I have two sets of surveys- (before and after) that were given to a group of students before and after a lecture. I can match before and after responses for each student. The survey consisted of Multiple Choice Questions, Likert Scales and Yes/No Questions. I am having difficulty deciding which SPSS tests to run. Any suggestions would be greatly appreciated.
Dear Alanna Grealy ,
Firstly, you should match the data into a single data file.
Secondly, it is a question of how you wish to combine your data into the variables you are going to use to analyse it. If you are planning to treat all your questions individually then you need to be led by your research questions and not look at every comparison as there will be a risk of Type I errors. If you wish to create combined (scale) variables the best technique is principal component analysis.
Thirdly, analysis techniques you could use are a paired t-test, a Wilcoxon signed rank test or a chi-squared test.
I agree with Peter Samuels that a lot depends on whether you want to analyze the data item-by-item or create multi-item scales. I disagree, however, with his recommendation to use principal component analysis, because the error assumptions are more appropriate for Principal Factor Analysis (which should be accompanied by an oblique rotation to allow for correlated factors).
If you do construct scales, then the simplest way to test the difference is with a paired-comparisons t-Test.
I see you have to develop scoring system in which you find out the level of all participants as sure you will not analyse it item by item , then you compare te levels before and after using the paired t test .
A t-test could be done to compare the average of mean and see if there is any difference in the mean before and after the control variable has been added.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS