust the political will of good natured partners and creates barriers to sustainable impact.

The Blockchain is new but we have a head start in thinking about it

The blockchain is an innovation, and it should be evaluated as such. True the blockchain could be revolutionary in its impact.  And yes this potential could grease the wheels of the runaway train thinking referenced above, but this potential does not moot the evidence we have around evaluating innovations.

Keeping the risk of the runaway train at bay includes MERL practitioners working with stakeholders to ask : is blockchain the right approach for this at all?  Only after determining the competitive advantage of the blockchain solutions over other possible solutions should MEL practitioners work with stakeholders to finalize design of the initial piloting.  The USAID Blockchain Primer is the best early thinking about this process and the criteria involved.  

Michael Quinn Patton and others have developed an expanded toolkit for MERL practitioners to best unpack the complexity of a project and design a MERL framework that responds to the decision making requirements on the scale up pathway.  Because the blockchain is an innovation, which by definition means there is less evidence on its application but great potential, it will require MEL frameworks that iteratively test and modify applications to inform the scale up pathway.  

The Principles for Digital Development highlight the need for iterative learning in technology driven solutions.  The overlapping regulatory, organizational and technological spheres further assist in unpacking the complexity using tools like Problem Driven Iterative Adaptation (PDIA) or other adaptive management frameworks that are well suited to testing innovations in each sphere.  

How Blockchain is different: Intended Impacts and Potential Spoilers

There will be intended and unintended outcomes from blockchain applications that MEL should account for.  This includes general intended outcomes of increased access to services and overall costs savings while “un-intended” outcomes include the creation of winners and losers.  

The primary intended outcomes that could be expected from blockchain applications are an increase in cost savings (by cutting out intermediaries) which results in increased access to whatever service/product (assuming any cost savings are re-invested in expanding access).  Or a possible increase in access that results from creating a service where none existed before (for example creating access to banking services in rural populations). Hence methods for measuring the specific type of cost savings and increased access that are already used could be applied with modification.  

However, the blockchain will be disruptive and when I say “un-intended” (using quotation marks) I do so because the cost savings from blockchain applications are the result of alleviating the need for some intermediaries or middlemen. These middlemen are third parties who could be some form of rent-seeker in providing a validation, accreditation, certification  or other type of service meant to communicate trust. For example, with m-Pesa,  banking loan and other services from banks were expanded to new populations. With a financial inclusion blockchain project these same services could be accessed by the same population but without the need for a bank, hence incurring a cost savings. However, as is well known in many a policy reform intervention, creating efficiencies usually means creating losers and in our example the losers are those previously offering the services that the blockchain makes more efficient.  

The blockchain can facilitate efficiencies, not elimination of all intermediary functions. With the introduction of any innovation, the need for new functions will emerge as old functions are mooted.  For example mPesa experienced substantial barriers in its early development until they began working with kiosk owners who, after being trained up, could demonstrate and explain mPesa to customers.  Hence careful iterative assessment of the ecosystem (similar to value chain mapping) to identify mooted functions (losers) and new functions (winners) is critical.

MERL practitioners have a value add in mitigating the negative effects from the creation of losers, who could become spoilers.  MERL practitioners have many analytical tools/skills that can not only help in identifying the potential spoilers (perhaps through various outcome mapping and stakeholder analysis tools) but also in mitigating any negative effects (creating user personas of potential spoilers to better assess how to incentivize targeted behavior changes).  Hence MEL might be uniquely placed to build a broader understanding amongst stakeholders on what the blockchain is, what it can offer and how to create a learning framework that builds trust in the solution.

Trust, the real innovation of blockchain

MERL is all about behavior change, because no matter the technology or process innovation,  it requires uptake and uptake requires behavior. Trust is a behavior, you trust that when you put your money in a bank it will be available for when you want to use it.  Without this behavior, stemming from a belief, there are runs on banks which in turn fail which further erodes trust in the banking system. The same could be said for paying money to a water or power utility and expecting that they will provide service, The more use, the more a relationship matures into a trustful one. But it does not take much to erode this trust even after the relationship is established, again think about how easy it is to cause a run on a bank or stop using a service provider.  

The real innovation of the blockchain is that it replaces the need for trust in humans (whether it is an individual or system of organizations) with trust in math. Just as any entity needs to build a relationship of trust with its targeted patrons, so will the blockchain have to develop a  relationship of trust not only with end users but with those within the ecosystem that could influence the impact of the blockchain solution to include beneficiaries and potential loser/spoilers.  This brings us back to the importance of understanding who these stakeholders are, how they will interact with and influence the blockchain, and their perspectives, needs and capacities.

