I’ve written a few data analysis programs that I use a lot. Try them out! (If you’d like the source code, just ask.)
Problem: You have a bunch of datafiles and you want to a) figure out which ones are analyzable and b) combine them into one big file.
Solution: Combine Files removes the drudgery from combining data files; it does it for you neatly and automatically and it can identify files with problems. It’ll remove duplicate column names, let you know how many rows are in each file so you can get rid of the incomplete files, and more. It also allows you to select problematic files and move them to a different folder for inspection.
Free Recall Scorer
Problem: You want to score free recall data.
Solution: Free Recall Scorer! You tell it the correct answers and what the subjects said (in any order). It’ll match them up, and it’ll even be lenient about spelling. It gives you detailed output including commission errors. Sorry, only single-word answers are allowed. Let the software do it’s job, save your humans time, and then let the humans inspect the data and make the tough calls.
Problem: You want to give subjects credit for correct responses even if they’re spelled wrong.
Solution: Lenient Scorer! You give it a list where each response is paired with the corresponding correct answer, and it will use a measure of letter overlap to tell you how accurate the response is. You can inspect the ones that aren’t quite perfect, or just call everything that scores > 74 correct.
Line Rows Up
Problem: Your subjects responded to the same stimulus (e.g., a question) multiple times and the responses are on separate rows in your dataset. You want them all in the same row (so you can compute correlations, etc).
Solution: Line Rows Up! You tell the program which columns of your dataset to use as keys. It will go through the whole thing, and it if it finds a set of keys it has seen before, it will put them in the same row as the previous one. It’s super easy to use and once you know how to use it it’s incredibly useful.
Problem: You want to compute a mean for each subject in your study (or maybe multiple means if you used a within-subject design). Pivot tables can do this, but maybe you want to compute a median, or a correlation, or a gamma correlation, for each subject. Or do other stuff.
Solution: Stat Buddy! You identify variables as dependent, between, or within, and then tell it what type of analysis you want (mean, correlation, etc). It will give you a nicely formatted output file automatically and without human error.