iNeed to keep it as simple as possible . I am using quota sampling and think a questionnaire might be easier to analyse than a likert . So my question is how do I analyse questionnaire and Likert?I see that non parametric t-test for likert but don’t understand how the data is extracted and analysed ?and not sure with questionnaire. I am absolute beginner so need to keep it simple .thank you! Btw measuring confidence in nurses ability to handover following the intervention of a handover tool.
By "Likert" do mean individual Likert-type items, or do you mean using several items, summed together, to make a "scale"? What's an example of what you mean by "questionnaire"?
I just read in research papers that people used a “likert scale “with questions in it and questionnaire is a questionnaire ,not sure how to expand on that,just questions to ask the sample before and after the intervention.thank you
How you analyze these kinds of responses will depend upon the number of questions, the type of questions, and maybe also what other demographic data you want to collect.
To keep things simple, you may want to avoid trying to compose a Likert scale. Doing this properly requires a few extra steps, probably isn't worth the bother, and may be subject to criticism if not done correctly.
Instead you can analyze questions individually. It's up to you what form these questions take, and the type of response will determine the appropriate analysis.
For example, you could have individual Likert-type item questions ("On a scale of 1 to 5..."); you could have yes/no questions; and so on.
For what you are doing, and considering that you are dealing with educated adults, having more-or-less numeric responses will be beneficial. I would probably consider "On a scale of 1 to 10" as numeric.
I recommend having a few questions that try to get at what you are thinking for your main point. For example, a question on "how confident are you" and a question on "how comfortable are you" may get different responses.
It's also helpful to have a couple of demographic questions. Maybe what gender the respondent is, what department they are in, and so on. This is helpful just to describe your respondents, but also in case it makes sense to compare different groups (Did females respond differently from males?).
In order to do the analysis correctly, it's important that you are able to tie the "before" response for a respondent to the "after" response for the same respondent. This is what allows you to do the pairing in the paired t-test or other paired test. If you are able to ask people's names or whatever, that makes it easy. If not, some code could be used, like initials and favorite color, or birthday and pet name.
Once you formulate your questions, you might ask for more details about the analysis. It's difficult to say without knowing what kind of questions they are.
- Badges
- Science topic
- Similar topics
- Citations