MERL practitioners who wish to use blockchain will need to pick up the latest thinking in behavioral sciences to understand this “trust” factor for each stakeholder and integrate it into an adaptive management framework.  The next blog in this series will go into further detail about the role of “trust” when evaluating a blockchain application.  

The Blockchain is different — don’t throw the baby out with the bath water

There will inevitably be mountains of pressure go to “full steam ahead” (part of me wants to add “and damn the consequences”) without sufficient data driven due diligence and ethical review, since blockchain is the next new shiny thing.  MERL practitioners should not only be aware of this unfortunate certainty, but they also need to pro-actively consider their own informed strategy on how they will respond to this pressure. MERL practitioners are uniquely positioned to advocate for data driven decision making and provide the data necessary to steer clear of misapplication of blockchain solutions.  There are already great resources for MEL practitioners on the ethical criteria and design implications for blockchain solutions.

The potential impact of blockchain is still unknown but if current thinking is to be believed, the impact could be paradigm shifting.  Given this potential, getting the initial testing right to maximize learning will be critical to cultivating the political will, the buy-in, and the knowledge base to kick start something much bigger.  

If you’d like to discuss this and related aspects, join us on September 5th in Washington, DC, for a one-day workshop on “What can the blockchain offer MERL?”

Michael Cooper is a former Associate Director at Millennium Challenge Corporation and the U.S. State Dept in Policy and Evaluation.  He now heads Emergence, a firm that specializes in MEL and Blockchain services. He can be reached at emergence.cooper@gmail.com or through the Emergence website.

Blockchain as an M&E Tool

by Mike Cooper and Shailee Adinofi

This is the second in a series of blogs aimed at discussing and soliciting feedback on how the blockchain can benefit MEL practitioners in their work.  The series includes: What does Blockchain Offer to MERL, this post (Blockchain as an M&E Tool), and future posts on the use of MEL to inform Blockchain maturation, evaluating for trust in Blockchain applications, and integrating  blockchain into MEL Practices. The series leads into a MERL Tech Pre-Workshop on September 5th, 2018 in Washington D.C.  that will go into depth on possibilities and examples of MEL blockchain applications. Register here!

Introducing the Blockchain as an M&E Tool   

Blockchain is a technology that could transform many of the functions we now take for granted in our daily lives. It could change everything from supply chain management to trade to the Internet of Things (IOT), and possibly even serve as the backbone for the next evolution of the internet itself.  Within international development there have already been blockchain pilots for refugee assistance and financial inclusion (amongst others) with more varied pilots and scaled applications soon to come.

Technological solutions, however, need uptake in order for their effects to be truly known. This is no different for the blockchain. Technology solutions are not self-implementing — their uptake is dependent on social structures and human decision making.  Hence, while on paper the blockchain offers many benefits, the realization of these benefits in the monitoring, evaluation and learning (MEL) space requires close working with MEL practitioners to hear their concerns, excitement, and feedback on how the blockchain can best produce these benefits.

Blockchain removes intermediaries, thus increasing integrity

The blockchain is a data management tool for achieving data integrity, transparency, and addressing privacy concerns. It is a distributed software network of peer-to-peer transactions (data), which are validated through consensus, using pre-established rules. This can remove the need for a middleman or “intermediaries”, meaning that it can “disintermediate” the holders of a traditional MEL database, where data is stored and owned by a set of actors.  

Hence the blockchain solves two primary problems:

  1.   It reduces the need for “middlemen” (intermediaries) because it is peer-to-peer in nature.  For MEL, the blockchain may thus reduce the need for people to be involved in data management protocols, from data collection to dissemination, resulting in cost and time efficiencies.
  2.  The blockchain maintains data integrity (meaning that the data is immutable and is only shared in the intended manner) in a distributed peer-to-peer network where the reliability and trustworthiness of the network is inherent to the rules established in the consensus algorithms of the blockchain.  

So, what does this mean?  Simply put, a blockchain is a type of distributed immutable ledger or decentralized database that keeps continuously updated digital records of data ownership. Rather than having a central administrator manage a single database, a distributed ledger has a network of replicated databases, synchronized via the internet, and visible to anyone within the network (more on control of the network and who has access permissions below).

Advantages over Current Use of Centralized Data Management  

Distributed ledgers are much less vulnerable to loss of control over data integrity than current centralized data management systems. Loss of data integrity can happen in numerous ways, whether by hacking, manipulation or some other nefarious or accidental use.  Consider the multiple cases of political manipulation of census data as recorded in Poor Numbers: How We Are Misled by African Development Statistics and What to Do about It because census instruments are designed and census data analyzed/managed in a centralized fashion with little to no transparency.

Likewise, within the field of evaluation there has been increasing attention on p-hacking, where initial statistical results are manipulated on the back side to produce results more favorable to the original hypothesis.  Imagine if cleaned and anonymized data sets were put onto the blockchain where transparency, without sacrificing PII, makes p-hacking much more difficult (perhaps resulting in increased trust in data sets and their overall utility/uptake).

Centralized systems can have lost and/or compromised data (or loss of access) due to computer malfunctions or what we call “process malfunctions” where the bureaucratic control over the data builds artificially high barriers to access and subsequent use of the data by anyone outside the central sphere of control. This level of centralized control (as in the examples above regarding manipulation of census design/data and p-hacking) introduces the ability for data manipulation.

Computer malfunctions are mitigated by the blockchain because the data does not live in a central network hub but instead “lives’ in copies of the ledger that are distributed across every computer in the network. This lack of central control increases transparency. “Hashing” (a form of version control) ensures that any data manipulations in the blockchain are not included in the blockchain, meaning only a person with the necessary permissions can change the data on the chain. With the blockchain, access to information is as open, or closed, as is desired.  

How can we use this technology in MEL?

All MEL data must eventually find its way to a digital version of itself, whether it is entered from paper surveys or it goes through analytical software or straight into an Excel cell, with varying forms/rigor of quality control.  A benefit of blockchain is its compatibility with all digital data. It can include data files from all forms of data collection and analytical methods or software. Practitioners are free to collect data in whatever manner best suits their mandates with the blockchain becoming the data management tool at any point after collection, as the data can be uploaded to the blockchain at any point. Meaning data can be loaded directly by enumerators in the field or after additional cleaning/analysis.  

MEL has  specific data management challenges that the blockchain seems uniquely suited to overcome including 1. protection of Personally Identifiable Information (PII)/data integrity, 2. mitigating data management resource requirements, and 3. lowering barriers to end use through timely dissemination and increased access to reliable data.  

Let’s explore each of these below:

1. Increasing Protection and Integrity of Data: There might be a knee jerk reaction against increasing transparency in evaluation data management, given the prevalence of personally identifiable information (PII) and other sensitive data. Meeting internal quality control procedures for developing and sharing draft results is usually a long arduous process — even more so if delivering cleaned data sets.  Hence there might be hesitation in introducing new data management techniques given the priority given to the protection of PII balanced against the pressure to deliver data sets in a timely fashion.

However, we should learn a lesson from our counterparts in healthcare records management, one of the more PII and sensitive data laden data management fields in the world.  The blockchain has seen piloting in healthcare records management precisely because it is able to secure the integrity of sensitive data in such an efficient manner.

Imagine an evaluator completes a round of household surveys, the data is entered, cleaned and anonymized and the data files are ready to be sent to whomever the receiver is (funder, public data catalog, etc.)  The funder requires that the data uploaded to the blockchain is done using a Smart Contract. Essentially a Smart Contract is a set of “if……then” protocols on the Ethereum network (a specific type of blockchain) which can say “if all data has been cleaned of PII and is appropriately formatted….etc….etc…, it can be accepted onto the blockchain.”  If the requirements written into the Smart Contract are not met, the data is rejected and not uploaded to the blockchain (see point 2 below). So, in the case where proper procedures or best or preferred practices are not met, the data is not shared and remains safe within the confines of a (hopefully) secure and reliable centralized database.

This example demonstrates one of the unsung values of the blockchain. When correctly done (meaning the Smart Contract is properly developed) it can ensure that only the data that is appropriate is shared and is in fact shared only with those meant to have it in a manner where the data cannot be manipulated.  This is an advantage over current practice where human error can result in PII being released or unuseable or incompatible data files being shared.

The blockchain also has inherent quality control protocols around version control that mitigate against manipulation of the data for whatever reason. Hashing is partly a summary labelling of different encrypted data sets on the blockchain where any modification to the data set results in a different hash for that data set.  Hence version control is automatic and easily tracked through the different hashes which are one way only (meaning that once the data is hashed it cannot be reverse engineered to change the original data). Thus, all data on the blockchain is immutable.

