SPEAR MAN CORRELATION IN SPSS
Spearman correlation is used to find the strength of linkage between ordinal variables.
Just as Pearson correlation coefficient measures the relationship between two scale variables or continuous bivariates, the spearman Correlation is no different as it measures the co-occurence between variables that are ordinal or ranked in nature.
Correlation analyses indicates the strength of relationship as said earlier on a number scale system indicated below.
-1----------------0--------------+1
When the relationship coefficient tends towards:
Positive 1 (+)
We agree there is a positive relationship. Meaning the larger one variable, we also have to increase the presence of the other variable. Or the larger the other.
Towards zero
When the relationship tends towards zero. We say there is little or not relationship between the variables. Depending on how close the coefficient is to zero. When there is no relationship, it also means both variables are does not occur together.
Negative 1 (-1)
When the relationship tends towards the axis of negative one, we say there is a negative relationship that exist between the variables. When there is a negative relationship, it means to get a larger variable, we have to decrease the other variable involved in the analyses. It is better known as smaller A larger B.
Just like the partial correlation that involves measured level data, we use ranks in spearman correlation. It is a non - paracontinous level test.
Questions that can be answered with spearman correlation
The following are typical questions that can be answered with spearman correlation Analysis test.
Does people people with higher level of education are influencial or not.
Let us analyze this question, there is a rank between people that are educated. Starting from the primary school level which is the first school leaving certificate, the secondary school level which is the senior school certificate , the university level which is bachelor's degree, the masters level which is masters degree and the doctorate level which is the doctors degree. My question is to find if all these people with those degrees are influencial to the society or not.
If I find they are influencial, then there is a need to improve in my status of education. But if they are not, then I don't see any need to advise an individual to obtain further education.
I am a fish farmer and I want to find out if my customers are more satisfied with fishes that are of high quality.
Statistical analysis is very good in business process, so as to guide the owner in order not to make mistake that deteriorates the business. I want to find out in my fish farm if my consumers are satisfied with fish of higher quality than that of a medium quality. Of course you know when the fish has high quality it tells you that more money is put into it than that of a medium quality. So if I carry my analysis to find out that they are satisfied with the higher quality one, I just only have to increase my price for the sale of fish. So I could be able to meet up expectation.
Spearman Correlation uses the same method for Pearson correlation but in this case, it converts the variables into ranks.
Null hypothesis for spearman correlation
The null hypothesis for spearman correlation is the same as for person which states that there is no relationship between the variables in consideration.
How to run spearman correlation Analysis test in SPSS
Go-to Analyze menu
Click on correlate
Select bivariate.
Select to the variable box the ordinal variables you wish to find their relationship.
Un-tick Pearson correlation while leaving spearman and kendali tau-b active.
Select flag significant correlation and two tailed.
Output of spearman correlation Analysis
The first result is the kendali tau-b which calculate how many concordant pairs exist in the variables at ordinal level.
The second coefficient is the spearman correlation which is also spearman rho for the calculation of the correlation strength.
Interprete and compare with the general significance value which is 0.05.
Note
if Ho which is the null hypothesis has its significance level > significance level of 0.05, there is no relationship.
If significance level of result < 0.05, there is a significant relationship that exist between variables.
Conclusion
The spearman correlation is a non paracontinous level test. Non paracontinous level test are test that does not assumes that variables approximate a normal distribution.
It is obvious in the analysis of relationship between rank variable that kendali tau-b is selected. This is a test associated with spearman correlation to find sequence of re -occurence or double occurrence. It helps to indicate a better result.