I'm doing a statistical analysis on a survey based on a likert scale and I start by saying that I'm not an expert of statistics.

What I want to understand is if there are significant differences in answers among age groups. Each question was to be answered choosing from 5 alternatives.

If I count each alternative as a count variable for every question (and so I have each age sample with five variables, should I make a square root transformation or a log transformation to make variance homogenous in order to perform a one way ANOVA and then check residuals normality?

I'm a bit confused about what to do. I've read that it is possible to consider the negative binomial model. If so, it would be possible to compare my samples without using ANOVA and how (generally speaking)?

Sorry for my English an thank you in advance


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy