Making data work accessible

April 2024 Issue

I sincerely believe that data matters. I believe that everybody everywhere could potentially live happier, more productive, and less stressful lives by learning to work with data – even small datasets and simple methods – to get insights into the things that are important to them. For example, I'm a university professor, and I was able to make better decisions about which classes to teach and how to teach them by creating a few bar charts. I was able to focus my non-university time more effectively after creating a simple spreadsheet of the return-on-investment, or ROI, of several projects in my life. And, recently, I was even able to use a basic analysis to show a local nonprofit here in Utah, Bicycle Collective, that for every 1,500 bicycles they make available for free or at reduced prices, they are able to provide health benefits equivalent to saving 1 life. (That conclusion was based on data from the World Health Organization's Health Economic Assessment Tool for Walking and Cycling, or HEAT).

I also know from my 30+ years of university teaching that many people face significant challenges in learning to work with data. Many different potential causes contribute to these challenges, such as:

• Lack of resources such as up-to-date textbooks, access to computers and internet, or safe and adequate classroom spaces
• Social and financial obligations, such as a need to care for people or work multiple jobs, that don't allow a person to spend as much time learning as they would like
• Limited or poorly-taught classes, which can fail to meet a person's interests or, even worse, derail a person's motivation to learn a topic

This list is, of course, not comprehensive. There are so many other challenges that people can face in their learning. I do know that all of the items I listed have affected my own students, both in their K-12 schools and in universities, and I have attempted to help my students with each of these situations. But, beyond those important issues, there is another set of challenges that I have become more acutely aware of and that I am specifically trying to address as I redesign my class, write a new textbook, and create new materials: learning differences that affect a person's ability to work productively with numbers and data.

In this newsletter, I'll explain what some of these differences are, how common they are, and a few suggestions for responding productively to these differences. But first I should explain my use of the term "differences."

Differences vs. disabilities

I recently started using the term "learning differences," as opposed to, say, "learning challenges" or "learning disabilities," when thinking about my university students. I made that choice because many of the characteristics that are labeled as "disabilities" may be either benign variations or even possible advantages in different situations. (See the 2022 LinkedIn article "Learning Disability vs Learning Difference" by Jess Arce for more information on this distinction.) The following are a few examples.

Height. A simple analogy to height might help. Being unusually tall or unusually short can make life challenging in a world that is primarily designed for people of moderate heights: taller people don't fit well in cars or theatre seats; shorter people may not be able to reach items on the top shelf at grocery stores or see over the people in front of them at a theatre; and both groups have trouble finding clothes that fit. (Then again, I have an adult daughter who is a petite person and has found that clothes in the boys sections at stores may fit better and be less expensive.) As such, being an outlier on height might seem like a problem, or maybe even, in extreme cases, a disability. However, the effect is context sensitive. When it comes to sports, taller people can have an advantage in basketball and volleyball, while shorter people can have an advantage in gymnastics and powerlifting. (See this Livestrong article for information about height and powerlifting.) As such, a characteristic that might be a challenge in one context may become a potential advantage in another.

ADHD. There is evidence that, in a similar manner, some of the characteristics that are often considered "learning disabilities" may have advantages in different contexts. For example, a 2021 article by Zia Sherrell in MedicalNewsToday was entitled "6 strengths and benefits of ADHD." Those potential strengths and benefits included: hyperfocus, resilience, creativity, conversational skills and humanity, spontaneity and courage, and high energy.

Autism. In a 2023 article on AutismBC entitled "Let’s Celebrate the Pros of Being Autistic," author Aly Laube listed the following potential strengths of people on the autism spectrum: pattern recognition, planning and sorting skills, hyperlexia, knowing oneself, strong sense of justice, fun with special interests, critical thinking skills, immersion in sensory experiences, bonding with other autistic people, and expression through creativity

Other differences. Similar lists of potential benefits have been proposed for other circumstances, including:

• Depression: "Can Depression be Good for You?"
• Anxiety: "The Benefits of Anxiety and Nervousness"
• Dyslexia: "The Advantages of Dyslexia"

An important qualification. I don't mean to be na?ve and suggest, in the words of Professor Pangloss in Voltaire's Candide, that "All is for the best in this best of all possible worlds." (As a reminder, Candide is a work of satire and this quote is direct dig by Voltaire at Gottfried Leibniz's philosophy of "theodicy.") That is, I don't pretend that these differences have only positive aspects and no negative aspects. I know from both personal experience and from the experiences of people around me that ADHD, autism, depression, anxiety, and/or dyslexia can make life extraordinarily challenging and that they can be profoundly distressing, even completely overwhelming. For that reason, I have emphasized the potential benefits, and have not given any estimate of how often things work out well for people. My point in mentioning these potential benefits is that some challenges are better understood as poor matches between the person and the context –?that is, as accessibility issues – and not as global deficits.

