Multiple Regression Analysis
What is a multiple regression analysis?
Multiple regression analysis examines whether, based on correlation of more than one independent variables with dependent variable, is a predictive relation and can be used to test a hypothesis. The multiple regression analysis uses continue or ordinal data, but next to that it is able to take into account one or more categoric variables as independent variable.
When do you use the multiple regression analysis?
You will use the multiple regression analysis to test whether one or more independent variables influence a dependent variable and if this effect is positief or negative. It's also possible to test interaction effects.
Example of a multiple regression analysis
You want to examine whether intelligence (independent) and the hours of study (independent) have any influence on the exam grade (dependent). For example you will test this hypothesis: the higher the intelligence and hours of study, the higher the exam grade (or: intelligence and hours of study have a positive influence on the exam grade). Next to that you could test the interaction effect between intelligence and hours of study by standardising the two and multiply them with each other and then use them as a new variable.
What's important when considering the multiple regressionanalysis?
The multiple regression analysis is a strong statistical test from which you can draw conclusions. The more independent variables you use, the more respondents you will need. Normal is N=50+8m (m being the number of independent variables). We think this number of respondents still is a little small, but it will do in most cases. It's also important to check multicorrelation between the (independent) variables. If the independent variables correlate more than .8, the software has difficulties to distinguish which of the independent variables has influence on the dependent variable. It's possible to include categoric data by using dummy variables. Bear in mind in that case
you will be using a reference group. Again, make sure the dependent variable is continue (?) or ordinal.