Testing for flatline responders in MTurk data

I am working with data collected on the MTurk platform that links directly to a Qualtrics survey. The data is easy to format, so there's not a lot of wrangling there.

I have many questionnaires with different lengths, all of which have Cronbach's alpha of over .7 and most are in the high .88-.97 range. I have control questions in my surveys to screen out bots and participants who are not conscientious responders.

On questionnaire using a Likert scale from 1-5, I know it's possible to get good alphas if someone responds with all 1, someone responds all 2, someone responds all 3, etc. I'm curious if anyone knows of a package or method for identifying these "flatliners" either quantitatively or possibly visually if you could think of how to plot this using a heatmap. Ideally, you could tell if someone was a flatline because their data would look more like Tetris blocks if colored in rather than confetti. Does this make sense?

Thank you!

This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.

If you have a query related to it or one of the replies, start a new topic and refer back with a link.