What statisical analyzation method should I use to analyse job satisfaction survey?

I am creating a job satisfaction survey and an employee knowledge survey to test the success of a training program. Participants are asked to fill out a series of questions pertaining to their knowledge and satisfaction in their job. They rate each question on a scale of 1-7. I'm looking for a good statiscal analysis method to test the reliability of the data and also present it.

What statisical analyzation method should I use to analyse job satisfaction survey?
The data collected can be plotted as a graph with the use of a frequency table. From this the three sigma levels can be worked out and also the 'trend' and 'shift' from the central tendency will give a clear picture from the bell curve you would have plotted. The 'kurtosis' of the graph if any will throw further light.


VR
Reply:Content Validity can be established by a panel of experts prior to administration.





Factor Analysis or more specificially principle components analysis can be used to establish the construct validity and identify the latent compenents.





Once you have the components and they make real world sense use Cronbach's alpha to test the internal consistency. After that if all is good you can consider getting into test-retest, split-half etc.
Reply:the question needs further clarification for a statistician to attempt to answer





1.) do you mean the reliability of the new method in producing better trained more satisfied workers ? if so do you have a control group (those without new training) to use in a test along with your treatment (those with new training) for a statistical comparison?


2.) are you going to use the same group ? if so you must first administer survey THEN the treatment THEN the survey again to have reliable data for stat. test


do this when you cannot have a control group because all employees must undergo the treatment (new training)





the crux of your issue concerns your definition of relability and what you wish to be reliable i suggest you go back think about this and ask your question again to a statistician your question is confusing and perhaps this is why you are not getting alot of reponses