Hi all,

I have a very me-specific question about formatting LDA output. There's a lot of exposition to this question, and I think it's important to read it all (unfortunately) in order to "get" it. Thanks so much for your time, and sorry about the text wall.

Here's a description of my experiment

I'm measuring brain response using EEG (microvolts) from research study participants as they're playing a button pressing game. The basic structure of the game is (1) the participant sees a cue [flash of light], (2) the participant has to press a response button when they detect the light. The outcome measure of the task is how fast they press (milliseconds)... a trial with "good" task performance is  one in which the participant presses the button promptly, and a trial with "bad" performance is one where it takes the participant a long time to press (say, because they've zoned out).

Here's the analysis I'm attempting

What I'm trying to do is classify trials into 'fastest 25% of trials," "middle 25-50%", "middle 50-75%" and "slowest 25%" based on the brain response using linear discrimination analysis. Since every person is different (some people are just faster or slower than others), I'm running one model per person, and then comparing model performance across people. What I'd ultimately like to see is if there are factors (like, say, participant age or clinical diagnosis) that influence how well the model performs. What I care about is "% correctly classified" (see Screenshot1, yellow highlighted section)

This is what I want from SPSS

Is there a way to have SPSS return only the % correctly classified? Even better: is there a way to have SPSS format output like this (see Screenshot2). Remember: my goal is to compare model performance across participants.

Thanks!
KW


------------------------------
Kathryn WT
------------------------------

#SPSSStatistics