Differences that affect data work

When it comes to working with numbers and data, there are some common learning differences that can make the process more difficult. These differences include:

• Low numeracy, where "numeracy" refers to "the ability to understand, reason with, and apply simple numerical concepts"
• Math difficulties, which refer to "the incomplete mastery of number facts, computational weakness, difficulty transferring knowledge," and so on
• Dyscalculia, which refers to "a disability resulting in difficulty learning or comprehending arithmetic, such as difficulty in understanding numbers, learning how to manipulate numbers, performing mathematical calculations, and learning facts in mathematics"
• Low graphicacy, where "graphicacy" refers to the ability to understand and present information in the form of sketches, photographs, diagrams, maps, plans, charts, graphs and other non-textual formats

Some of these differences are more common than others. For example, dyscalculia pertains to approximately 3% to 6% of the population (Wikipedia: "Dyscalculia"). On the other hand, as much as 30% of the adult population in the US has low levels of numeracy, as shown in the following graph from the US National Center for Education Statistics:

In fact, according to research by the Organization for Economic Cooperation and Development (OECD), only 12% – 1 out of 8 – of US adults aged 16 to 65 performed at the highest proficiency level (a 4 or 5 out of 5) on problem solving in math, based on the 2012 Program for the International Assessment of Adult Competencies, or PIAAC. A sample task at this level was to "review two stacked-column bar graphs representing how many years of schooling men and women in Mexico have had by decade, then identify the percentage of men who had more than 6 years of schooling in 1970." This is the kind of task that can help people get useful insights from data, so it is clear that there is a great need for making data work more accessible to a wider range of people.

Dylan Lynn: A statistician with dyscalculia

As evidence that learning differences don't automatically mean that a person should avoid a field like statistics or data science, there is the wonderful example of Dylan Lynn, who is a statistician with dyscalculia. That learning difference made the calculations involved in her classes particularly challenging, and also led many of her teachers and advisors to suggest that she drop out of math and statistics. However, that advice ignored the things that made data work attractive to her, such as interpreting and applying the insights that came from analyses. (I mentioned Dylan's story briefly in my newsletter for January of 2024, entitled "Making ourselves useful.")

In the video above, Adrianne Meldrum of Made for Math interviews Dylan Lynn. She describe the collection of steps she used to make math and statistics more accessible for herself. It's exasperating that she had to find ways to make her own accommodations, often fighting with her teachers and advisors along the way, but her insights will hopefully spare other people from having to go through the same struggles. She describes these solutions in greater detail in an article she co-wrote for Education Sciences and on the website Dyscalculia.org. Her solutions included practices like using grid paper, colored mechanical pencils, working through one step at a time, consistent symbols, and so on. I was then surprised to learn that several of my students with various math differences had discovered many of the same approaches, which leads me to the next section: what to do.

What to do

You may have a learning difference that makes data work difficult or unappealing for you. Or, you may be in a position to make your data work more accessible to people with learning differences and math differences. I certainly am: every semester at my university, I teach 300 to 400 students in multiple sections of the introductory statistics course for students in the behavioral sciences. Each semester I receive a collection of accommodations letters from the Accessibility Services Office, and I know that those letters underestimate the number of students who could potentially benefit from a more accessible approach to statistics and data analysis.

I'm currently in the process of writing a new textbook for my class, which I hope to have available later this year as a free, Open Educational Resource (OER) text, along with videos, presentations, and other course materials. I had been working on this text for approximately a year before I gained a greater understanding of learning and math differences. Consequently, I started the project over again, with several principles in mind. Many of the principles came from Dylan Lynn's experiences. Additional principles came from the UK's Design in Government blog, which discussed "Designing for people with dyscalculia and low numeracy." These principles are summarized in this chart:

These principles have had a profound effect on how I teach data work. For example, I now attempt to explain every concept and procedure 3 times using different modalities:

1. Visually, using graphs and images for each step
2. Verbally, using written descriptions of each step in a numbered list
3. Symbolically, using standard symbols and formulas

My goal is that a person would be able to understand and conduct the procedure using any of the 3 approaches, depending on what they are most comfortable with. (I considered putting the symbolic, formula-based explanations in the appendix, until one of my students reminded me that many people, including himself as a person with autism, preferred the formula-based approach. He then recommended that I keep it in the main text as an option, which I have done.) As expected, this approach requires more effort, but I believe that it will make the content more accessible to my students and it will make the benefits of data more accessible in their professional work.

In addition, I do the following as often as possible:

• Describe the data. I say what is being measured –?number of children, average time spent on a website, distance traveled on public transportation – and give the units, such as kilometers, miles, and so on. I do this even when I'm using made-up numbers for a quick demonstration.
• Use consistent symbols. There are often many different names or symbols that can be used for the same concept, such as the mean: M (the APA's preferred symbol), X with an overbar, a Greek mu, and so on. Unless I am specifically giving the symbolic, formula-based approach to something, I use words, but when I use symbols, I use the same ones throughout.
• Use consistent formulas. For example, I universally indicate multiplication with parentheses, and I organize division vertically, with a numerator on top, a horizontal fraction bar in the middle, and the denominator on the bottom.
• Use brackets instead of subscripts for indexing. When referring to a data point in an array, I use square brackets – a convention from computer science – instead of subscripts – a convention from mathematics – because it keeps the index numbers the same size as other text and is less likely to be confused with other operations.
• Include the research question, as well as the interpretation and applications of the analysis. This places the work into a specific, meaningful context, and it also invites the students to use their own logic and experience to evaluate the results. Again, I try do this even when I'm using made-up numbers for a quick demonstration.

February Office Hour

My online Office Hour in February of 2022 was dedicated to this topic. You can see the video of that session here:

I will share additional tips and insights on accessibility as I continue my work. If you have your own experiences with data work either being difficult to access or made more accessible, it would be wonderful if you could share those below. And, as always, thanks for joining me.