Digging into Approaches to Using Data
[Written on March 3, 2014] These days, in the world of international development and governance, “citizen engagement” is like the southwest print of fashion: it is “all the rage”.
Everyone who I have chatted with in the past few months, no matter if they work in philanthropies, multilaterals, academia or nonprofits, has invoked this term in some way or another. Some take citizen engagement to broadly mean “political participation” while others specifically point to “feedback”.
Any way that you slice it, the central concern is the same. In ’digital civil society’, as Stanford University’s Lucy Bernholz calls the universe of actors using technologies for social good, the key concern is with data usage. We have data, now what? How is data being used for reform? If it isn’t, how do we get there?
Today, governments are releasing datasets more than ever, nonprofits are generating data samples at rapid rates, and advocacy around open access to information is rising. One side of the equation seems to be well under way. The other — the analysis and strategic communication of this information to create change — remains underemphasized, under pursued, or at best poorly understood.
Given this scenario, it makes sense that “closing the citizen engagement loop” is where our attention lies today. However, my fear is that we will reinvent the wheel instead of iterating upon what we know works and what we know does not.
In an attempt to help push us in a new and improved direction, I will start by stressing one point: we need to scale approaches, not entire models.
(I am not a fan of the word ‘scale’ because it suggests that a development project can be successfully replicated when surely there are a million and one contextual reasons why it probably won’t work. In any case, there seems to be an obsession with the term, so perhaps it is best to redefine it to better associate it with replication that can be useful.)
Let’s take the example of a single data collection and usage model to better understand the approach versus model distinction. The State Government of Bihar, India requires village level development officers to send daily mobile reports according to specified metrics on 10 public services. This data is fed into a dashboard for state level government officers to analyze and act upon.
We would consider this a successful intervention model with evidence of state officers using the data to spot poor services in a specific locality, studying why the quality of service delivery is low, and taking action to mitigate the problem. Even so, it would be unwise to duplicate this model in a different Indian state like Madhya Pradesh (MP), where the same commitment from top leadership to enforce data collection may be lacking.
Instead, there is a higher likelihood of boosting outcomes in Madhya Pradesh by introducing for example, the training methods that Bihar state officials used to encourage local data collection to existing dysfunctional information gathering efforts in MP.
As digital civil society discusses and implements ways to address our shared concern of data usage for reform, I hope that past efforts are shared and learned from, and most importantly, examined for tried techniques that work.