2. Decreasing Data Management Resources: Current data management practice is very resource intensive for MEL practitioners.  Data entry, creation of data files, etc. requires ample amounts of time, mostly spent guarding against error, which introduces timeliness issues where processes take so long the data uses its utility by the time it is “ready” for decision makers.  A future post in this series will cover how the blockchain can introduce efficiencies at various points in the data management process (from collection to dissemination). There are many unknowns in this space that require further thinking about the ability to embed automated cleaning and/or analytical functions into the blockchain or compatibility issues around data files and software applications (like STATA or NIVIVO).  This series of posts will highlight broad areas where the blockchain can introduce the benefits of an innovation as well as finer points that still need to be “unpacked” for the benefits to materialize.

3. Distributed ledger enables timely dissemination in a flexible manner:  With the increased focus on the use of evaluation data, there has been a correlated increase in discussion in how evaluation data is shared.

Current data dissemination practices include:

  • depositing them with a data center, data archive, or data bank
  • submitting them to a journal to support a publication
  • depositing them in an institutional repository
  • making them available online via a project or institutional website
  • making them available informally between researchers on a peer-to-peer basis

All these avenues of dissemination are very resource intensive. Each avenue has its own procedures, protocols, and other characteristics that may not be conducive to timely learning. Timelines for publishing in journals is long with incentives towards only publishing positive results, contributing to a dismal utilization rates of results.  Likewise, many institutional evaluation catalogs are difficult to navigate, often incomplete, and generally not user friendly. (We will look at query capabilities on the blockchain later in the blog series).

Using the blockchain to manage and disseminate data could result in more timely and transparent sharing.  Practitioners could upload data to the chain at any point after collection, and with the use of Smart Contracts, data can be widely distributed in a controlled manner.  Data sets can be easily searchable and available in much timelier and user-friendly fashion to a much larger population. This creates the ability to share specific data with specific partners (funders, stakeholders, the general public) in a more automated fashion and on a timelier basis.  Different Smart Contracts can be developed so that funders can see all data as soon as it is collected in the field, while a different Smart Contract with local officials allows them to see data relevant to their locality only after it is entered, cleaned, etc.).

With the help of read/write protocols, anyone can control the extent to which data is shared. Use of the data is immutable, meaning it cannot be changed (in contrast to current practice where we hope the PDF is “good enough” to guard against modification but most times data are pushed out in excel sheets, or something similar, with no way to determine what the “real” data when different versions appear).   

Where are we?

We are in the early stages of understanding, developing and exploring the blockchain in general and with MEL in particular. On September 5th, we’ll be leading a day-long Pre-Conference Workshop on What Blockchain Can Do For MERL. The Pre-Conference Workshop and additions to this blog series will focus on how:

  • The blockchain can introduce efficiencies in MEL data management
  • The blockchain can facilitate “end use” whether it is accountability, developmental, formative, etc.
  • To work with MEL practitioners and other stakeholders to improve the uptake of the blockchain as an innovation by overcoming regulatory, organizational and cultural barriers.  

This process is meant to be collaborative so we invite others to help inform us on what issues they think warrant further exploration.  We look forward to collaborating with others to unpack these issues to help develop thinking that leads to appropriate uptake of blockchain solutions to MEL problems.  

Where are we going?

As it becomes increasingly possible that blockchain will be a disruptive technology, it is critical that we think about how it will affect the work of MEL practitioners.  To this end, stay tuned for a few more posts, including:

  • How can MEL inform Blockchain maturation?
  • Evaluating for Trust in Blockchain applications
  • How can we integrate blockchain into MEL Practices?

We would greatly benefit from feedback on this series to help craft topics that the series can cover.  Please comment below or contact the authors with any feedback, which would be greatly appreciated.

Register here for the September 5th workshop on Blockchain and MERL!

Michael Cooper is a former Associate Director at Millennium Challenge Corporation and the U.S. State Dept in Policy and Evaluation.  He now heads Emergence, a firm that specializes in MEL and Blockchain services. He can be reached at emergence.cooper@gmail.com or through the Emergence website.

