Some relationship occurs with an influencing variable which is neither the dependent or independent variable.
This article will focus on how to carryout partial correlation analysis test in SPSS. If you are familiar with SPSS then this guide is for you.
In the course of this guide, I will define partial correlation, state the example of case where partial correlation can be applied and ways to carry it out in SPSS.
What is partial correlation?
Partial correlation is a type of correlation analysis test that tends to find the relationship between a dependent and independent variables considering a third variable also known as the confounding factor.
According to statistics solutions manual, the third variable is called the hidden factor, suppressor, mediating variable or control variable. It is also a method to correct for the overlap of the moderating variable.
Examples of question that can be answered by using partial correlation
The following questions can be answered using partial correlation. They are;
Does farming in Nigeria improve yeild of crops? What is the relationship between crops planted and the yield when considering for the amount of rainfall.
You can see in this example that having an increased amount of rainfall may increase the plant yeild.
How to carryout partial correlation in SPSS
Before you can carry out partial correlation in SPSS, ensure you find the linearity of relationship using the scatter plot simply drag the variables to each axis to find out if the direction of relationship which is either positive or negative.
The following are the steps on how to carryout partial correlation in SPSS. They are,
Go-to Analyze
Click on correlate
Select partial
Then a dialogue box will be open where you have to specify for the controlling or confounding variable and the variables you wish to find the relationship for.
You can choose to add descriptive statistics which includes the mean and standard deviation.
Remember to select the zero order correlation and exclude cases listwise.
Output of the correlation Analysis
The output of the correlation analysis presents a single table which shows the relationship between each of the variables including the co founding factor and also the variables in consideration.
This will enable the statistician view the variable that are highly related.
Below the relationship is when considering for the confounding factor.
The null hypothesis for correlation analysis states that there is no relationship between the variables in consideration.
You have to check for the level of significance and compare to 0.05 or 5% meaning that five samples out of the total rejects the null hypothesis.
Conclusively
This article has been able to present a brief in partial correlation analysis in SPSS. It is essential to ensure there is multivariate normality of your data before applying to find the relationship between them.
