Most people have heard about the IQ test. Its popularity makes it an easy example for statistics classes. However, measuring something as vast as intelligence is difficult. In this post, I talk about how a podcast episode about intelligence testing can be used as an example of statistics and research methods in the real world.
Then I began teaching Introductory Statistics. The course was pre-prepped by a faculty member at my graduate institution (thanks, Dr. Wood!), and used IQ scores in examples throughout the semester. IQ scores are an easy, familiar construct that many students could get behind, and translate well to many different portions of the course. Distributions, the normal curve, z-scores, etc. The list goes on and on.
But for all its teaching utility, using IQ scores as an example always irked me. I was using it as a tool for teaching key stats concepts, but I wasn’t addressing any of the misconceptions that students have about intelligence testing. What is an IQ score? What does it actually measure? Why and how are IQ scores used?
Enter Radiolab’s G series. It is all about the concept of general intelligence. It talks about the good and the bad related to standardized intelligence testing. There are three episodes in the series (all linked below). I’m going to focus on the first episode, and its relevance to statistics and research methods concepts.
G: The Miseducation of Larry P.
I won’t go into a detailed summary about this episode (Radiolab does a great job of that on their own, check out their website). Basically, the episode centers around California’s Larry P. case. The Larry P. case banned the use of intelligence testing on African American students in the California school system. The case demonstrated intelligence testing was largely biased, such that White children tended to score higher than African American children. This caused a disproportionate number of African American children to be erroneously placed in special education classrooms. [Side note: the article I linked above by Powers, Hagans-Murillo, and Restori (2004) makes the case that this issue is on-going.] In The Miseducation of Larry P. Radiolab provides personal perspective, talking with the child (now a man) that ignited the case, and discussing the eye-opening IQ data that biases against people of color.
While listening to the episode, I had a lightbulb moment. Around 45 minutes into the episode, they talk about the IQ data from the Larry P. case, discussing problematic research methods and misinterpreted data. Here are where two major teachable moments come in:
1. The IQ tests at the time (1979) were normed with White samples, not minorities. Teachable moment: generalizability, sampling, populations.
2. Some African American students scored high on the tests, and some scored low, as did White students. The range of the distribution of scores was similar for the two groups. However, the average score for African American students was lower than that of the White students. Teachable moment: distribution of data (Could the data for either group be skewed? If so, how? If not, why?), data interpretation (How does your interpretation of the data differ if you use the range of scores versus the mean for each group?)
I envision an assignment in an Intro Stats or Research Methods course prompting students to connect this episode to key concepts from class. The episode also touches on psychometrics, eugenics, civil rights, and social justice, among other topics. Many different classes could use this episode as a real-world example of research methods and statistics at work.
The episode that follows, G: Problem Space, addresses some of the questions that the first episode left on the table. Have the tests been changed since the 1970’s? Is there still a ban on intelligence tests for African American students in the California school system? Does the ban only apply to African Americans? Are there ways other than intelligence testing to assess whether students need special accommodations in the classroom? Spoiler alert: the answer to all those questions is yes. They also present some compelling reasons for why intelligence testing is still important, despite the numerous flaws (I’ll leave you on a cliff hanger for this one, go listen to the series already!).
In the last episode, G: Relative Genius, Radiolab tells the story of Albert Einstein. Or rather, the story of Albert Einstein’s brain. This episode could be particularly useful for a Neuro course, covering the evolution of neuroscience as a field. It could also prompt interesting nature vs. nurture discussions.
If you’re like me and love podcasts, or are intrigued by the idea of using podcasts in your courses, check out my list of “teachable” podcasts here. You can also check out Ciara’s list of video resources here. If you have video or podcast recommendations of your own, you can send them to us via a form on our Content Resources page.
How else could Radiolab’s G series be used in the classroom? Have you assigned podcasts to your students? What is your go-to example for Stats and Research Methods? Please share your thoughts! Comment below or email us at firstname.lastname@example.org.
Written by Karly Schleiche