Frequency and Correlation procedure in SAS

August 13, 2021

Frequency and Correlation procedure in SAS

Hello readers;

In this blog there is demonstration of proc frequency and correlation using SAS on Demand Academics.

Frequency procedure

·        Tables statement to mention variables combination
·        Chi-squared test
·        Fischer Exact T test

About Fischer exact T test

The p-value which we take as 0.05 (by default) means there is 5% of risk that we would reject null hypothesis. Apart from this default procedure, when we have to calculate p-value there are certain tests involved one of them is Fischer Exact T-test.

This test is used when we have two nominal variables and we have to check whether the proportion of one variable is different of that other variable.

Take an example that there are two group one is control and other one s treatment group and in rate wise manner you got records that due to certain drug the treatment group showed lower blood sugar levels than control group.

So, the values of blood sugar levels are significantly different or not in those two groups this hypothesis is validated in Fischer Exact T test.

Cars dataset have three origin categories Asia Europe and USA so three frequency tables and plots will be generated in results tab.

Only Asia’s frequency table and plots are shown for reference purpose.

RESULTS

Fischer test using proc frequency in SAS

DATA

Correlation in SAS

HOEFFDING: tests whether the two random variables are dependent or independent of each other (distance statistic).
KENDALL: nonparametric test used to determine the strength and direction of association between variables.
PEARSON: measures the strength of linear association between variables.
SPEARMANN: used in place of Pearson in non-parametric cases.