I am looking for resources (books, online, articles) which have a step-by-step explanation of how to test hypothesis using Likert scale/item data. So far, I've learnt that non-parametric methods are best (Kruskal-Wallis, Mann Whitney, ordinal regression, etc.) However, I cannot find any source which has a comprehensive step by step guide on how to do any of these tests. Can anyone point me in the right direction? or give me step-by-step example? I need to understand the reasoning behind each step of the process, not just the code to use in R/SPSS
$\begingroup$
$\endgroup$
3
-
$\begingroup$ Please clarify what type of hypothesis tests you are planning to conduct. For example, comparing responses for different groups, providing distributions, looking for latent variable measures, etc. $\endgroup$– Gregg HCommented Apr 26, 2021 at 21:39
-
$\begingroup$ Just like @Gregg H said, regarding the hypothesis - what problem drove you to survey? What questions are you trying to answer? There are many ways you could statistically analyze likert data, but how you determine which method or methods are appropriate is driven by a problem you're trying to solve, a question you're trying to answer, or a hypothesis you're trying to test. $\endgroup$– KatCommented Apr 27, 2021 at 6:08
-
$\begingroup$ my hypothesis is as follows: @GreggH H0: There is no difference in the perceptions of the effectiveness of drills between military and civilian crews. H1: There is a difference in the perceptions of the effectiveness of drills between military and civilian crews. The questions used to collect data were 5-point Likert items, so there are multiple items. the survey was administered to both military and civilian crew members, my main aim is to compare responses for the two groups. $\endgroup$– JamesCommented Apr 27, 2021 at 7:17
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
Based on the follow-up in the comments, I would suggest that the appropriate strategy to analyze the data to answer this research question is to use latent variable model to obtain a single measure form the Likert responses. In brief, you will want to run a model that (1) estimates a single score for each individual from the survey item responses, and (2) compares the average of those latent scores. This can be accomplished by a multi-group confirmatory factor analysis (CFA).
A very good reference for how to do this in R is Latent Variable Modeling Using R by Beaujean. I have taught from this book, and students have shared that it is a great reference and is very accessible. It does require a basic understanding of how R works.
Happy to share more if requested.
