Pre-Mid-Post significance test

Question

Is there a significance test for pre-mid-post comparisons? For example, participants took a survey answering questions about different systems for a variable such as likeability and they answered these questions before seeing the systems, in the middle of two experiences, and then after the final experience. How would I look at these ratings and examine whether there were significant differences between pre-mid-post?

Answer 1

You could do an ANOVA analysis and look at whether the groups were the same or different. Then if you find that there is significant differences you could look into where those differences are.

Answer 2

Typically for data like this you would use a one way repeated measures ANOVA (see https://statistics.laerd.com/spss-tutorials/one-way-anova-repeated-measures-using-spss-statistics.php for details). However, you mentioned using a likert-scale, so it's worth noting that if you want to look at changes on an individual item, your response variable will be ordinal (unlike in the case of a sum or average of several items, which approximate continuous variables) and therefore unsuited for a standard ANOVA approach. In these cases you might want to use a Friedman test instead (see https://www.sciencedirect.com/topics/medicine-and-dentistry/friedman-test for more details). Alternatively, if you're expecting your response variable to consistently move in the same direction over time, linear trend estimation (https://en.wikipedia.org/wiki/Linear_trend_estimation) may also be appropriate.