Hello all,
I have 20 patients with the survey, including the Likert scale; I want to have a "Wilcoxson sign rand test" to do comparison for total score not the sub-score; is that correct to say the total score is the total of each row for the Likert scale?
pre_ Q1 ………..pre_Q5 total_pre post_Q1 ………….. post_Q5 total_post
1
2
3
.
.
.
10
or should I add up the ordinal data associated with each value sentiment?
Thank you!
That would depend on the scoring system laid out by the scale if it is a standardised one. If it is a scale that you yourself have developed, you can add all the items in the scale and arrive at the total. I have a YouTube video on how to do this via SPSS/PSPP.
https://www.youtube.com/watch?v=vVfHjNRxQOw
Hey Bahar
I probably doesn't really matter if you sum the scores or compute a mean, this won't make a difference in any way when you do correlations etc. The mean might be better because it is in the scale of the Likert response format. For example, if you are using a 1-5 Likert scale and the mean is 1.2 you know that the sample is scoring low, but if it is 4.5 you know they are scoring high. It's tricker to think about summed scores as you have to do a bit of arithmetic to work out the range of possible scores. Using the mean also handles missing data a bit better as you can divide the summed total by the number of items completed, not always the total number of items.
Mark
Bahar Ysr You need to add all the items in the scale as seen in the video. When you do that, the total score of each of the respondent/patient will appear automatically.
- Badges
- Science topic
- Similar topics
- Filing