Potential new example for chapter 19 #501
Description
I always lament having to find real one-sample tests where the data are accessible.
Today, I found Natural behaviour is learned through dopamine-mediated reinforcement, which, brace yourselves, does left-tailed and right-tailed one-sample t-tests and a two-tailed paired-sample t-test all in one paper! (See Figure 2g)
I was able to write a fun lab that walks students through conducting a power analysis, collecting the data from the journal website (which seems deliberately annoying so I attached a .csv), visually summarizing the data, conducting resampling to assess whether a t-test is appropriate (n=25 < 30 < infinity), and actually performing the t-tests (effect sizes and confidence intervals, too). Doing all three tests gives them ample practice and enables them to compare and contrast across different test types.
We'll wrap up the semester learning to conduct bootstrap HT/CI and permutation HT/CI for a population mean, so these data will be able to follow the students through that.
Anyway, I thought I'd share because I often spend way too much time searching for these types of examples and I hope I can help someone save an hour or so. 8]
From William Cipolli on Isostats 4/8/2025. I have the data in an email folder saved in IMS called "exercises".