By Jessica Hartnett, Ph.D., Gannon University
On the very first day of my Introduction to Statistics class, I show my students this and tell them that upon successful completion of my course, they should add this to their resumes:
Novice data analysis using JASP software, including descriptive statistics, t-tests, ANOVA, chi-square, regression, and correlation
Over the course of the semester, I work with them, talking about different statistical tests, analyzing and interpreting countless examples using JASP, and learning basic, regimented APA style Method, Results, and Discussion section standards so my students can “talk” statistics. I want them to live up to that special skill claim and to feel comfortable doing statistics. I teach like this because I believe that statistics instructors are in a unique position to teach a core proficiency within our discipline that is also a specific, highly marketable skill.
Notice that I didn’t include performing statistics by hand anywhere in that paragraph. I do very little by hand calculations in my classes. Why? Because statisticians don’t. And if students need to understand the guts of data analysis, they will go on to graduate school in a quantitative field. And guess what? Most of our BA/BS students ARE NOT doing that. The American Psychological Association Center for Workforce Studies (APA, n.d.) counted up 3.5 million people in the US who have bachelor’s degrees in Psychology. So you know how many of them have PhDs in psychology? 4% of them do, and 10% of them have a Master’s in psychology. A full 53% stopped with the bachelor's degree and the remainder pursued post-baccalaureate degrees outside of psychology. As such, what skills do we need to teach to serve the majority of our students? Basic, novice stats skills, with the assumption that they will learn more in-depth statistics if they pursue graduate study that requires it.
Teach them statistics so that they can keep up with data and research within their careers and non-quantitative MS/MA programs. Teach them research methodology along with statistics so they can be valued by their employer and trusted to run the occasional correlation or ANOVA, or so that they can help with someone else’s data collection. Teach them enough about statistics that they are not going to fall for click-bait headlines that poorly summarize research.
Another benefit of not belaboring by-hand calculations is that it leaves you time to do other things. Like mastering analytic software, calculating alternatives to p-values, and teaching your students how to talk statistics. Rather than focusing on by-hand calculations, my students leave our class feeling confident in JASP and able to produce rough APA reports that include effect sizes and confidence intervals. Which brings me to my next point: Picking appropriate software for novice statisticians who probably aren’t going the academic route.
Most of your students are not going to graduate school and will probably never see SPSS again if you do use SPSS. I would wager that the small proportion of students you are teaching who do go on to graduate school likely won’t see SPSS again, either. And, your students aren’t rich and they are on the go, so let’s use free software options they can run off their own machines and, maybe, even tablets and mobile devices. You could use JASP, PSPP, R, Google Sheets, or Jamovi. Both JASP and Google Sheets can be used via web browser, for added flexibility. For a free-ish option, you can teach students to conduct statistical analyses via MS Excel, especially using the data analysis add-ins. I use JASP. It is intuitive and doesn’t take up a lot of RAM. Plus, if your students are graduate school bound, they can use the free JASP/R hybrid program, JAMOVI.
In terms of my mini-APA style reports, I have them create a Methods, Results, and Discussion section for each test we run. The Methods and Discussion are only one to two sentences long (I said mini) and the Result section teaches them the basics of in-text statistical reporting. My Methods sections also include effect sizes, CIs, and p-values. As I tell my students, the point of conducting statistical analysis is to share your efforts with others, many of whom do not understand statistics at all. Mini-APA style reports teach them this skill, and also allow me to assess my students’ understanding of the analysis.
*If you have any questions about teaching stats or need any help, feel free to email me: email@example.com. And then re-email me a week later because I can be horrible at email. Or just DM/follow me on Twitter, @notawful.
**Important caveats relevant to my argument: You may be teaching stats as part of a sequence. Consider what role you have in changing that sequence. How can you work with other instructors to maintain consistency for your students? In my department, we just switched to teaching our Intro Psych class with JASP, and we’re teaching our Stats Lab, Multivariate, and Psychometrics classes with R/Jamovi. This was a big shift but everyone was on board with the change. Now, we don’t have to charge our students a lab fee for our statistics class, either, which I think is an improvement. We also have small class sizes, 30 or less, and significantly smaller advanced classes, which are required for our Bachelor’s of Science track. Additionally, I pick my own textbook and I don’t teach as part of a stats/research sequence. I also teach mostly non-psychology majors.
American Psychological Association. (n.d.). CWS data tool: Degree pathways in psychology. Retrieved May 8, 2019, from https://www.apa.org/workforce/data-tools/degrees-pathways
Jessica L. Hartnett is an associate professor of psychology at Gannon University in Erie, PA. She enjoys studying novel methods for teaching statistics and research methods, best practices in obtaining informed consent, and positive psychology. In her spare time, she reflects upon how lucky she is to have a philosopher husband who understands the demands of an academic career and two beautiful sons who doesn’t care about the demands of her academic career in the least.