Research and Statistical Support

MODULE 5

VII. Correlation, Chi-square, & t-tests

The following covers some of the most commonly used SAS procedures with which you can run some basic statistical analyses. Use the File, Import Data... to import the Example Data 1 file using the Import Wizard with SPSS File (*.sav) source and member name example1 as was done previously.

1. PROC CORR

In SAS, we use PROC CORR to examine bi-variate relationships.

```PROC CORR DATA=example1;
VAR age recall1 recall2;
RUN;```

If we had missing data in at least one of the variables, then we could use a 'nomiss' optional statement to do listwise deletion. As an example, use the Import Wizard to import the Example Data 2 file using the Excel File (*.xls) source as was done previously. Remember to use 'example2' as the member name to differentiate it from example1.

PROC CORR DATA=example2;
VAR age recall1 recall2;
RUN;

PROC CORR DATA=example2 NOMISS;
VAR age recall1 recall2;
RUN;

2. Chi-square tests using PROC FREQ

We can use proc freq to examine class standing (cl_st, where 1 = freshman, 2 = sophomore, 3 = junior, & 4 = senior) by Sex (males coded 1, females coded 0).  Using the chi2 option we can request a chi-square test that tests if these two variables are independent, as shown below.

```PROC FREQ DATA=example1;
TABLES cl_st*Sex / CHISQ;
RUN;```

We can also get the Fisher's exact test statistic using the modifier EXACT.

```PROC FREQ DATA=example1;
TABLES cl_st*Sex / CHISQ EXACT ;
RUN;```

3. PROC TTEST

We can use PROC TTEST to examine differences between two groups. Notice in the output, we get t-values for variances assumed equal and variances not assumed equal.

First, we can run an independent groups t-test.

```PROC TTEST DATA=example1;
CLASS candy;
VAR recall1;
RUN;```

Second, we can run a dependent groups t-test (sometimes called: related groups t-test, paired samples t-test, or repeated measures t-test).

```PROC TTEST DATA=example1;
PAIRED recall1*recall2;
RUN;```

Third, we can run the seldom used single sample t-test (here the null supposes the true mean of recall 1 = 15).

```PROC TTEST DATA=example1 HO=15;
VAR recall1;
RUN;```