Shailee Adinolfi is an international development professional with over 15 years of experience working at the intersection of financial services, technology, and global development. Recently, she performed business development, marketing, account management, and solution design as Vice President at BanQu, a Blockchain-based identity platform. She held a variety of leadership roles on projects related to mobile banking, financial inclusion, and the development of emerging markets. More about Shailee 

What does Blockchain offer to MERL?

by Shailee Adinolfi

By now you’ve read at least one article on the potential of blockchain, as well as the challenges in its current form. USAID recently published a Primer on Blockchain: How to assess the relevance of distributed ledger technology to international development, which explains that distributed ledgers are “a type of shared computer database that enables participants to agree on the state of a set of facts or events (frequently described as an “authoritative shared truth”) in a peer-to-peer fashion without needing to rely on a single, centralized, or fully trusted party”.

Explained differently, the blockchain introduces cost savings and resource efficiencies by allowing data to be entered, stored and shared in an immutable fashion by substituting the need for a trusted third party with algorithms and cryptography.

The blockchain/Distributed Ledger Technology (DLT) industry is evolving quickly, as are the definitions and terminology. Blockchain may not solve world hunger, but the core promises are agreed upon by many – transparency, auditability, resiliency, and streamlining. The challenges, which companies are racing to be the first to address, include scale (speed of transactions), security, and governance.

It’s not time to sit back wait and see what happens. It’s time to deepen our understanding. Many have already begun pilots across sectors. As this McKinsey article points out, early data from pilots shows strong potential in the Agriculture and Government sectors, amongst others. The article indicates that scale may be as little as 3-5 years away, and that’s not far out.

The Center for Global Development’s Michael Pisa argues that the potential benefits of blockchain do not outweigh the associated costs and complexities right now. He suggests that the development community focus its energies and resources on bringing down barriers to actual implementation, such as standards, interoperability, de-siloing data, and legal and regulatory rules around data storage, privacy and protection.

One area where blockchain may be useful is Monitoring, Evaluation, Research and Learning (MERL). But we need to dig in and understand better what the potentials and pitfalls are.

Join us on September 5th for a one-day workshop on Blockchain and MERL at Chemonics International where we will discuss what blockchain offers to MERL.

This is the first in a series of blogs at discussing and soliciting feedback on how the blockchain can benefit MEL practitioners in their work.  The series includes this post (What does Blockchain Offer to MERL),  Blockchain as an M&E Tool, and future posts on the use of MEL to inform Blockchain maturation, evaluating for trust in Blockchain applications, and integrating  blockchain into MEL Practices. 

 

Integrating Big Data into Evaluation: a conversation with Michael Bamberger and Rick Davies

At MERL Tech London, 2018, we invited Michael Bamberger and Rick Davies to debate the question of whether the enthusiasm for Big Data in Evaluation is warranted. At their session, through a formal debate (skillfully managed by Shawna Hoffman from The Rockefeller Foundation) they discussed whether Big Data and Evaluation would eventually converge, whether one would dominate the other, how can and should they relate to each other, and what risks and opportunities there are in this relationship.

Following the debate, Michael and Risk wanted to continue the discussion — this time exploring the issues in a more conversational mode on the MERL Tech Blog, because in practice both of them see more than one side to the issue.

So, what do Rick and Michael think — will big data integrate with evaluation — or is it all just hype?

Rick: In the MERL Tech debate I put a lot of emphasis on the possibility that evaluation, as a field, would be overwhelmed by big data / data science rhetoric. But since then I have been thinking about a countervailing development, which is that evaluative thinking is pushing back against unthinking enthusiasm for the use of data science algorithms. I emphasise “evaluative thinking” rather than “evaluators” as a category of people, because a lot of this pushback is coming from people who would not identify themselves as evaluators. There are different strands to this evaluative response.

One is a social justice perspective, reflected in recent books such as “Weapons of Math Destruction”, “Automated Inequality”, and “Algorithms of Oppression” which emphasise the human cost of poorly designed and or poorly supervised use of algorithms using large amounts of data to improve welfare and justice administration. Another strand is more like a form of exploratory philosophy, and has focused on how it might be possible to define “fairness” when designing and evaluating algorithms that have consequences for human welfare[ See 1, 2, 3, 4]. Another strand is perhaps more technical in focus, but still has a value concern. This is the literature on algorithmic transparency. Without transparency it is difficult to assess fairness [See 5, 6, ] Neural networks are often seen as a particular challenge. Associated with this are discussions about “the right to explanation” and what this means in practice[1,]

In parallel there is also some infiltration of data science thinking into mainstream evaluation practice. DFID is funding the World Bank’s Strategic Impact Evaluation Fund (SIEF) latest call for “nimble evaluations” [7]. These are described as rapid and low cost and likely to take the form of an RCT but ones which are focused on improving implementation rather than assessing overall impact [8]. This type of RCT is directly equivalent to A/B testing used by the internet giants to improve the way their platforms engage with their users. Hopefully these nimble approaches may bring a more immediate benefit to the people’s lives than RCTs which have tried to assess the impact of a whole project and then inform the design of subsequent projects.

Another recent development is the World Bank’s Data Science competition [9], where participants are being challenged to develop predictive models of household poverty status, based on World Bank Household Survey data.  The intention is that they should provide a cheaper means of identifying poor households than simply relying on what can be very expensive and time consuming nationwide household surveys. At present the focus on the supporting website is very technical. As far as I can see there is no discussion of how the winning prediction model will be used and an how any risks of adverse effects might be monitored and managed.  Yet as I suggested at MERLTech London, most algorithms used for prediction modelling will have errors. The propensity to generate False Positives and False Negatives is machine learning’s equivalent of original sin. It is to be expected, so it should be planned for. Plans should include systematic monitoring of errors and a public policy for correction, redress and compensation.

Michael:  These are both important points, and it is interesting to think what conclusions we can draw for the question before us.  Concerning the important issue of algorithmic transparency (AT), Rick points out that a number of widely discussed books and articles have pointed out the risk that the lack of AT poses for democracy and particularly for poor and vulnerable groups. Virginia Eubanks, one of the authors cited by Rick, talks about the “digital poorhouse” and how unregulated algorithms can help perpetuate an underclass.  However, I think we should examine more carefully how evaluators are contributing to this discussion. My impression, based on very limited evidence is that evaluators are not at the center — or even perhaps the periphery — of this discussion. Much of the concern about these issues is being generated by journalists, public administration specialists or legal specialists.  I argued in an earlier MERL Tech post that many evaluators are not very familiar with big data and data analytics and are often not very involved in these debates.  This is a hypothesis that we hope readers can help us to test.

Rick’s second point, about the infiltration of data science into evaluation is obviously very central to our discussion.  I would agree that the World Bank is one of the leaders in the promotion of data science, and the example of “nimble evaluation” may be a good example of convergence between data science and evaluation.  However, there are other examples where the Bank is on the cutting edge of promoting new information technology, but where potential opportunities to integrate technology and evaluation do not seem to have been taken up.  An example would be the Bank’s very interesting Big Data Innovation Challenge, which produced many exciting new applications of big data to development (e.g. climate smart agriculture, promoting financial inclusion, securing property rights through geospatial data, and mapping poverty through satellites). The use of data science to strengthen evaluation of the effectiveness of these interventions, however, was not mentioned as one of the objectives or outputs of this very exciting program.  

It would also be interesting to explore to what extent the World Bank Data Science competition that Rick mentions resulted in the convergence of data science and evaluation, or whether it was simply testing new applications of data science.

Finally, I would like to mention two interesting chapters in Cybersociety, Big Data and Evaluation edited by Petersson and Breul (2017, Transaction Publications).  One chapter (by Hojlund et al) reports on a survey which found that only 50% of professional evaluators claimed to be familiar with the basic concepts of big data, and only about 10% reported having used big data in an evaluation.  In another chapter, Forss and Noren reviewed a sample of Terms of Reference (TOR) for evaluations conducted by different development agencies, where they found that none of the 25 TOR specifically required the evaluators to incorporate big data into their evaluation design.

It is difficult to find hard evidence on the extent to which evaluators are familiar with, sympathetic to, or using big data into their evaluations, but the examples mentioned above show that there are important questions about the progress made towards the convergence of evaluation and big data.  

We invite readers to share their experiences both on how the two professions are starting to converge, or on the challenges that slow down, or even constrain the process of convergence.

Take our survey on Big Data and Evaluation!

Or sign up for Michael’s full-day workshop on Big Data and Evaluation in Washington, DC, on September 5th, 2018! 

Big data or big hype: a MERL Tech debate

by Shawna Hoffman, Specialist, Measurement, Evaluation and Organizational Performance at the Rockefeller Foundation.

Both the volume of data available at our fingertips and the speed with which it can be accessed and processed have increased exponentially over the past decade.  The potential applications of this to support monitoring and evaluation (M&E) of complex development programs has generated great excitement.  But is all the enthusiasm warranted?  Will big data integrate with evaluation — or is this all just hype?

