Analyzing Customer Satisfaction Indexes

[rafael_josem]

Hello,

I work for a company and we recently started to measure our customer satisfaction index. We chose the sample size based on a well-known formula and then we chose the interviewees randomly.

Anyways, In six months from now we would like to do the research again and compare both studies to see if there is any statistical change. I thought of using hypothesis testing, but im not sure which to use:

- Paired t test?
- two samples testing?

I Supposed that if I choose the first one, I'll have to survey the same people from the first research, and if I use the second one I'll need to draw a new sample?

Im bit confused...

Thanks in advance,

Rafael J.

 

[Antonio Vieira]

I believe what you have to do is what we are trying to perform now - check the reliability of the questionnaire by using some statistical techniques such as testing the internal consistency of each quality dimension (group of questions for the same subject) in the questionnaire.
We are trying to use the Cronbach's Alpha estimate method, but the results are not being what we expected...