Posts Tagged ‘evaluation’

Traditional development evaluation has been characterized as ‘backward looking’ rather than forward looking and too focused on proving over improving. Some believe applying an ‘agile’ approach in development would be more useful — the assumption being that if you design a program properly and iterate rapidly and constantly based on user feedback and data analytics, you are more likely achieve your goal or outcome without requiring expensive evaluations. The idea is that big data could eventually allow development agencies to collect enough passive data about program participants that there would no longer be a need to actively survey people or conduct a final evaluation, because there would be obvious patterns that would allow implementers to understand behaviors and improve programs along the way.

The above factors have made some evaluators and data scientists question whether big data and real-time availability of multiple big data sets, along with the technology that enables their collection and analysis, will make evaluation as we know it obsolete. Others have argued that it’s not the end of evaluation, but rather we will see a blending of real-time monitoring, predictive modeling, and impact evaluation, depending on the situation. Big questions remain, however, about the feasibility of big data in some contexts. For example, are big data approaches useful when it comes to people who are not producing very much digital data? How will the biases in big data be addressed to ensure that the poorest, least connected, and/or most marginalized are represented?

The Technology Salon on Big Data and Evaluation hosted during November’s  American Evaluation Association Conference in Chicago opened these questions up for consideration by a roomful of evaluators and a few data scientists. We discussed the potential role of new kinds and quantities of data. We asked how to incorporate static and dynamic big data sources into development evaluation. We shared ideas on what tools, skills, and partnerships we might require if we aim to incorporate big data into evaluation practice. This rich and well-informed conversation was catalyzed by our lead discussants: Andrew Means, Associate Director of the Center for Data Science & Public Policy at the University of Chicago and Founder of Data Analysts for Social Good and The Impact Lab; Michael Bamberger, Independent Evaluator and co-author of Real World Evaluation; and Veronica Olazabal from The Rockefeller Foundation. The Salon was supported by ITAD via a Rockefeller Foundation grant.

What do we mean by ‘big data’?

The first task was to come up with a general working definition of what was understood by ‘big data.’ Very few of the organizations present at the Salon were actually using ‘big data’ and definitions varied. Some talked about ‘big data sets’ as those that could not be collected or analyzed by a human on a standard computer. Others mentioned that big data could include ‘static’ data sets (like government census data – if digitized — or cellphone record data) and ‘dynamic’ data sets that are being constantly generated in real time (such as streaming data input from sensors or ‘cookies’ and ‘crumbs’ generated through use of the Internet and social media). Others considered big data to be real time, socially-created and socially-driven data that could be harvested without having to purposely collect it or budget for its collection. ‘It’s data that has a life of its own. Data that just exists out there.’ Yet others felt that for something to be ‘big data’ multiple big data sets needed to be involved, for example, genetic molecular data crossed with clinical trial data and other large data sets, regardless of static or dynamic nature. Big data, most agreed, is data that doesn’t easily fit on a laptop and that requires a specialized skill set that most social scientists don’t have. ‘What is big data? It’s hard to define exactly, but I know it when I see it,’ concluded one discussant.

Why is big data a ‘thing’?

As one discussant outlined, recent changes in technology have given rise to big data. Data collection, data storage and analytical power are becoming cheaper and cheaper. ‘We live digitally now and we produce data all the time. A UPS truck has anywhere from 50-75 sensors on it to do everything from optimize routes to indicate how often it visits a mechanic,’ he said. ‘The analytic and computational power in my iPhone is greater than what the space shuttle had.’ In addition, we have ‘seamless data collection’ in the case of Internet-enabled products and services, meaning that a person creates data as they access products or services, and this can then be monetized, which is how companies like Google make their money. ‘There is not someone sitting at Google going — OK, Joe just searched for the nearest pizza place, let me enter that data into the system — Joe is creating the data about his search while he is searching, and this data is a constant stream.’

What does big data mean for development evaluation?

Evaluators are normally tasked with making a judgment about the merit of something, usually for accountability, learning and/or to improve service delivery, and usually looking back at what has already happened. In the wider sense, the learning from evaluation contributes to program theory, needs assessment, and many other parts of the program cycle.

This approach differs in some key ways from big data work, because most of the new analytical methods used by data scientists are good at prediction but not very good at understanding causality, which is what social scientists (and evaluators) are most often interested in. ‘We don’t just look at giant data sets and find random correlations,’ however, explained one discussant. ‘That’s not practical at all. Rather, we start with a hypothesis and make a mental model of how different things might be working together. We create regression models and see which performs better. This helps us to know if we are building the right hypothesis. And then we chisel away at that hypothesis.’

Some challenges come up when we think about big data for development evaluation because the social sector lacks the resources of the private sector. In addition, data collection in the world of international development is not often seamless because ‘we care about people who do not live in the digital world,’ as one person put it. Populations we work with often do not leave a digital trail. Moreover, we only have complete data about the entire population in some cases (for example, when it comes to education in the US), meaning that development evaluators need to figure out how to deal with bias and sampling.

Satellite imagery can bring in some data that was unavailable in the past, and this is useful for climate and environmental work, but we still do not have a lot of big data for other types of programming, one person said. What’s more, wholly machine-based learning, and the kind of ‘deep learning’ made possible by today’s computational power is currently not very useful for development evaluation.

Evaluators often develop counterfactuals so that they can determine what would have happened without an intervention. They may use randomized controlled trials (RCTs), differentiation models, statistics and economics research approaches to do this. One area where data science may provide some support is in helping to answer questions about counterfactuals.

More access to big data (and open data) could also mean that development and humanitarian organizations stop duplicating data collection functions. Perhaps most interestingly, big data’s predictive capabilities could in the future be used in the planning phase to inform the kinds of programs that agencies run, where they should be run, and who should be let into them to achieve the greatest impact, said one discussant. Computer scientists and social scientists need to break down language barriers and come together more often so they can better learn from one another and determine where their approaches can overlap and be mutually supportive.

