Is post-hoc analysis necessary in this case?

Sorry for the lengthened explanation, I tried to be as clear as possible:

I am required to assess a paper with a cross sectional study design, in which the authors surveyed participants about certain perceived barriers and certain attitudes.
Descriptive statistics (mean and standard deviation of the likert-scale scores resulting from the survey questions) and significance tests were done to assess the relationship between perceived barriers that affect certain attitudes and demographical factors like age, sex, ethnicity etc…
A significant relationship was found between one of the perceived barriers and ethnicity after conducting a one-way ANOVA to compare the means of perceived barriers scores between ethnicity categories and then the authors proceeded to conduct a post-hoc analysis using LSD test to make comparisons between mean perceived barrier score for every ethnicity with all the others ethnicity scores.

My question is why was it necessary to conduct a post-hoc analysis in this case?
Wasn’t it enough that we found a significant association after conducting the ANOVA test?
Also, I read that authors looking to present results from a post hoc analysis should always pair it with the power of the study to find differences and that power of the study is an important metric that gives credibility to any post hoc analysis.

Thank you

1 Like

because they want to understand where the difference lies among the ethnicity categories. It’s pretty standard eg some stats software would produce it even if you don’t request it

1 Like

Thank you, much appreciated!

1 Like