On exploratory research


My first thread was not that all successful in terms of replies (probably due to my limited knowledge of statistics and consequently a not-so-clear thread), so I taught of trying once again.

I am analysing data from 56 respondents on 33 Likert scale (1-5: Disagree – Agree) questions that are divided into three sections, as part of an exploratory research. Basically through the research I want to confirm my belief that a certain correlation exists.

1. Apart from basic descriptive statistics what other statistics tools can I use to draw more information from responses? I will be using t-test to analyse responses according to different groups such as male/female

2. I tested the reliability of the questions for each section separately (cronbach) and got the following readings Sec 1 - .738 Sec 2 - .653
Sec 3 - .662. If I recall correctly .7 is the acceptable minimum but I was told that according to someone surnamed Nunnally .6 is considered as suitable for exploratory research.

3. Other advice…. more than welcome.

I am using SPSS


TS Contributor
What correlations are you trying to show? I would focus on that. This will lead you directly to the types of analyses / tests you should be running.

Your reliability metrics seem pretty good to me.
Thank you John!

The research is part of a dissertation wherein I am investigating the relationship between "conventional marketing", "entrepreneurial marketing" and "own creativity perception". The dissertaion title is "Between marketing and creativity tools which are the most suitable for SME".

The reserach was intended to verify my belief that there exists a

-ve correlation between conv.marketing and entr.marketing (-.68)
-ve correlation between conv.marketing and creativity perception (-.29)
+ve correlation between entrp.marketing and creativity perception (.49)

In brackets are the correlation that actually resulted following SPSS analysis.

Basically I asked about some other statistical tools to beef up my analysis chapter (that is if I am able to understand what statistics they return)..