A recent debate that I chaired at MERL Tech London explored these very questions. Alongside two skilled debaters (who also happen to be seasoned evaluators!) – Michael Bamberger and Rick Davies – we sought to unpack whether integration of big data and evaluation is beneficial – or even possible.

Before we began, we used Mentimeter to see where the audience  stood on the topic:

Once the votes were in, we started.

Both Michael and Rick have fairly balanced and pragmatic viewpoints; however, for the sake of a good debate, and to help unearth the nuances and complexity surrounding the topic, they embraced the challenge of representing divergent and polarized perspectives – with Michael arguing in favor of integration, and Rick arguing against.

“Evaluation is in a state of crisis,” Michael argued, “but help is on the way.” Arguments in favor of the integration of big data and evaluation centered on a few key ideas:

  • There are strong use cases for integration. Data science tools and techniques can complement conventional evaluation methodology, providing cheap, quick, complexity-sensitive, longitudinal, and easily analyzable data.
  • Integration is possible. Incentives for cross-collaboration are strong, and barriers to working together are reducing. Traditionally these fields have been siloed, and their relationship has been characterized by a mutual lack of understanding of the other (or even questioning of the other’s motivations or professional rigor).  However, data scientists are increasingly recognizing the benefits of mixed methods, and evaluators are seeing the potential to use big data to increase the number of types of evaluation that can be conducted within real-world budget, time and data constraints. There are some compelling examples (explored in this UN Global Pulse Report) of where integration has been successful.
  • Integration is the right thing to do.  New approaches that leverage the strengths of data science and evaluation are potentially powerful instruments for giving voice to vulnerable groups and promoting participatory development and social justice.   Without big data, evaluation could miss opportunities to reach the most rural and remote people.  Without evaluation (which emphasizes transparency of arguments and evidence), big data algorithms can be opaque “black boxes.”

While this may paint a hopeful picture, Rick cautioned the audience to temper its enthusiasm. He warned of the risk of domination of evaluation by data science discourse, and surfaced some significant practical, technical, and ethical considerations that would make integration challenging.

First, big data are often non-representative, and the algorithms underpinning them are non-transparent. Second, “the mechanistic approaches offered by data science, are antithetical to the very notion of evaluation being about people’s values and necessarily involving their participation and consent,” he argued. It is – and will always be – critical to pay attention to the human element that evaluation brings to bear. Finally, big data are helpful for pattern recognition, but the ability to identify a pattern should not be confused with true explanation or understanding (correlation ≠ causation). Overall, there are many problems that integration would not solve for, and some that it could create or exacerbate.

The debate confirmed that this question is complex, nuanced, and multi-faceted. It helped to remind that there is cause for enthusiasm and optimism, at the same time as a healthy dose of skepticism. What was made very clear is that the future should leverage the respective strengths of these two fields in order to maximize good and minimize potential risks.

In the end, the side in favor of integration of big data and evaluation won the debate by a considerable margin.

The future of integration looks promising, but it’ll be interesting to see how this conversation unfolds as the number of examples of integration continues to grow.

Interested in learning more and exploring this further? Stay tuned for a follow-up post from Michael and Rick. You can also attend MERL Tech DC in September 2018 if you’d like to join in the discussions in person!

Evaluating ICT4D projects against the Digital Principles

By Laura Walker McDonald,  This post was originally published on the Digital Impact Alliance’s Blog on March 29, 2018.

As I have written about elsewhere, we need more evidence of what works and what doesn’t in the ICT4D and tech for social change spaces – and we need to hold ourselves to account more thoroughly and share what we know so that all of our work improves. We should be examining how well a particular channel, tool or platform works in a given scenario or domain; how it contributes to development goals in combination with other channels and tools; how the team selected and deployed it; whether it is a better choice than not using technology or using a different sort of technology; and whether or not it is sustainable.

At SIMLab, we developed our Framework for Monitoring and Evaluation of Technology in Social Change projects to help implementers to better measure the impact of their work. It offers resources towards a minimum standard of best practice which implementers can use or work toward, including on how to design and conduct evaluations. With the support of the Digital Impact Alliance (DIAL), the resource is now finalized and we have added new evaluation criteria based on the Principles for Digital Development.

