Need some help with the statistics in my public health project?

Question

I have some data from questionnaires asking people their opinions on the national health service in the UK.

I have asked them all for basic details such as age, s*x and number of hospital visits.

I then asked people to give a score from 1-5 on how big they think various problems are.

I want to use this info to initially find which problems people thought were the biggest and see if the score difference are significant and then to see if different groups of people had significantly different opinions by comparing mean scores for different problems between groups.

I know that I need to compare the means for each problem but don't know which tests to use particularly when comparing multiple groups

My knowledge of stats is basic and I'm so confused about what tests to use.

I'm trying to use the stats package SPSS but its not really giving me any clues!

Answer 1

You will want to do a t test for independent means when comparing two variables.  If you are comparing more than that you will want to use a anova.  You could do a chisquare if you thought you should be able to accurately predict what an answer should be (if answers were expected to be randomly distributed, but I doubt that is the case).  

You might also want to consider grouping your responses into sets.  This often sets off data and makes differences more perceptible.  For example, instead of comparing 1,2,3,4,5 as potential responses include 1&2,3,4&5 this clustering of responses will likely find more statistical significance, especially with smaller samples.  

I hope that helps.  Feel free to email me with questions.  Also, the help section of SPSS if indeed helpful although not necessarily in finding what exactly you want.