Stats Doesn't Have to be Scary!

Looking for a free, open-source, easy-to-use statistics program? If you haven't heard about JASP before, then I suggest you read my blog more often! 😂 I've already promoted the use of JASP in the classroom and in research, despite a few limitations such as the inability to edit graphs.

Check out the full playlist: 

https://www.youtube.com/playlist?list=PLvPJKAgYsyoJxXZn-fTmqxF_GfOQ-htvJ

In trying to probe the limits of JASP, I uploaded a dataset with over 40,000 rows and 6 or 7 columns of data. It took a minute to upload such a large file, but there was no problem with running analyses even on a large dataset such as that one!

I think most behavioral researchers could spend their entire careers using this program, as long as they also have another program handy to generate publication-quality graphs. R is a popular solution for this (though you'll have to learn a bit of programming to use it).

Benjamin Nanes, MD, PhD, recommends additional options such as ImageJ or Inkscape. They're free, which is a big plus for impoverished graduate students and/or those who simply want to avoid the hassle of trying to get a license for SPSS or another such program on their personal computers.

Though I haven't tried these myself (yet!), I trust Dr. Nanes' recommendations and plan to try them out soon. Another option, also recommended by Dr. Nanes, is Inkscape: this could be used to add text (such as axis labels) to the graph generated by JASP and export it in a vector format that your journal will accept.

Everything I've said so far about JASP also goes for jamovi, another free and open-source program with a user-friendly interface. JASP started development before jamovi, so it's a little further along in its capabilities, but the original lead programmer for JASP is the lead programmer for jamovi, so there are many similarities between the programs--and I like both of them! jamovi does have a few features that JASP lacks, including the ability to see the R syntax for a given operation. This makes jamovi a great bridge for those who would like to learn R!

Since both JASP and jamovi are based on R but provide a far more visually appealing user interface, the analyses are trustworthy (though I've double-checked some analyses myself) and the programs themselves are easy to use.

In any case, if you're wondering why I like JASP so much, I made and edited a series of videos yesterday showing how to install JASP, upload files, and run most of the common tests in JASP 0.8.6.0. I've compiled these videos into a YouTube playlist; note that the instructions for jamovi are going to be quite similar.

screenshot of JASP 0.8.6.0 from my own computer

If you haven't already tried JASP or jamovi, what are you waiting for?

***

Wondering about the social media usage of actual college students? 

Check out the results of this totally informal—but real—survey.

In case you missed it, I review some fantastic, easy-to-use, and FREE stats programs here.

For more help explaining statistical concepts and when to use them, 

please download my freely available PDF guide here!

The Simple Life: Graphs to aid understanding of the results

In my first publication, my advisor and I created graphs according to the guidelines listed on the Judgment and Decision Making journal's webpage. My advisor and I both maintain a general preference to see results presented visually, in the form of easily-understood, properly-labeled graphs. However, in correspondence with JDM's editor, he felt that the results were simple enough to understand, and the graphs therefore weren't really necessary.

I think the editor's idea was that:
a) not including the graphs would save space, and
b) including the graphs would probably involve a bunch of difficult/time-consuming formatting. I certainly don't blame the editor for advising us to exclude the graphs; I understand perfectly where he's coming from!

So, to save time, space, and effort, our graphs were never published. Until now.

The graph for Experiment 1, which fits best with the information presented near the top of p. 305, presents the proportion of participants' decisions that were consistent with pre-exposure (bar on left) and with the recognition heuristic (as determined by participants' self-reported recognition; bar on right). Error bars represent 95% confidence intervals.



The graph for Experiment 2, below, illustrates the data presented on p. 307-308. The dots represent the mean proportion of choices that were consistent with the recognition heuristic, for participants in each of the three training conditions. Again, error bars indicate 95% confidence intervals, which I recommend reporting as part of a more complete picture of your data.

Here's a link to the full-size .jpg image for the first graph, and here's the link for the second graph.