Are we all going to be using big data?

Not everyone needs to use big data. Not everyone has the capacity to use it, and it doesn’t exist for offline populations, so we need to be careful that we are not forcing it where it’s not the best approach. As one discussant emphasized, big data is not magic, and it’s not universally applicable. It’s good for some questions and not others, and it should be considered as another tool in the toolbox rather than the only tool. Big data can provide clues to what needs further examination using other methods, and thus most often it should be part of a mixed methods approach. Some participants felt that the discussion about big data was similar to the one 20 years ago on electronic medical records or to the debate in the evaluation community about quantitative versus qualitative methods.

What about groups of people who are digitally invisible?

There are serious limitations when it comes to the data we have access to in the poorest communities, where there are no tablets and fewer cellphones. We also need to be aware of ‘micro-exclusion’ (who within a community or household is left out of the digital revolution?) and intersectionality (how do different factors of exclusion combine to limit certain people’s digital access?) and consider how these affect the generation and interpretation of big data. There is also a question about the intensity of the digital footprint: How much data and at what frequency is it required for big data to be useful?

Some Salon participants felt that over time, everyone would have a digital presence and/or data trail, but others were skeptical. Some data scientists are experimenting with calibrating small amounts of data and comparing them to human-collected data in an attempt to make big data less biased, a discussant explained. Another person said that by digitizing and validating government data on thousands (in the case of India, millions) of villages, big data sets could be created for those that are not using mobiles or data.

Another person pointed out that generating digital data is a process that involves much more than simple access to technology. ‘Joining the digital discussion’ also requires access to networks, local language content, and all kinds of other precursors, she said. We also need to be very aware that these kinds of data collection processes impact on people’s participation and input into data collection and analysis. ‘There’s a difference between a collective evaluation activity where people are sitting around together discussing things and someone sitting in an office far from the community getting sound bites from a large source of data.’

Where is big data most applicable in evaluation?

One discussant laid out areas where big data would likely be the most applicable to development evaluation:

Screen Shot 2015-11-23 at 9.32.07 AM

It would appear that big data has huge potential in the evaluation of complex programs, he continued. ‘It’s fairly widely accepted that conventional designs don’t work well with multiple causality, multiple actors, multiple contextual variables, etc. People chug on valiantly, but it’s expected that you may get very misleading results. This is an interesting area because there are almost no evaluation designs for complexity, and big data might be a possibility here.’

In what scenarios might we use big data for development evaluation?

This discussant suggested that big data might be considered useful for evaluation in three areas:

  1. Supporting conventional evaluation design by adding new big data generated variables. For example, one could add transaction data from ATMs to conventional survey generated poverty indicators
  2. Increasing the power of a conventional evaluation design by using big data to strengthen the sample selection methodology. For example, satellite images were combined with data collected on the ground and propensity score matching was used to strengthen comparison group selection for an evaluation of the effects of interventions on protecting forest cover in Mexico.
  3. Replacing a conventional design with a big data analytics design by replacing regression based models with systems analysis. For example, one could use systems analysis to compare the effectiveness of 30 ongoing interventions that may reduce stunting in a sample of villages. Real-time observations could generate a time-series that could help to estimate the effectiveness of each intervention in different contexts.

It is important to remember construct validity too. ‘If big data is available, but it’s not quite answering the question that you want to ask, it might be easy to decide to do something with it, to run some correlations, and to think that maybe something will come out. But we should avoid this temptation,’ he cautioned. ‘We need to remember and respect construct validity and focus on measuring what we think we are measuring and what we want to measure, not get distracted by what a data set might offer us.’

What about bias in data sets?

We also need to be very aware that big data carries with it certain biases that need to be accounted for, commented several participants; notably, when working with low connectivity populations and geographies or when using data from social media sites that cater to a particular segment of the population. One discussant shared an example where Twitter was used to identify patterns in food poisoning, and suddenly the upscale, hipster restaurants in the city seemed to be the problem. Obviously these restaurants were not the sole source of the food poisoning, but rather there was a particular kind of person that tended to use Twitter.

‘People are often unclear about what’s magical and what’s really possible when it comes to big data. We want it to tell us impossible things and it can’t. We really need to engage human minds in this process; it’s not a question of everything being automated. We need to use our capacity for critical thinking and ask: Who’s creating the data? How’s it created? Where’s it coming from? Who might be left out? What could go wrong?’ emphasized one discussant. ‘Some of this information can come from the metadata, but that’s not always enough to make certain big data is a reliable source.’ Bias may also be introduced through the viewpoints and unconscious positions, values and frameworks of the data scientists themselves as they are developing algorithms and looking for/finding patterns in data.

What about the ethical and privacy implications?

Big Data has a great deal of ethical and privacy implications. Issues of consent and potential risk are critical considerations, especially when working with populations that are newly online and/or who may not have a good understanding of data privacy and how their data may be used by third parties who are collecting and/or selling it. However, one participant felt that a protectionist mentality is misguided. ‘We are pushing back and saying that social media and data tracking are bad. Instead, we should realize that having a digital life and being counted in the world is a right and it’s going to be inevitable in the future. We should be working with the people we serve to better understand digital privacy and help them to be more savvy digital citizens.’ It’s also imperative that aid and development agencies abandon our slow and antiquated data collection systems, she said, and to use the new digital tools that are available to us.

How can we be more responsible with the data we gather and use?

