Wednesday, December 5, 2007

More on data analysis

Cronbach Alpha Reliability
- to find out how reliable a questionnaire is
- to find out if all the items in the questionnaire are reliable ( ~0.6-0.9 would be good)
- to find out if the items can be repeatedly used
( Low reliability might be due to: context used, wordings of items, interpretation of statement by participants etc)
- Too high a cronbach alpha reliability is not desirable. It could mean that the particular item is redundant.

Paired Sample t-test
The Paired Samples T Test compares the means of two variables. It computes the difference between the two variables for each case, and tests to see if the average difference is significantly different from zero.

------------------------------------------------------------------------------------------------
Preparing the data...
In this case, we will pair Pref_SC with Act_SC, Pref_IS with Act_IS etc.
In SPSS,
(1) Data input the mean values (for both pref and act)
(2) Check that all the data are of the same decimal places
(3) To analyse: Click on analyze --> compare means --> paired-sample T-test
(4) Choose the different pairs of data for comparison
(5) Under options: Set confidence interval to 95% --> 'ok'

To save the data in a different format, go to the left hand side, select the data that we want to save, right click and choose the file format as required.

------------------------------------------------------------------------------------------------
Analysing the data...
As we are using 95% confidence interval, an alpha value, p <0.05 means that the difference is significant.
For eg, if p = 0.001, it means that there is a significance difference between the actual mean and preferred mean (a gap exists). So, some interventions will be needed to address these differences.

No comments: