My thesis title is : Important of user-generated contact in tourism decision making.

I am worried about my methodology. I am not sure what kind of data analysis that can I use.

I will use survey method. In my questionnaire, consist of 4 sections : Demographic, media habit of visitor, internet behavior of visitor, credibility of source information. For example, I will ask

-How do you rate the credibility of information from contents created by other users online? -Would you agree that information from other consumers on Internet is more reliable than advertising? -What us the most influential source of information when you are about to make a purchase? -How often do you post contents (rating, comments, picture, video....) on Internet about a purchase.

My questionaire is not Likert Scale.

So far, I only know I can use SPSS to analyse. But I really no idea at all what should I use to analyse. One type of analysis is enough for thesis?

  • This is not a good question for this site. – Thomas Oct 20 '19 at 11:45
  • You should ask your supervisor, they're supposed to help you with this kind of question. – Erwan Oct 20 '19 at 14:42

Since you are using 'will use' in your question I assume there is still time to make asjustments and will just quickly mention some things that you might still want to change in order your get it 'passed' as your thesis.

  • You say you do not use a Likert-scale. Why not still change it to a Likert scale? (I will also suggest you use a neutral (midpoint) option for all bipolar scale questions you use)
  • Another important change to make in your questions is that you are now (judging from the few example questions you posted) biasing your questions by how you formulate the stem. You are doing this atm in two manners (you should change this):
    • "..on Internet is more reliable than advertising?" should be something like "..on Internet is more or less reliable than advertising?"
    • "Would you agree that.." is likely to lead to acquiescence bias, you can change this to something like "do you think that..".

Getting back to your actual question.

The analysis is dependent on what you wish to 'test' and the 'sort' of data you have.
Most commonly you would have a research hypothesis that is 'tested' by seeing if scores (e.g. averages) on question/construct A are associated with (/'can predict' but not te be confused with causality) differences on question/construct B.
(e.g. you ask people how tall the they are in questions A and what size of pants they wear in question B, to test the hypothesis 'taller people tend to have pants in larger sizes. This would be tested by taking both averages and going for a simple t-test (if the variable would have been categorical for example we would have needed to choose a chi-square test instead)

So the actual test you would choose (e.g. SPSS) depends on this design/hypothesis (or you using 1 variable to predict 1 other variable?) and the datatypes of both the predictor and the dependent variable (nominal, ordinal, ratio etc).

This makes it difficult to answer this question without having a lot more information.
Here is a link to a simple in-browser tool that asks all those question about your research design and variables and rolls out the acceptable significance test (tool).

Hope this helps!

| improve this answer | |

Not the answer you're looking for? Browse other questions tagged or ask your own question.