Development and humanitarian agencies do need be more responsible with data policies and practices, however. Big data approaches may contribute to negative data extraction tendencies if we mine data and deliver it to decision-makers far away from the source. It will be critical for evaluators and big data practitioners to find ways to engage people ‘on the ground’ and involve more communities in interpreting and querying their own big data. (For more on responsible data use, see the Responsible Development Data Book. Oxfam also has a responsible data policy that could serve as a reference. The author of this blog is working on a policy and practice guide for protecting girls digital safety, security and privacy as well.)

Who should be paying for big data sets to be made available?

One participant asked about costs and who should bear the expense of creating big data sets and/or opening them up to evaluators and/or data scientists. Others asked for examples of the private sector providing data to the social sector. This highlighted additional ethical and privacy issues. One participant gave an example from the healthcare space where there is lots of experience in accessing big data sets generated by government and the private sector. In this case, public and private data sets needed to be combined. There were strict requirements around anonymization and the effort ended up being very expensive, which made it difficult to build a business case for the work.

This can be a problem for the development sector, because it is difficult to generate resources for resolving social problems; there is normally only investment if there is some kind of commercial gain to be had. Some organizations are now hiring ‘data philanthropist’ positions that help to negotiate these kinds of data relationships with the private sector. (Global Pulse has developed a set of big data privacy principles to guide these cases.)

So, is big data going to replace evaluation or not?

In conclusion, big data will not eliminate the need for evaluation. Rather, it’s likely that it will be integrated as another source of information for strengthening conventional evaluation design. ‘Big Data and the underlying methods of data science are opening up new opportunities to answer old questions in new ways, and ask new kinds of questions. But that doesn’t mean that we should turn to big data and its methods for everything,’ said one discussant. ‘We need to get past a blind faith in big data and get more practical about what it is, how to use it, and where it adds value to evaluation processes,’ said another.

Thanks again to all who participated in the discussion! If you’d like to join (or read about) conversations like this one, visit Technology Salon. Salons run under Chatham House Rule, so no attribution has been made in this summary post.

Read Full Post »

Screen Shot 2015-09-02 at 7.38.45 PMBack in 2010, I wrote a post called “Where’s the ICT4D distance learning?” which lead to some interesting discussions, including with the folks over at TechChange, who were just getting started out. We ended up co-hosting a Twitter chat (summarized here) and having some great discussions on the lack of opportunities for humanitarian and development practitioners to professionalize their understanding of ICTs in their work.

It’s pretty cool today, then, to see that in addition to having run a bunch of on-line short courses focused on technology and various aspects of development and social change work, TechChange is kicking off their first Diploma program focusing on using ICT for monitoring and evaluation — an area that has become increasingly critical over the past few years.

I’ve participated in a couple of these short courses, and what I like about them is that they are not boring one-way lectures. Though you are studying at a distance, you don’t feel like you’re alone. There are variations on the type and length of the educational materials including short and long readings, videos, live chats and discussions with fellow students and experts, and smaller working groups. The team and platform do a good job of providing varied pedagogical approaches for different learning styles.

The new Diploma in ICT and M&E program has tracks for working professionals (launching in September of 2015) and prospective Graduate Students (launching in January 2016). Both offer a combination of in-person workshops, weekly office hours, a library of interactive on-demand courses, access to an annual conference, and more. (Disclaimer – you might see some of my blog posts and publications there).

The graduate student track will also have a capstone project, portfolio development support, one-on-one mentorship, live simulations, and a job placement component. Both courses take 16 weeks of study, but these can be spread out over a whole year to provide maximum flexibility.

For many of us working in the humanitarian and development sectors, work schedules and frequent travel make it difficult to access formal higher-level schooling. Not to mention, few universities offer courses related to ICTs and development. The idea of incurring a huge debt is also off-putting for a lot of folks (including me!). I’m really happy to see good quality, flexible options for on-line learning that can improve how we do our work and that also provides the additional motivation of a diploma certificate.

You can find out more about the Diploma program on the TechChange website  (note: registration for the fall course ends September 11th).




Read Full Post »

The private sector has been using dashboards for quite some time, but international development organizations face challenges when it comes to identifying the right data dashboards and accompanying systems for decision-making.

Our May 29th, 2015, Technology Salon (sponsored by The Rockefeller Foundation) explored data dashboards and data visualization for improved decision making with lead discussants John DeRiggi, Senior Data Architect, DAI; Shawna Hoffman, Associate Manager, Evaluation and Learning at The MasterCard Foundation; Stephanie Evergreen, Evergreen Data.

In short, we learned at the Salon that most organizations are struggling with the data dashboard process. There are a number of reasons that dashboards fail. They may never get off the ground, they may not deliver what was promised, they may deliver but no one uses them, or they may deliver but the data is poor and bad decisions are made. Using data for better decision-making is an ongoing process – not a task or product to complete and then relegate to automation. Just getting a dashboard up and running doesn’t guarantee that it’s a success – it’s critical to look deeper to see if the data and its visualization have actually improved decisions and how. Like with any ICT tool, user centered design and ongoing iteration are key. Successful dashboards are organized, useful, include targets, and have trends and predictions. Organizational culture and change management are critical in the process.

Points discussed in detail*:

1) Ask whether you actually need a dashboard

The first question to ask is whether a dashboard is needed or possible. One discussant, who specializes in data visualization, noted that she’s often brought in because someone wants to do data visualization, and she then needs to work backwards with the organization through a number of other preparatory steps before getting to the part on data visualization. It’s critical to have data dashboard discussions with different parts of the organization in order to understand real needs and expectations. Often people will say they need a dashboard because they want to make better decisions, noted another lead discussant. “But what kind of decisions, and what information is needed to make those decisions? Where does that information come from? Who will get it?”

