What is factor analysis?
Factor analysis is used to examine underlying factors within variables or items. It searches for patterns and correlation within the different items or variables and creates new factors out of similar variables.
When is factor analysis to be used?
If you want to examine the underlying structure of a group of variables you'll need to use factor analysis. This test is not to be used to test a hypothesis, but merely to reduce the factors of your data.
Example factor analysis
You want to examine the influence of intelligence and hours of study on the grade. Let's assume we have a dataset with 8 items. 4 items cover 'intelligence' (for example the question: I'm more intelligent than my fellow students) and 4 items cover 'hours of study' (for example: I study longer for an exam than my fellow students). By testing the 8 items with a factor analysis, you can find out whether the 4 items for intelligence can be reduced to 1 factor for intelligence and therefore can be grouped. The same goes for the factors on 'hours of study'; after the factor analysis you can establish whether these 4 factors also can be reduced to 1 on 'hours of study' (factor analysis can conduct the grouping for you!). After this analysis the new factors can be checked with Cronbach’s
alpha (explained later).
What's important when considering factor analysis?
You will have to consider the possibility of factor analysis at the start of your project. Too often a lot of questions or items are being included in the research, taking in account the possibility to reduce the items afterwords. Bear in mind that SPSS will factor on correlation, thus despite the fact that two items are very similar (for example two similar types of leadership of a manager) SPSS isn't always capable of discovering them. So, keep in mind that the different factors you would like to examine are to be seen by SPSS; otherwise you will have a useless data-matrix. Watch out for advise like: 'just conduct a factor analysis and use it's outcome...'