Evening, RGers!
I have an 8-item variable in my questionnaire measured using a 5-point Likert scale, ranging from strongly disagree to strongly agree, and I have to do descriptive statistical analysis for this variable, which is named State of Readiness.
What is most appropriate to do: compute the overall median for this variable, or each item's median separately? If separately, do I mention all items, a few, or highest medians only? What norm should I follow?
P.S. This variable answers the question: How well is the sample ready ...?
Thank you.
So, you have only one variable and you want to make descriptive analysis. In these case, the two approaches are valid but when computing the overall median for the variable, it will allow for a more detailed analysis of strengths and weaknesses across different dimensions of readiness. You can apply the both approaches to obtain a comperhensive analysis.
It really just depends on what is of interest to you or your readers...
Your description suggests that you have 8 questions that together measure "State of Readiness"....
If that's the case, it seems that you would want to determine the median for each participant across the 8 questions. And call that "State of Readiness".
From there, you would present the summary statistics (minimum, 25th percentile, median, 75th percentile, maximum) for this "State of Readiness" variable.
But it's okay to look at the individual questions as well.
Do you just want to do descriptive statistics? If so, you then why not report both kinds of medians? If, however, you want to use the total scale as an independent or dependent variable, then you will need to do more work to assess its reliability and validity.
- Badges
- Science topic