2) Define the audience and type of dashboard

People often think that they can create one dashboard that will fulfill everyone’s needs. As one discussant put it, they will say the audience for the dashboard is “everyone – all decision makers at all levels!” In reality most organizations will need several dashboards for different levels of decision-making. It’s important to know who will own it, use it, keep it up, and collect the data. Will it be internal or externally facing? Discussing all of this is a key part of the process of thinking through the dashboard. As one discussant outlined, dashboards can be strategic, analytical or operational. But it’s difficult for them to be all three at once. So organizations need to come to a clear understanding of their data and decision-making needs. What information, if available, would help different teams at different levels with their decision making? One dashboard can’t be everything to everyone. Creating a charter that outlines what the dashboard project is and what it aims to do is a way to help avoid mission creep, said one discussant.

3) Work with users to develop your dashboard

To start off the process, it’s important to clearly identify the audience and find out what they need – don’t assume you know, recommended one discussant. But also, as a Salon participant pointed out, don’t assume that they know either. Have a conversation where their and your expertise comes together. “The higher up you go, the less people may understand about data. One idea is to just take the ‘data’ out of the conversation. Ask decision-makers what questions they are trying to answer, what problems they are trying to solve. Then find out how to collect and visualize the data that helps them answer their questions,” suggested another participant. Create ownership and accountability at all levels – with users, with staff who will input the data, with project managers, with grantees – you need cooperation from all levels noted others. Clear buy-in will also help with data quality. If people see the results of their data coming out in a data visualization, they may be more inclined to provide quality data. One way to involve users is to gather different teams to talk about their data and to create ‘entity relationship models’ together. “People can get into the weeds, and then you can build a vocabulary for the organization. Then you can use that model to build the system and create commonality across it,” said one discussant. Another idea is to create paper prototypes of dashboards with users so that they can envision them better.

4) Dashboards help people engage with the data they’ve collected

A dashboard is a window into your data, said one participant. In some cases, seeing their data visualized can help staff to see that they have been providing poor quality data. “People didn’t realize how bad their data was until they saw their dashboard,” said one discussant. Another noted that people may disagree with what the data tells them in the dashboard and feel motivated to provide better data. On the other hand, they may realize that their data was actually good, and instead they need to improve ineffective programs. A danger is that putting a dashboard on top of bad data shines a light on the data, said one participant, and this might create an incentive for people to manipulate their data.

5) Don’t be over-ambitious

Align the dashboard with indicators that link to strategic goals and directions and stay focused, recommended one discussant. There is often a temptation to over-complicate with tons of data and visuals. But extraneous data leads to misinterpretation or distraction. Dashboards should make complex data available in an accessible way to users, she said. You can always make more visuals if needed, but you want a concise story told in the data and visuals that you’re depicting. Determine what is useful, productive and credible and leave out what is exciting but extraneous. “Don’t try to have 30 indicators.”

6) Be clear about your data categories and indicators

Rolling up data from a large number of different programs into a dashboard is a huge challenge, especially if different sites or programs are using different data models. For example, if one program is describing an activity as a ‘workshop’ and the other uses ‘training session,’ said one discussant, you have a problem. A Salon participant explained that her organization started with shallow but important common denominators across programs. Over time they aim to go deeper to begin looking at outcomes and impact.

7) Think through how you’ll sustain the dashboard and related system(s)

One discussant said that her organization established three different teams to work on the dashboard process: a) Metrics – Where do we have credible representative data? Where do we have indicators but we don’t have data? b) Plumbing: Where are the data sources? How do they feed into each other? Who is responsible, and can this be aggregated up? And c) Visualization: What visual would help different decision makers make their decisions? Depending on where the organization is in its stage of readiness and its existing staff capacities, different combinations of skill sets may be required to supplement existing ones. Data experts can help teams understand what is possible, yet program or management teams and other dashboard users also need to be involved so that they can identify the questions they are trying to answer with the data and the dashboard.

8) Don’t underestimate the time/resources needed for a functional dashboard

People may not realize that you can’t make a dashboard without data to support it, noted one participant. “It’s like a power point presentation… a power point doesn’t just appear out of nowhere. It’s a result of conversations, research, data, design and more. But for some reason, people think a dashboard will just magically create itself out of thin air.” People also seem to think you can create and launch a dashboard and then put it on autopilot, but that is not the case. The dashboard will need constant changes and iteration, and there will be continual work to keep it up. The questions being asked will also likely change over time and so the dashboard may need to shift to take this into consideration. Time will be required to get buy-in for the dashboard and its use. One Salon participant said that in her former organization, they met quarterly to present, use and discuss the dashboard, and it took about 2 years in order for it to become useful and for people to become invested in it. It’s very important, said one participant, to ensure that management knows that the dashboard is not a static thing – it will need ongoing attention and management.

9) Be selective when it comes to the technology

People tend to think that dashboards are just visual, said a Salon participant. They think they are really cool, business solution platforms. Often senior leadership has seen been pitched something really expensive and complicated, with all kinds of bells and whistles, and they may think that is what they need. It’s important to know where your organization is in terms of capacity before determining which technology would be the best fit, however, noted one discussant. She counseled organizations to use whatever they have on hand rather than bringing in new software that takes people 6 months to learn how to use. Simple excel-based dashboards might be the best place to start, she said.

10) Legacy systems can be combined with new data viz capabilities

One discussant shared how his company’s information system, which was set up over 15 years ago, did not allow for the creation of APIs. This meant that the team could not build derivative software products from their massive existing database. It is too expensive to replace the entire system, and building modules to replace some of it would lead to fragmenting the user experience. So the team built a thin web service layer on top of the existing system. This exposed the data to friendly web formats from which developers could build interactive products.

11) Be realistic about “real time” and “data quality”

One question that came up was around the the level of evidence needed to make good decisions. Having perfect data served up into a perfect visualization is utopian, said one Salon participant. The idea is that we could have ‘real time’ data to inform our decisions, she explained, yet it’s hard to quality check data so quickly. “So at what level can we say we’ll make decisions based on a level of certainty – is it when we feel 80% of the data is good quality? Do we need to lower that to 60% so that we have timely data? Is that too low?” Another question was around the kinds of decisions that require ‘real time’ data versus those that could be made based on data that is 3 to 6 months old. Salon participants said this will depend on the kind of program and the type of decision. The sector in which one is working may also determine the level of comfort with real time and with data quality – for example, the humanitarian sector may need more timely data and accept a lower level of verification whereas the development sector may be the opposite.

Another point was that dashboards should include error bars and available metadata, as well as in some cases a link to raw data for those who want to dig into the data and understand what is behind the dashboard. Sometimes the dashboard process will highlight that there is simply not much quality data available for some programs in some countries. This can be an opportunity to work with staff on the ground to strengthen capacity to collect it.

12) Relax

As one discussant said, “much of the concern about data quality is related to our own hang-ups as data nerds and what we feel comfortable putting out there for people to use to make decisions. We always say ‘we need more research.’” But here the context is different. “Stakeholders and management want the answer. We need to just put the data out there with some caveats to help them.” One way to offer more context for a dashboard is creating a dashboard report that provides some narrative alongside the visualization. Dashboards should also show trends, not only what has happened already, she said. People need to see trends towards the future so that decisions can be made. It was also pointed out that a dashboard shouldn’t be the only basis for decisions. Like a car dashboard – these data dashboards signal that something is changing but you still need to look under the hood to see what it is. The dashboard should trigger questions – it should be a launch pad for discussion.

13) Organizational culture is a huge part of this process

The internal culture and people’s attitudes towards data are embedded into how an organization operates, noted one Salon participant. This varies depending on the type of organization – an evaluation focused organization vs. a development organization vs. a contractor vs. a humanitarian organization, for example. Outside consultants can help you to build a dashboard, but it will be critical to have someone managing organizational change on the inside who knows the current culture and where the organization is aiming to go with the dashboard process. The process is getting easier, however. Many organizations are thirsty for data now, noted one lead discussant. “Often the research or evaluation team create a dashboard and send it to the management team, and then everyone loves it and wants one. People are ready for it now.”

More resources on data dashboards and visualization.

Special thanks to our lead discussants and to our hosts for this Salon! If you’d like to join our Salon discussions in the future, sign up at the Technology Salon site.

*Salons run under Chatham House Rule, so no attribution has been made in this post.

Read Full Post »

Today as we jump into the M&E Tech conference in DC (we’ll also have a Deep Dive on the same topic in NYC next week), I’m excited to share a report I’ve been working on for the past year or so with Michael Bamberger: Emerging Opportunities in a Tech-Enabled World.

The past few years have seen dramatic advances in the use of hand-held devices (phones and tablets) for program monitoring and for survey data collection. Progress has been slower with respect to the application of ICT-enabled devices for program evaluation, but this is clearly the next frontier.

In the paper, we review how ICT-enabled technologies are already being applied in program monitoring and in survey research. We also review areas where ICTs are starting to be applied in program evaluation and identify new areas in which new technologies can potentially be applied. The technologies discussed include hand-held devices for quantitative and qualitative data collection and analysis, data quality control, GPS and mapping devices, environmental monitoring, satellite imaging and big data.

While the technological advances and the rapidly falling costs of data collection and analysis are opening up exciting new opportunities for monitoring and evaluation, the paper also cautions that more attention should be paid to basic quality control questions that evaluators normally ask about representativity of data and selection bias, data quality and construct validity. The ability to use techniques such as crowd sourcing to generate information and feedback from tens of thousands of respondents has so fascinated researchers that concerns about the representativity or quality of the responses have received less attention than is the case with conventional instruments for data collection and analysis.

Some of the challenges include the potential for: selectivity bias and sample design, M&E processes being driven by the requirements of the technology and over-reliance on simple quantitative data, as well as low institutional capacity to introduce ICT and resistance to change, and issues of privacy.

None of this is intended to discourage the introduction of these technologies, as the authors fully recognize their huge potential. One of the most exciting areas concerns the promotion of a more equitable society through simple and cost-effective monitoring and evaluation systems that give voice to previously excluded sectors of the target populations; and that offer opportunities for promoting gender equality in access to information. The application of these technologies however needs to be on a sound methodological footing.

The last section of the paper offers some tips and ideas on how to integrate ICTs into M&E practice and potential pitfalls to avoid. Many of these were drawn from Salons and discussions with practitioners, given that there is little solid documentation or evidence related to the use of ICTs for M&E.

Download the full paper here! 

Read Full Post »

Earlier this month I attended the African Evaluators’ Conference (AfrEA) in Cameroon as part of the Technology and Evaluation stream organized by Pact with financial support from The Rockefeller Foundation’s Evaluation Office and The MasterCard Foundation.

A first post about ICTs and M&E at the Afrea Conference went into some of the deliberations around using or not using ICTs and how we can learn and share more as institutions and evaluators. I’ve written previously about barriers and challenges with using ICTs in M&E of international development programs (see the list of posts at the bottom of this one). Many of these same conversations came up at AfrEA, so I won’t write about these again here. What I did want to capture and share were a few interesting design and implementation thoughts from the various ICT and M&E sessions. Here goes:

1) Asking questions via ICT may lead to more honest answers. Some populations are still not familiar with smart phones and tablets and this makes some people shy and quiet, yet it makes others more curious and animated to participate. Some people worry that mobiles, laptops and tablet create distance between the enumerator and the person participating in a survey. On the other hand, I’m hearing more and more examples of cases where using ICTs for surveying actually allow for a greater sense of personal privacy and more honest answers. I first heard about this several years ago with relation to children and youth in the US and Canada seeking psychological or reproductive health counseling. They seemed to feel more comfortable asking questions about sensitive issues via online chats (as opposed to asking a counselor or doctor face-to-face) because they felt anonymous. This same is true for telephone inquiries.

In the case of evaluations, someone suggested that rather than a mobile or tablet creating distance, a device can actually create an opportunity for privacy. For example, if a sensitive question comes up in a survey, an enumerator can hand the person being interviewed the mobile phone and look away when they provide their answer and hit enter, in the same way that waiters in some countries will swipe your ATM card and politely look away while you enter your PIN. Key is building people’s trust in these methods so they can be sure they are secure.

At a Salon on Feb 28, I heard about mobile polling being used to ask men in the Democratic Republic of Congo about sexual assault against men. There was a higher recorded affirmative rate when the question was answered via a mobile survey than when the question had been asked in other settings or though other means. This of course makes sense, considering that often when a reporter or surveyor comes around asking whether men have been victims of rape, no one wants to say publicly. It’s impossible to know in a situation of violence if a perpetrator might be standing around in the crowd watching someone getting interviewed, and clearly shame and stigma also prevent people from answering openly.

Another example at the AfrEA Tech Salon, was a comparison study done by an organization in a slum area in Accra. Five enumerators who spoke local languages conducted Water, Sanitation and Hygiene (WASH) surveys by mobile phone using Open Data Kit (an open source survey application) and the responses were compared with the same survey done on paper.  When people were asked in person by enumerators if they defecated outdoors, affirmative answers were very low. When people were asked the same question via a voice-based mobile phone survey, 26% of respondents reported open defecation.

2) Risk of collecting GPS coordinates. We had a short discussion on the plusses and minuses of using GPS and collecting geolocation data in monitoring and evaluation. One issue that came up was safety for enumerators who carry GPS devices. Some people highlighted that GPS devices can put staff/enumerators at risk of abuse from organized crime bands, military groups, or government authorities, especially in areas with high levels of conflict and violence. This makes me think that if geographic information is needed in these cases, it might be good to use a mobile phone application that collects GPS rather than a fancy smart phone or an actual GPS unit (for example, one could try out PoiMapper, which works on feature phones).

In addition, evaluators emphasized that we need to think through whether GPS data is really necessary at household level. It is tempting to always collect all the information that we possibly can, but we can never truly assure anyone that their information will not be de-anonymized somehow in the near or distant future, and in extremely high risk areas, this can be a huge risk. Many organizations do not have high-level security for their data, so it may be better to collect community or district level data than household locations. Some evaluators said they use ‘tricks’ to anonymize the geographical data, like pinpointing location a few miles off, but others felt this was not nearly enough to guarantee anonymity.

3) Devices can create unforeseen operational challenges at the micro-level. When doing a mobile survey by phone and asking people to press a number to select a particular answer to a question, one organization working in rural Ghana to collect feedback about government performance found that some phones were set to lock when a call was answered. People were pressing buttons to respond to phone surveys (press 1 for….), but their answers did not register because phones were locked, or answers registered incorrectly because the person was entering their PIN to unlock the phone. Others noted that when planning for training of enumerators or community members who will use their own devices for data collection, we cannot forget the fact that every model of phone is slightly different. This adds quite a lot of time to the training as each different model of phone needs to be explained to trainees. (There are a huge number of other challenges related to devices, but these were two that I had not thought of before.)

4) Motivation in the case of poor capacity to respond. An organization interested in tracking violence in a highly volatile area wanted to take reports of violence, but did not have a way to ensure that there would be a response from an INGO, humanitarian organization or government authority if/when violence was reported. This is a known issue — the difficulties of encouraging reporting if responsiveness is low. To keep people engaged this organization thanks people immediately for reporting and then sends peace messages and encouragement 2-3 times per week. Participants in the program have appreciated these ongoing messages and participation has continued to be steady, regardless of the fact that immediate help has not been provided as a result of reporting.

5) Mirroring physical processes with tech. One way to help digital tools gain more acceptance and to make them more user-friendly is to design them to mirror paper processes or other physical processes that people are already familiar with. For example, one organization shared their design process for a mobile application for village savings and loan (VSL) groups. Because security is a big concern among VSL members, the groups typically keep cash in a box with 3 padlocks. Three elected members must be present and agree to open and remove money from the box in order to conduct any transaction. To mimic this, the VSL mobile application requires 3 PINs to access mobile money or make transactions, and what’s more, the app sends everyone in the VSL Group an SMS notification if the 3 people with the PINs carry out a transaction, meaning the mobile app is even more secure than the original physical lock-box, because everyone knows what is happening all the time with the money.


As I mentioned in part 1 of this post, some new resources and forthcoming documentation may help to further set the stage for better learning and application of ICTs in the M&E process. Pact has just released their Mobile Technology Toolkit, and Michael Bamberger and I are finishing up a paper on ICT-enabled M&E that might help provide a starting point and possible framework to move things forward.

Here is the list of toolkits, blog posts and other links that we compiled for AfrEA – please add any that are missing!

Previous posts on ICTs and M&E on this blog:

Read Full Post »

I attended the African Evaluators’ Conference (AfrEA) in Cameroon last week as part of the Technology and Evaluation strand organized by Pact with financial support from The Rockefeller Foundation’s Evaluation Office and The MasterCard Foundation. The strand was a fantastic opportunity for learning, sharing and understanding more about the context, possibilities and realities of using ICTs in monitoring and evaluation (M&E). We heard from a variety of evaluators, development practitioners, researchers, tool-developers, donors, and private sector and government folks. Judging by the well-attended sessions, there is a huge amount of interest in ICTs and M&E.

