Which statistical test to use for Likert scale questionaire I need help on determining what tests to use in SPSS. Let me provide some background on my research:


*

*I am doing a research on system automation of leave management in a company (which is still using a manual paper form to apply for leave).

*My hypothesis validates the automation of the different processes (independent variables) on the overall improvement of the leave management process (dependent). My control variables are age, gender and dept.

*I have developed a prototype of a system and provided a questionnaire in the form of a 5-point Likert scale to the participant. (Strongly Agree, Agree, Neutral, Disagree, Strongly Disagree).

*The questions measures their response on the statement provided on the test evaluation of the system. These questions are categorised based on the hypothesis.


Where do I start? 
 A: It sounds like your response (or "dependent") variable is perception of improvement eg individuals' response to a statement "leave management is better now than it was" (agree/disagree etc).  Assuming this is the case, you first need to clarify that you don't really test the hypothesis of improvement of the system, only retrospective perceptions of it.  
Putting that aside and given the data that you apparently have, it sounds like the technique you need is ordinal regression.  A quick google suggests this is possible in SPSS (I'm not an SPSS user so I won't try to recommend any of the links).
A: "Where do I start?"
You should always start a research study by clearly defining your objective (supported by literature review) and then identify your hypothesis to be tested, if any. Based on these, you should collect data to analyze. What's your objective?
Objective:


*

*To study the factors affecting the  perceptions of users about overall improvement of the leave management process by automation of the different processes. 

*To study the impact of demographic characteristics on these factors  


If these are your objectives, then you can use the following techniques:


*

*Factor Analysis to identify the factors (for objective -1)

*Three ANOVA tests corresponding to each demographic characteristic (age, gender, dept) [Note: for categories of input variable like gender (M/F), t- test can also be used, the results would be the same.
A: Try Fishers exact test to assess the association between the dependent and independent variables; whenever there is an association go for Contingency coefficient to see the strength of association between the variables. Go for correlation coefficient to see the relationship between demographic variables and perception improvement. But you have to do scoring for perception improvement
