Microsoft Office Tutorials and References

In Depth Information

**Testing to the T**

The Student t-distribution has nothing to do with students. The originator of

the method was not allowed to use his real name due to his employer’s rules.

Instead, he used the name Student.

Testing to the T

The TTEST function returns the probability that two samples come from

populations that have the same mean. For example, a comparison of the

salaries of accountants and professors in New York City is under way. Are the

salaries, overall (on average), the same for these two groups? Each group is a

separate population, but if the means are the same, then the average salaries

are the same.

Polling all the accountants and professors isn’t possible, so a sample of each

is taken. Twenty-five random members of each group divulge their salary

in the interest of the comparison. Figure 10-1 shows the salaries of the two

groups, as well as the results of the TTEST function.

Student t-testing is generally used when the sample size is 30 items or fewer.

Figure 10-1:

Comparing

salaries.

The TTEST function returns 10.6 percent (0.106437) based on how the

arguments to the function were entered. This percentage says there is a 10.6

percent probability that the mean of the underlying populations are the same.

Said another way, this is the likelihood that the mean of all accountant

salaries in New York City matches the mean of all professor salaries in New York

City. The formula in cell E8 is =TTEST(A2:A26,B2:B26,2,2).