Rather than repeat what’s I’ve written in other posts (see links at the bottom), I’ll focus here on some of the more relevant, interesting, and/or new information from the AfrEA discussions. This first post will go into institutional issues and the ‘field’ of ICTs and M&E. A second post will talk about design and operational tips I learned /was reminded of at AfrEA.

1) We tend to get stuck on data collection –Like other areas (I’m looking at you, Open Data) conversations tend to revolve around collecting data. We need to get beyond that and think more about why we are collecting data and what we are going to do with it (and do we really need all this data?). The evaluation field also needs to explore all the other ways it could be using ICTs for M&E, going beyond mobile phones and surveys. Collecting data is clearly a necessary part of M&E, but those data still need to be analyzed. As a participant from a data visualization firm said, there are so many ways you can use ICTs – they help you make sense of things, you can tag sentiment, you can visualize data and make data-based decisions. Others mentioned that ICTs can help us to share data with various stakeholders, improve sampling in RCTs (Randomized Control Trials), conduct quality checks on massive data sets, and manage staff who are working on data collection. Using big data, we can do analyses we never could have imagined before. We can open and share our data, and stop collecting the same data from the same people multiple times. We can use ICTs to share back what we’ve learned with evaluation stakeholders, governments, the public, and donors. The range of uses of ICTs is huge, yet the discussion tends to get stuck on mobile surveys and data collection, and we need to start thinking beyond that.

2) ICTs are changing how programs are implemented and how M&E is done — When a program already uses ICTs, data collection can be built in through the digital device itself (e.g., tracking user behavior, cookies, and via tests and quizzes), as one evaluator working on tech and education programs noted. As more programs integrate digital tools, it may become easier to collect monitoring and evaluation data with less effort. Along those lines, an evaluator looking at a large-scale mobile-based agricultural information system asked about approaches to conducting M&E that do not rely on enumerators and traditional M&E approaches. In his program, because the farmers who signed up for the mobile information service do not live in the same geographical community, traditional M&E approaches do not seem plausible and ICT-based approaches look like a logical answer. There is little documentation within the international development evaluation community, however, on how an evaluator might design an evaluation in this type of a situation. (I am guessing there may be some insights from market research and possibly from the transparency and accountability sectors, and among people working on “feedback loops”).

3) Moving beyond one-off efforts — Some people noted that mobile data gathering is still done mostly at the project level. Efforts tend to be short-term and one-off. The data collected is not well-integrated into management information systems or national level processes. (Here we may reference the infamous map of mHealth pilots in Uganda, and note the possibility of ICT-enabled M&E in other sectors going this same route). Numerous small pilots may be problematic if the goal is to institutionalize mobile data gathering into M&E at the wider level and do a better job of supporting and strengthening large-scale systems.

4) Sometimes ICTs are not the answer, even if you want them to be – One presenter (who considered himself a tech enthusiast) went into careful detail about his organization’s process of deciding not to use tablets for a complex evaluation across 4 countries with multiple indicators. In the end, the evaluation itself was too complex, and the team was not able to find the right tool for the job. The organization looked at simple, mid-range and highly complex applications and tools and after testing them all, opted out. Each possible tool presented a set of challenges that meant the tool was not a vast improvement over paper-based data collection, and the up-front costs and training were too expensive and lengthy to make the switch to digital tools worthwhile. In addition, the team felt that face-to-face dynamics in the community and having access to notes and written observations in the margins of a paper survey would enable them to conduct a better evaluation. Some tablets are beginning to enable more interactivity and better design for surveys, but not yet in a way that made them a viable option for this evaluation. I liked how the organization went through a very thorough and in-depth process to make this decision.

Other colleagues also commented that the tech tools are still not quite ‘there’ yet for M&E. Even top of the line business solutions are generally found to be somewhat clunky. Million dollar models are not relevant for environments that development evaluators are working in; in addition to their high cost, they often have too many features or require too much training. There are some excellent mid-range tools that are designed for the environment, but many lack vital features such as availability in multiple languages. Simple tools that are more easily accessible and understandable without a lot of training are not sophisticated enough to conduct a large-scale data collection exercise. One person I talked with suggested that the private sector will eventually develop appropriate tools, and the not-for-profit sector will then adopt them. She felt that those of us who are interested in ICTs in M&E are slightly ahead of the curve and need to wait a few years until the tools are more widespread and common. Many people attending the Tech and M&E sessions at AfrEA made the point that use of ICTs in M&E would get easier and cheaper as the field develops, tools get more advanced/appropriate/user-friendly and widely tested, and networks/ platforms/ infrastructure improves in less-connected rural areas.

5) Need for documentation, evaluation and training on use of ICTs in M&E – Some evaluators felt that ICTs are only suitable for routine data collection as part of an ongoing program, but not good for large-scale evaluations. Others pointed out that the notions of ‘ICT for M&E’ and ‘mobile data collection/mobile surveys’ are often used interchangeably, and evaluation practitioners need to look at the multiple ways that ICTs can be used in the wider field of M&E. ICTs are not just useful for moving from paper surveys to mobile data gathering. An evaluator working on a number of RCTs mentioned that his group relies on ICTs for improving samples, reducing bias, and automatically checking data quality.

There was general agreement that M&E practitioners need resources, opportunities for more discussion, and capacity strengthening on the multiple ways that ICTs may be able to support M&E. One evaluator noted that civil society organizations have a tendency to rush into things, hit a brick wall, and then cross their arms and say, “well, this doesn’t work” (in this case, ICTs for M&E). With training and capacity, and as more experience and documentation is gained, he considered that ICTs could have a huge role in making M&E more efficient and effective.