Last week at MERL Tech London, DIAL was able to formally launch this product by sharing a 2-page summary available at the event and engaging attendees in a conversation about how it could be used. At the event, we joined over 100 organizations to discuss Monitoring, Evaluation, Research and Learning related to technology used for social good.

Why evaluate?

Evaluations provide snapshots of the ongoing activity and the progress of a project at a specific point in time, based on systematic and objective review against certain criteria. They may inform future funding and program design; adjust current program design; or to gather evidence to establish whether a particular approach is useful. They can be used to examine how, and how far, technology contributes to wider programmatic goals. If set up well, your program should already have evaluation criteria and research questions defined, well before it’s time to commission the evaluation.

Evaluation criteria provide a useful frame for an evaluation, bringing in an external logic that might go beyond the questions that implementers and their management have about the project (such as ‘did our partnerships on the ground work effectively?’ or ‘how did this specific event in the host country affect operations?’) to incorporate policy and best practice questions about, for example, protection of target populations, risk management, and sustainability. The criteria for an evaluation could be any set of questions that draw on an organization’s mission, values, principles for action; industry standards or other best practice guidance; or other thoughtful ideas of what ‘good’ looks like for that project or organization. Efforts like the Principles for Digital Development can set useful standards for good practice, and could be used as evaluation criteria.

Evaluating our work, and sharing learning, is radical – and critically important

While the potential for technology to improve the lives of vulnerable people around the world is clear, it is also evident that these improvements are not keeping pace with the advances in the sector. Understanding why requires looking critically at our work and holding ourselves to account. There is still insufficient evidence of the contribution technology makes to social change work. What evidence there is often is not shared or the analysis doesn’t get to the core issues. Even more important, the learnings from what has not worked and why have not been documented and absorbed.

Technology-enabled interventions succeed or fail based on their sustainability, business models, data practices, choice of communications channel and technology platform; organizational change, risk models, and user support – among many other factors. We need to build and examine evidence that considers these issues and that tells us what has been successful, what has failed, and why. Holding ourselves to account against standards like the Principles is a great way to improve our practice, and honor our commitment to the people we seek to help through our work.

Using the Digital Principles as evaluation criteria

The Principles for Digital Development are a set of living guidance intended to help practitioners succeed in applying technology to development programs. They were developed, based on some pre-existing frameworks, by a working group of practitioners and are now hosted by the Digital Impact Alliance.

These nine principles could also form a useful set of evaluation criteria, not unlike OECD evaluation criteria, or Sphere standards. Principles overlap, so data can be used to examine more than one criterion, and ot every evaluation would need to consider all of the Digital Principles.

Below are some examples of Digital Principles and sample questions that could initiate, or contribute to, an evaluation.

Design with the User: Great projects are designed with input from the stakeholders and users who are central to the intended change. How far did the team design the project with its users, based on their current tools, workflows, needs and habits, and work from clear theories of change and adaptive processes?

Understand the Existing Ecosystem: Great projects and programs are built, managed, and owned with consideration given to the local ecosystem. How far did the project work to understand the local, technology and broader global ecosystem in which the project is situated? Did it build on existing projects and platforms rather than duplicating effort? Did the project work sensitively within its ecosystem, being conscious of its potential influence and sharing information and learning?

Build for Sustainability: Great projects factor in the physical, human, and financial resources that will be necessary for long-term sustainability. How far did the project: 1) think through the business model, ensuring that the value for money and incentives are in place not only during the funded period but afterwards, and 2) ensure that long-term financial investments in critical elements like system maintenance and support, capacity building, and monitoring and evaluation are in place? Did the team consider whether there was an appropriate local partner to work through, hand over to, or support the development of, such as a local business or government department?

Be Data Driven: Great projects fully leverage data, where appropriate, to support project planning and decision-making. How far did the project use real-time data to make decisions, use open data standards wherever possible, and collect and use data responsibly according to international norms and standards?

Use Open Standards, Open Data, Open Source, and Open Innovation: Great projects make appropriate choices, based on the circumstances and the sensitivity of their project and its data, about how far to use open standards, open the project’s data, use open source tools and share new innovations openly. How far did the project: 1) take an informed and thoughtful approach to openness, thinking it through in the context of the theory of change and considering risk and reward, 2) communicate about what being open means for the project, and 3) use and manage data responsibly according to international norms and standards?

For a more complete set of guidance, see the complete Framework for Monitoring and Evaluating Technology, and the more nuanced and in-depth guidance on the Principles, available on the Digital Principles website.