In a new paper led by @jmbh.bsky.social & Alberto Jover, we assigned n~160 in a 2 week study to either ๐1-7 Likert or ๐๏ธ1-100 Visual Analogue Scale and compared results using Bayesian multilevel models.
1/
VAS (compared to Likert) showed
- higher within-person means
- higher lag-0/1 correlations
- lower within-person skew & response frequencies of exact 0s.
Most effects weak.
Detailed plots per item & category available in the paper, I'll show the means as an example:
Interesting study! Did you take a look at careless responding and response tendencies? It might be harder to do things like straightlining in VAS than in likert. I have to say this changed my understanding of the topic, since I was tought there isn't much of a difference. Are there moderators?