Proving a hypothesis- how on earth?!!

B

I'm very new to spss so please excuse my total ignorance on how to do this.

I'm hoping someone can please help!
I want to prove a hypothesis- brand trust impacts positively on brand loyalty.

I have 6 questions for BT which were answered under a 5 point likert scale.
I have 10 questions for BL which were answered under a 5 point likert scale.

1 is strongly agree, 5 strongly disagree.

How on earth do I go about proving the hypothesis? I'm a complete beginner to spss and all tutorials/YouTube vids seem to only use one variable to prove theories. What test should I do, how do I get all variables down to one each, then go about doing test? Please please help a hopelessly lost person 😩

T

You need to get a basic research methods book. Make a question (e.g., Does X impact on Y), and a hypothesis (e.g., I believe X will result in more Y), and operationalize it - which means makes the variables measurable. So for example, you need to say what you mean by brand trust, and what you mean by brand loyalty, and how you're measuring these things. That will allow you to test your hypothesis.

A research methods book should tell you about how to deal with the questionnaire data.

C

SPSS for Psychologists is a good text which talks you through the processes as well as giving examples of how you can apply the different stats SPSS can do. However, I'd also recommend a chat with your supervisor (assuming you have one) about how you're going to approach the analysis.

G

If you are a begginer in SPSS, you should explore the option of seeking guidance from a seasoned data analyst who will guide you on which tests are most appropriate for testing the hypotheses in order to save time.

T

Also you don't prove the hypothesis as such (the starting chapters of any research methods book will explain). You can only reject the null hypothesis (that brand trust has no positive impact on brand loyalty).

P

Some would say (I think increasingly?) that you shouldn't try and prove hypotheses / reject the null at all, but should instead use other measures of confidence? But let's not turn this into a philosophy of science thread :).

If this is for PhD work I think you need to speak to somebody (e.g. your supervisor) quite soon to discuss the kinds of statistics that would be the most use for you. You also do need to read some beginner / intermediate books on research methodology, (Neuman is a good start) before you begin analysing any data. This is important, because if you begin writing the thesis / doing the analysis in order to prove a hypothesis, you will lose all credibility.

R

I should be of help. Message me.

49512