Diversity in statistics

Chelsea-Parlett · May 1, 2017 01:17 PM

Uncertainty creates a need for statistics – if you are certain about something, you don’t need to use statistics on it. And uncertainty is created by variation. If everyone was the exact same height, what good would the mean and standard deviation be? We'd already know what we wanted to know without statistics, and don’t even think about doing a t-test (it would require you to divide by zero, which is a big no-no).

The variation in statistics parallels our need for variation in our statisticians. April was National Mathematics and Statistics Awareness Month, and many different groups worked to create spaces and activities to raise awareness about math and stats.

But sometimes I get the feeling people are pretty aware of stats....It’s what they think about it that is a problem.

When I tell people that that I love stats they say something like, “Oh, wow, I hated stats when I was [insert time when you learn stats].” And while that means job availability and exciting project opportunities for those of us who have somehow come out the other side loving stats, it also shows how many others are slipping through the cracks.

Students at all levels of education are experiencing dread and anxiety over required statistics courses, and women and minorities may suffer the most [1]. The effects of statistics anxiety reaches far beyond just our field since most STEM majors require one or more statistics courses. Keeping students in the STEM “pipeline” is hard, and dread and anxiety over required statistics courses deepens the problem. The anxiety comes from the idea that statistics is convoluted and difficult to understand. It’s so pervasive that educational research articles on “dread courses” usually look at a statistics course.

If we want to utilize the talents, backgrounds, and perspectives of a diverse group of statisticians, we need to get a diverse group of people into the field in the first place. We need Frequentists and Bayesians, biostatisticians and epidemiologists. We need to make changes to the way we teach statistics before students even reach college. We need to fight against the dislike of and anxiety over statistics that causes otherwise inclined students to struggle. If we let a majority of women, minorities, or people who “don’t like math” to walk away from the field of statistics, we miss out on their unique opinions and strengths.

So what can you do? The options are endless, and we all have different experiences and backgrounds, so we will all have different ideas about how to improve statistics’ reputation (#diversity), but here are a few ideas to start us out:

Explain or leave out mathematical and statistical jargon when you can. It's easy to get lost when someone starts talking about Borel sets or oblique rotations. Technical language can be useful and save time, but leaving it out when you can allows newcomers and seasoned experts alike to understand the concepts at hand.
Invite people in to your statistical projects. There's nothing like working with real data, and letting someone in to your process helps create a place where they can ask questions as they explore real life statistical concepts.
Be a thoughtful statistical teacher and learner. As I learn more and more, I zoom out my statistical perspective; I can see the bigger pictures (seriously, it's all regression) and can think about how concepts fit together instead of memorizing seemingly unrelated formulas. Continue to learn, and then turn around and teach others the insight you've gained!

[1] Onwuegbuzie, A.J. (1998). The dimensions of statistics anxiety Louisiana Ed Research Journal, 23, 23-40. Onwuegbuzie, A. J. et al.,. (2003). Statistics Anxiety. Teaching in Higher Ed, 8(2).