The same multiple choice questions will be asked to both populations. They aim to assess knowledge, attitudes and use of a certain technology.
Some questions have Yes or No, others have more than 2 categories.
For yes/no questions you can use an independent samples t-Test (in this case, a test for "differences in proportions"). If your questions with multiple categories are ordinal, you may be able to combine them in scales that you could use in t-Tests. Otherwise, you will need to use non-parametric statistics.
For categorical data (e.g., yes/no questions), it is likely that chi-square analyses would be appropriate. What you do with your other data depends on whether those data are individual items or, as David Morgan above suggests might be possible, you can form scales.
Certainly independent-samples t-tests are a possibility. Depending on the nature of your data (you don't provide sufficient information about this), Mann-Whitney U-tests might also be appropriate.
Before answering this question I'd want to know what you mean by compare (e.g., is one a lot more X), and how are you going to use this (e.g., deciding which company to work for, where to holiday, etc.). Also more info about the instrument and samples (and how each sampled could be of value. Cristian Ramos-Vera , David L Morgan , and Robert Trevethan are likely making some assumptions or found these out from emails from you, so they might also be able to provide answers for me too.
Daniel Wright. Thanks for contributing. In my case, you're right: I certainly made some assumptions - and I acknowledge that those assumptions might not have had appropriate foundations. I did not have any information additional to what was in the question.
I was primarily trying to just "keep things moving". I agree with you that there's insufficient information in the question (sorry, Kelsie) for others to step in and be really helpful.
Kelsie, pls to feel free to get back with more information about your aims and data. We might then be able to help you more. Apart from that, if you are a research student at a university, your supervisors should be able to help you.
FIRSTLY, you should assess normality distribution after that decided the appropriate statistical test if your data are parametric use independent t-test if non-parametrric Mann-Whitney u test to compare between two different population but considering your sample and other characteristics that may affect your results as demographic.
- Badges
- Science topic
- Similar topics
- Statistics
- Survey