Multiple regression

Multiple Regression is one of the most widely used statistical tools, the reason for this is convincing: it is remarkably efficient for answering questions involving many variables.

Application

This technique is applicable to many subject fields, from agriculture to medicine.

Example

The data in this example is obtained from a study of the relationship between the amount of rainfall on the yield of corn in the USA.

The raw data was presented in the following format:

Year Rainfall (in/year) Yield (bu/acre)
1890
1891



1927
 
9.6
12.9



10.4
 
24.5
33.7



32.6
 

Summary of statistical findings

There is strong evidence that the relationship between yield and rainfall is not linear. The term in the regression model which describes the curvature is significant (p=0.014). A graphical representation of corn yield vs rainfall can be seen as follows:

Yearly corn yield vs rainfall

 

 

Contact information
Mr Coos Bosma
Statistical Consultant and Analyst
Tel: 27 41 504 9902
coos.bosma@mandela.ac.za