One evaluator, however, questioned whether having better, cheaper, higher quality data is actually leading to better decisions and outcomes. Another evaluator asked for more evidence of what works, when, with whom and under what circumstances so that evaluators could make better decisions around use of ICTs in M&E. Some felt that a decision tree or list of considerations or key questions to think through when integrating ICTs into M&E would be helpful for practitioners. In general, it was agreed that ICTs can help overcome some of our old challenges, but that they inevitably bring new challenges. Rather than shy away from using ICTs, we should try to understand these new challenges and find ways to overcome/work around them. Though the mHealth field has done quite a bit of useful research, and documentation on digital data collection is growing, use of ICTs is still relatively unexplored in the wider evaluation space.

6) There is no simple answer. One of my takeaways from all the sessions was that many M&E specialists are carefully considering options, and thinking quite a lot about which ICTs for what, whom, when and where rather than deciding from the start that ICTs are ‘good and beneficial’ or ‘bad and not worth considering.’ This is really encouraging, and to be expected of a thoughtful group like this. I hope to participate in more discussions of this nature that dig into the nuances of introducing ICTs into M&E.

Some new resources and forthcoming documentation may help to further set the stage for better learning and application of ICTs in the M&E process. Pact has just released their Mobile Technology Toolkit, and Michael Bamberger and I are finishing up a paper on ICT-enabled M&E that might help provide a starting point and possible framework to move things forward. The “field” of ICTs in M&E is quite broad, however, and there are many ways to slice the cake. Here is the list of toolkits, blog posts and other links that we compiled for AfrEA – please add any that you think are missing!

(Part 2 of this post)

Previous posts on ICTs and M&E:

Read Full Post »

This is a cross-post from Tom Murphyeditor of the aid blog A View From the Cave. The original article can be found on Humanosphere. The post summarizes discussions at our November 21st New York City Technology Salon: Are Mobile Money Cash Grants the Future of Development?  If you’d like to join us for future Salons, sign up here.

by Tom Murphy

Decades ago, some of the biggest NGOs simply gave away money to individuals in communities. People lined up and were just given cash.

The once popular form of aid went out of fashion, but it is now making a comeback.

Over time, coordination became extremely difficult. Traveling from home to home costs time and money for the NGO and the same problem exists for recipients when they have to go to a central location. More significant was the shift in development thinking that said giving hand outs was causing long term damage.

The backlash against ‘welfare queens’ in the US, UK and elsewhere during the 1980s was reflected in international development programming. Problem was that it was all based on unproven theories of change and anecdotal evidence, rather than hard evidence.

Half a decade later, new research shows that just giving people money can be an effective way to build assets and even incomes. The findings were covered by major players like NPR and the Economist.

While exciting and promising, cash transfers are not a new tool in the development utility belt.

Various forms of transfers have emerged over the past decade. Food vouchers were used by the World Food Programme when responding to the 2011 famine in the Horn of Africa. Like food stamps in the US, people could go buy food from local markets and get exactly what they need while supporting the local economy.

The differences have sparked a sometimes heated debate within the development community as to what the findings about cash transfers mean going forward. A Technology Salon hosted conversation at ThoughtWorks in New York City last week, featured some of the leading researchers and players in the cash transfer sector.

The salon style conversation featured Columbia University and popular aid blogger Chris Blattman, GiveDirectly co-founder and UCSD researcher Paul Neihaus and Plan USA CEO Tessie San Martin. The ensuing discussion, operating under the Chatham House Rule of no attribution, featured representatives from large NGOs, microfinance organizations and UN agencies.

Research from Kenya, Uganda and Liberia show both the promise and shortcomings of cash transfers. For example, giving out cash in addition to training was successful in generating employment in Northern Uganda. Another program, with the backing of the Ugandan government, saw success with the cash alone.

Cash transfers have been argued as the new benchmark for development and aid programs. Advocates in the discussion made the case that programs should be evaluated in terms of impact and cost-effectiveness against just giving people cash.

That idea saw some resistance. The research from Liberia, for example, showed that money given to street youth would not be wasted, but it was not sufficient to generate long-lasting employment or income. There are capacity problems and much larger issues that probably cannot be addressed by cash alone.

An additional concern is the unintended negative consequences caused by cash transfers. One example given was that of refugees in Syria. Money was distributed to families labeled for rent. Despite warnings not to label the transfer, the program went ahead.

As a result, rents increased. The money intended to help reduce the cost incurred by rent was rendered largely useless. One participant raised the concern that cash transfers in such a setting could be ‘taxed’ by rebels or government fighters. There is a potential that aid organizations could help fund fighting by giving unrestricted cash.

The discussion made it clear that the applications of cash transfers are far more nuanced than they might appear. Kenya saw success in part because of the ease of sending money to people through mobile phones. Newer programs in India, for example, rely on what are essentially ATM cards.

Impacts, admitted practitioners, can go beyond simple incomes. There has been care to make sure that implementing cash transfer programs to not dramatically change social structures in ways that cause problems for the community and recipients. In one case, giving women cash allowed for them to participate in the local markets, a benefit to everyone except for the existing shop oligarchs.

Governments in low and middle-income countries are seeing increasing pressure to establish social programs. The success of cash transfer programs in Brazil and Mexico indicate that it can be an effective way to lift people out of poverty. Testing is underway to bring about more efficient and context appropriate cash transfer schemes.

An important component in the re-emergence of cash transfers is looking back to previous efforts, said one NGO official. The individual’s organization is systematically looking back at communities where the NGO used to work in order to see what happened ten years later. The idea is to learn what impacts may or may not have been on that community in order to inform future initiatives.

“Lots of people have concerns about cash, but we should have concerns about all the programs we are doing,” said a participant.

The lessons from the cash transfer research shows that there is increasing need for better evidence across development and aid programs. Researchers in the group argued that the ease of doing evaluations is improving.

Read the “Storified” version of the Technology Salon on Mobiles and Cash Transfers here.

Read Full Post »

Older Posts »


Get every new post delivered to your Inbox.

Join 894 other followers