Struggling with statistical tests of my bachelor thesis

Are there any experts on statistical tests on this board?

I‘m writing my bachelor thesis about early retirement and made a poll with 280 participants. Now I need to test my 4 hypothesis that I formulated in the beginning. But I‘m unsure about the right statistical tests. Anyone who could help me out?

1 Like

I assume the best person should be your supervisor?

3 Likes

He is rather unresponsive and his answers are always vague. But you are right, I should reach out to a professor.

Yes, if he is not doing his job, reach out to the professor.

You can also ask ChatGPT and use that as a starting point for more research.

3 Likes

Here is a starting point: Null hypothesis significance testing: a short tutorial - PMC

3 Likes

Not anymore :(.
But at the end you need to calculate the P-values and interpret them. (that should be the hard way) several tools can calculate the results.

I believe that you can also play with your data in some pre-built models.

2 Likes

Hi @Cortana

Sorry to hear that your supervisor is unresponsive.

I will try to help, however, it very much depends on what you want to test and a bit more info would be beneficial.
In general, 280 is a good amount of responses for a bachelor thesis!
Now the question is, what do you want to show with your 4 hypothesis?

Generally, start maybe with some general descriptive statistics about your dataset and the variables (you can easily do this with R, XL-STAT or SPSS).
You could also do a chi-square test to see whether there is a difference in the observed frequencies.
Then it depends a bit, maybe you wanna do some regression testing, or correlation testing (e.g., pearson and spearman). You could do a student t-test to check for differences (e.g., have women and men answered a question differently).

Hope that helps already. If not, let me know.

2 Likes

People might laugh about that idea, but the answer will surely be better than the advice you get from anyone not deep into the field. ^^
I would do exactly this. Get some inspiration there and why it would make sense to use a specific test for your data, and then check in with your supervisor whether he agrees this makes sense.

Out of experience I can tell you that he/she will be much more willing to help you if they see you have something on the table already rather than asking an empty question like “how should i continue?”.

3 Likes