t-test with excel

Statistical test to determine equality of variance

Inorder to do a T-test it is must that two sample groups have equal variance. F-test is done to compare the variance in the two groups.

• Select Tools/ Data Analysis / F-Test Two Sample for Variance.
  

In the F-Test Two Sample for Variance dialog box: For the Input Range for Variable 1, highlight the nine values from $A$2:$A$10.

For the input range for Variable 2, highlight the seven values from $B$2:$B$10. Leave the other items at their default selections. This dialog box is shown below. Click OK.

The following results are produced by Excel:

Probability p=0.43

This is a one-tail p-value associated with the test for equality of variance. Generally, if this value is less than 0.05 you assume that the variances are NOT equal.

A. If the variances are assumed to NOT be equal, proceed with the t-test that assumes non-equal variances.

B. If the variances are assumed to be equal, proceed with the t-test that assumes equal variances.

T-test

The t-test in Excel is similar for both the equal and unequal variances case – the main difference is which version you select from the menu. Assuming that you have selected the unequal version of the two-sample t-test – this is how you proceed.

Select Tools/ Data Analysis/ t-Test: Two Sample assuming equal Variances

For the Input Range for Variable 1, highlight the nine values from $A$2:$A$10. For the input range for Variable 2, highlight the nine values from $B$2:$B$10. Leave the other items at their default selections. This dialog box is shown below. Click OK.

The following output is created:

The two tailed calculated t-statistic is 0.875 and the highlighted p-value for this test is p=0.39 Since the p-value is more than 0.05, this provides evidence to accept the null hypothesis of equal means.

 

>>Continue reading: Anova with Excel, Descriptive statics