We’ve said it before: When you need to know more than just a general average, you need an Equity Gap Score. An average tells you what happened and an Equity Gap Score tells you for who. Each piece of information needs the other to give meaning. If you have a mandate that includes equity between any category of people, whether race, sex, income, education, geography, or whether or not they like sugar in their tea, you need an Equity Gap Score.
An Equity Gap Score is simply a number that helps to contextualize a statistic. It will help add meaning to any data result and can help you track equity issues in any data project. At We All Count, we think they should be included as standard practice any time you see an average involving people.
They look like this:
“The city has an average yearly income of $54,000, with an equity gap score of 3.7 between the women and men”
Or
“The rate of death by drowning has fallen by 17% in the last three years, while the equity gap score has worsened from 1.4 to 1.8 between poor neighbourhoods and rich neighbourhoods”.
The first number tells you the overall information, and the second number tells you how the equity between categories of people is doing, comparing the lowest to the highest numbers that are in that average. An equity gap ratio of 1 is a perfect equity score, indicating that all group numbers were equal to each other (like a set of 3 and 3 and 3 has an average of 3) getting worse the higher the score is (indicating a big difference between highest and lowest groups like a set of 1 and 1 and 7, which also has an average of 3).
They couldn’t be easier to calculate:
Best in category / Worst in category = Equity Gap Score.
CASE STUDY:
We were so thrilled when our friend and colleague Corey Newhouse, Founder at Public Profit, told us that she and her team were using these Equity Gap Scores (EGSs) in their work with great success.
We talked to them about the details so we could learn what was working best and how the team was able to introduce the idea to their clients.
The Public Profit team used the Equity Gap Scores recently in two large projects – one on the topic of the relationships between health and housing and one on the best practices for organizational culture change around diversity, equity inclusion.
Both projects involved surveying respondents but each survey was unique. The Public Profit team built Equity Gap Scores for survey items, individual or combined, where it made sense and then introduced the idea to their clients by sharing some of the results. In the health and housing survey, they created Equity Gap Scores for things like self-reported level of current health and level of access to public transportation. In the second survey, they applied the tool to items such as level of access to professional training activities and sense of belonging in the workplace.
By building EGSs for these individual items as well as creating equity gap scores for combinations of items they call domains, they were able to provide the report audience with a meaningful metric of equity that they could use to think about trends across topics and domains. This empowers the audience to get a solid sense of what parts of their work are potentially in need of the most immediate and urgent strategic adjustments in terms of equity. By including these scores in the reports is a visual way – which is largely how Public Profit did this – they allow readers to immediately understand broad trends in equity happening in their organization as well as specific groups or clusters of people who might be experiencing the most disparate impacts consistently. Once these trends are understood, of course, it’s always possible to dig down into the details of the data underlying the gaps.
Since Equity Gap Scores were new to both their clients, the Public Profit team took several steps to orient them to the idea and how to engage with it. First, they mentioned in the early stages of analysis that they would be using EGSs as part of the reporting. Second, before any of the main meaning-making sessions with clients they had a pre-meeting designed to provide an introduction to the idea and get everyone on the same page. At this pre-meeting, the Public Profit team explained what Equity Gap Scores are, how they could be helpful, how to read them, and how to make sense of them.
By the time the group meaning-making sessions (what WAC would call the Interpretation step of the Data Equity Framework) came around and the report containing the Equity Gap Scores was presented, they could easily make sense of the scores and use them to start all kinds of equity focussed conversations around their findings.
One of the reasons Corey and the Pubic Profit team think the EGSs worked so well with their clients is that they are a practical and understandable tool to address our growing awareness of equity issues in data. Corey says:
I think because it was such a natural fit for the analysis needs, that there was just a huge step forward in everyone’s ability to understand equity with the data…..One of the things I loved about it so much was, it allows us to just be mindful that we all have biases in our heads; I know I can look at one set of data be like, “Oh, that’s just five percentage points different…”, but then look at another, another set of data and say, “Woah! That’s five percentage points different!” And so just to have a really systematic, simple way to be like: bigger is worse. A bigger gap = not good. To be able to normalize that, and both of these were very hefty surveys, so to be able to do some of that pattern-finding at the domain level of like, “Whoa, there’s something happening over here, let’s look a little bit deeper…” I think that that was definitely a big lesson learned for us and our clients…..from, a pattern-finding perspective. And it worked amazingly well.
I need a little help here. Equity gap makes so much more sense than ‘achievement gap’, where the latter is putting assessment results all on the student. But the word ‘equity’ to me suggests that the gap is caused by ‘inequities’, which to me means the reason for the gap is outside the student. But the gap is from many sources — schools, parents, community, teachers, AND the student, etc.
Why am I hung up on the word ‘equity’ and what it communicates? Your definition is that ‘average’ says what happened, and ‘equity gap’ says to whom (not by whom). Is it this simple?
Hi Claire! Great comment. I think you make a good point about the nuance of the word “equity”. I think we may need to do some refining on the name. I think we need to spend time thinking about the distinction between differences, disproportionalities, disparities, and inequities.
I like the concept of ‘Equity Gap’ to quantify the degree of inequities over a spectrum of different background variables. However, it appears to me that the calculation is for data based on the whole population than a sample of individuals. I have tried to look for such summary measure in statistics text books I have been familiar with to understand how to construct and interpret confidence intervals when we have data only for a sample of individuals but I am not able to access one so far. I would appreciate what approach you recommend in such cases of using sample data and calculating estimation stat (confidence intervals) to infer to the general population.
Hi Wondu. Thank you for your note. We use the Equity Gap measure with data that is from either samples or entire populations. You are absolutely correct that for either one, if you want to understand the level of certainty around the gap, you’d need to calculate a confidence interval for the point estimate. We strongly encourage this!