What I’ve Learned about Data Advocacy

A very special thanks to Lucy Chambers from Open Knowledge Foundation for inspiration, review and advice on this blog post. You can find Lucy on twitter @lucyfedia.

Over the last two decades of helping advocates and activists use technologies more effectively, the use of data in campaigns has always been constant.  While the tools for gathering, storing and packaging data has changed significantly, there are quite a few things that haven’t.

Data is everywhere

The key to effective data advocacy is to understand the difference between data and evidence.  Data is discrete pieces of information, such as prices, measurements, dates, names of places and people, and addresses.  Evidence is when data is used to establish facts or expose truth.  For activists and advocates, the ability to take available relevant data and turn it into evidence can be the key to winning campaigns

For example: this Apple.Apple

Your first thought might be, ‘yummy apple.’ But if you are encountering the apple in a store, you’ll probably come across data about the apple.  Such as:

Type: Cox       Location Picked: Kent           Size: 50 grams      Date picked: 10/09/2013        Price: 50 pence

 You might also be interested in nutritional data, such as:


Now take a bushel of apples:


The data we first want to know might be similar to when we go to buy an apple:

Type: Braeburn         Location picked: Kent, United Kingdom     Date: 10/09/2013

However if I’m a wholesale buyer and want to know what sort of profit I can get, I’ll be interested in the following:

Price per kilo from farmer: 1 pound 25 pence     Price per kilo at market: 2 pounds

If I’m a farm labourer, there is different data about the bushel that I would want to know:

Price paid per bushel picked: 75 pence     Average time taken to pick a bushel: 8.5 minutes

Relevant data varies from the consumer, the merchant to the farm labourer.  The data needed for a campaign will change dramatically depending on the issue. Whether advocating for better eating habits, or consumer protection or better pay for farm labourers.

There are plenty of sources of data around, what you use and where you find it, will very much depend on what you are going to do with the data.

Data projects should start with a really good question

Find the question that asks ‘what is happening’ and is related to the core of your campaign.  For transparency and accountability groups this might be asking a question like: ‘does the government’s budget match it’s actual expenditure?’ or ‘are the promised services actually available to citizens?’  For human rights organisations working with marginalised communities the question might be, “has violence increased or decreased?’ or ‘who is committing violence?’  The question should resonate and engage those who are impacted.

The basic steps in using data in advocacy


(Courtesy of the Open Knowledge Foundation)

You need to know who your stakeholders are, how they will react to the question itself and what you want them to do in relation to the data.

  • Are you wanting to mobilise allies?  How will your data motivate them to action?  How can you engage them in either collecting or distributing data or both?

  • Are you wanting to educate neutral parties?  How can the data persuade them to become allies or reconsider your oponnents positions?

  • Are you countering opponents?  Is your data disproving their policies, statements or messaging?

Trying to answer “why?” is where data can be problematic

It’s quite easy for data to answer a question about what is happening, but it’s much harder to answer why.  Why do farm labourers get paid so little for their bushel of apples?  This can be the achilles heel of using data in advocacy.  We even see this in the public debate about climate change, where one data set indicates that climate is changing and temperatures are rising. A small group of ‘climate change deniers’ claim that correlating a separate data set about carbon emissions does not indicate a connection.  For more see the Data Driven Intelligence blog post on ‘If correlation doesn’t imply causation, then what does?‘.  Also some examples on ‘Correlation or Causation?‘ from Bloomberg Business Week.

I once had a sex worker advocate in India say to me “we’ve documented thousands of cases of violence against sex workers and we can prove it happens. But, we can’t prove why it happens. Why does a police officer beat up a sex worker?  What compels a doctor, whom the sex worker has come to for treatment, to beat her?” The data about violence against sex workers can prove it is happening and by whom.  The data that a large majority of violence happening against sex workers is at the hands of the police has so much power because it is disproving a widely held belief that clients are often the main perpetrators of violence against sex workers.  But once you are proving it happens, then what? Sometimes in order to address a difficult issue, you have to prove why and that is much harder.

Make sure you understand exactly what your data does prove.

Working with data is a journey

Be prepared to have your own assumptions tested. Rebecca Chiao of Harassmap recently told me that their journey in collecting data about sexual harassment on the streets of Cairo took a major turn when they learned that they discovered that the reports they were getting were not just about women experiencing sexual harassment but also men. “We were reminded that no one let young boys get into cabs by themselves.” Suddenly, what they assumed was a women’s issue became something bigger. “The data made us see our own blind spot. It was surprising to be confronted by our own stereotypes on this issue!” This meant that the picture of who their stakeholders were changed significantly and they had to change their outreach accordingly.

So you will often start with a question, but as you gather data, the answer to the question may take you on a different path and the question itself might change.

Great resources for using data:

More Resources from Fabriders:

What have you learned about using data in advocacy?  Or do you have any great resources about data advocacy to share?