Collaborating with Numbers
I don’t like numbers, but not because I loathe math. For me, numbers are those random roots that pop up in the middle of a hike and trip me up. It’s not the fault of the roots, they’re simply existing as they were designed. Nevertheless, I judge and blame them for being in my way. They force me to walk more cautiously, looking down at the ground instead of up and around to view the big picture of nature. I forget, however, that the roots are nature too. Without them working down in the ground, everything above could not exist. In “How to Make the World Add Up,” Tim Harford reminds me both are needed, and that “even to see that there is a problem requires a statistical perspective, even quite a simple one.” [1]
Still, I am naturally inclined to prefer qualitative rather than quantitative information. The quality of an experience gives me a deeper understanding of a problem. Though quantity is important, counting how many people are experiencing that problem is only impactful when the number is significant enough to notice. If the numbers are too low, no one will care enough to try and solve the problem. That’s another reason why I’m not a numbers person; they hold too much power over decisions that determine who gets what and how much. But again, it’s not the fault of the numbers – they aren’t making the decisions.
Although Harford’s rules convinced me to approach statistics with less trepidation, I still don’t like them. Perhaps it would be more accurate to say that I don’t trust them, or more precisely, I don’t trust the stories told around numbers. I discern whether or not to accept the stories based on the context of the storytellers: who they work for, what they stand for, and how their claims fit their agenda. I appreciate that Harford doesn’t shy away from such politics, noting how common it is to make assumptions “from the context in which we are usually presented with the number.” [2] Context makes it possible to spin any number “to support various political outlooks.” [3]
I had to wrestle with context to tell a story recently in my current role as an education specialist at the Nonprofit Association of Oregon. We are the umbrella organization that supports all nonprofit organizations in our state, providing education, resources and advocacy. Nonprofit leaders turn to us as the experts in all things nonprofit. As part of my role, I gather and teach the latest research and information related to diversity, equity, inclusion and belonging (DEIB). Last summer, I developed an online survey designed to measure the effectiveness of DEIB initiatives in nonprofit organizations in the state of Oregon. The survey was made available to all 1,558 members of our association. We received 109 responses which we deemed a sufficient number to start building a story around organizational efforts to create more diverse, equitable and inclusive workplace environments. We input the data into spreadsheets, created pie charts and graphs, made comparisons of responses between demographics, and took note of recognizable patterns. I also employed Harford’s first rule: search your feelings. This is instinctual for me. I’m a super feeler – if that’s a thing – and my gut was telling me there was something off with the numbers. Though the data were showing that equity initiatives are effective, my report concluded that those numbers were not telling the whole story. Though Harper might have agreed with my conclusions, I imagine he would have challenged my feelings: “expertise needs to be complemented by control of your own emotional reactions to the statistical claim you see.” [4]
These emotions were based on my personal experience which Harford does urge the reader to ponder. The data did not show my experience as a person of color working in predominantly white workplace environments, nor did it reflect the experience of my colleagues of color working in nonprofits in the Portland metro area. If leaders were to rely on the raw data without such reflections they would feel good about their present progress and proceed as if what they are doing is sufficient, however “future generations need to know whether we feel better because of the steps we’ve taken, or whether they were empty rituals that did no good, cost money, wasted time and produced unwelcome side effects.” [5]
Had I read Harford’s book prior to writing up my report, the process of learning from the data might have been smoother. First, I would have designed my questions more carefully, avoiding premature enumeration.[6] This explains why I kept coming back to certain words in two of the questions we asked. One question read: “Do you feel the working environment in your nonprofit is safe to be your authentic self?” How do we define authentic self? The other question read: “Do you feel your nonprofit celebrates diverse peoples, beliefs and ideas?” Answers would differ based on what the word “celebrate” meant or what “diversity” looked like to the respondent. So in these two cases, the answers weren’t the problem, it was the questions.
Second, I could have used rule number six to explain why the data wasn’t truly accurate: ask who is missing. For example, our data showed that 96% responded “yes” when asked if they would advise someone of similar culture, orientation, and/or gender identity to work in the present work environment at their organization. Ninety-six percent is great, until we put that number in context by looking at the demographics of the respondents: 83% white and 63% women. So while it’s true an overwhelming majority said yes, (which is probably why we have the same cultural make-up in nonprofits,) white women are advising other white women – so we don’t get much diversity from internal advisement. Only 10% of respondents identified as a person of color. How might the story have changed if we’d heard more of these voices?
The data also showed that 67% of respondents said that there is enough oversight to ensure policies surrounding equity are actionable. When we divide the numbers up by seniority, however, those numbers change. Those at entry or mid-level did not feel as confident in the oversight as senior leadership. Though context can be twisted, it is critical to understanding what is being measured. We have to step back and ask more questions. “Rather than focusing on the latest data point, get some context. What has happened to this data series over the past year? Past decade? What is happening elsewhere? Are there any comparisons which help make sense of the number?” [7]
If my agenda was to make DEIB efforts look good, I might have let the numbers speak for themselves. I could have ignored the voices not heard. I could have dismissed the difference between seniority. But statistics should “illuminate reality with clarity and honesty”.[8] One of the biggest problems in DEIB initiatives is too much effort is spent trying to “look” equitable rather than taking actual action. This creates an illusion of success. Illusions don’t solve problems, they exacerbate them, making the issue that much harder to address.
I’ve noticed that people decide to take action to fix a problem when they feel it at the heart level rather than just the mind. Qualitative data are the narratives that reach this level. They are easier to wrap our minds around. They satisfy the senses, help us see and hear what’s really happening. Problems come alive and we can more easily relate. Numbers, on the other hand, are cold and hard. When the complexity of problems in life are reduced to numbers, the lived experience is deadened. No matter how many fancy graphs are added to their presentation, numbers have the power to flatten or outright erase the problem. This is how I viewed statistics before Harford. From his book I learned a more collaborative approach. “We get information from graphs and spreadsheets. We also get information from the rich and vivid experiences all around us.” What he learned from Hans Rosling feels true: “the best insights come when we are able to combine the two.” [9]
[1] Tim Harford. How to make the World Add Up, (University of Oxford, Mathematical Institute, Oxford Mathematics Public Lectures, October 8, 2020) https://www.youtube.com/watch?v=0ZxV1WhqE90
[2] Tim Harford, How to Make the World Add Up (New York: Riverhead Books, 2020).
[3] Ibid.
[4] Ibid.
[5] Ibid.
[6] Ibid.
[7] Ibid.
[8] Ibid.
[9] Ibid.
About the Author
Erica Briggs
One response to “Collaborating with Numbers”
Leave a Reply
You must be logged in to post a comment.
Hi Erica, great post and I’m glad you are working in the DEIB sphere, especially considering how much it is under attack in various states like Oklahoma, where I live. Your role requires an extra dose of diplomacy and message framing. How do you navigate conveying results with context, which might not look as favorable as stand-alone statistics?