I have been using JASP with my students, this quarter. So far they like it – as far as students actually ‘like’ statistical programs. They seem to prefer it over SPSS.
Things students like:
- the interface is easy
- the results show up immediately (on the right half of the screen). No more “confusing” working with separate output files
- it starts fast, it runs fast
- it runs off an USB-stick if needed
- the OSX -version does not look like an afterthought
- clicking an H1 hypothesis (with corresponding p-value) is handy and promotes insight
- The plots are useful easy to make (just click “Plots”)
In terms of usability, the program is much better than SPSS for the kind of analyses our students do. Most of the (classical statistics) things they need is in it.
Two small wishes
- an easy way to make a code book (1=man; 2=woman and so on). Perhaps we overlooked it.
- The classic confidence intervals are given in absolute terms. Let’s say I do a one-sample t-test on variable is age, against test value 40 years, then the classical output contains information such as sample mean (let’s say 36 years) , t-value and 95%-confidence (perhaps 35.1 ….36.9 years).
But if I do a Bayesian One-Sample t-test, my (95%) Credible Intervals is only given in terms of effect size δ. I can calculate the absolute numbers from that output, but it would be handy if they were presented in the output already.