Chapter Contents |
Previous |
Next |

The TTEST Procedure |

For each variable in the analysis, the TTEST procedure displays the following summary statistics for each group:

- the name of the dependent variable
- the levels of the classification variable
- N, the number of nonmissing values
- Lower CL Mean, the lower confidence bound for the mean
- the Mean or average
- Upper CL Mean, the upper confidence bound for the mean
- Lower CL Std Dev, the lower confidence bound for the standard deviation
- Std Dev, the standard deviation
- Upper CL Std Dev, the upper confidence bound for the standard deviation
- Std Err, the standard error of the mean
- the Minimum value, if the line size allows
- the Maximum value, if the line size allows
- upper and lower UMPU confidence bounds for the standard deviation, displayed if the CI=UMPU option is specified in the PROC TTEST statement

Next, the results of several *t* tests are given. For
one-sample and paired observations *t* tests, the TTEST
procedure displays

- t Value, the
*t*statistic for testing the null hypothesis that the mean of the group is zero - DF, the degrees of freedom
- Pr > |t|, the probability of a greater
absolute value of
*t*under the null hypothesis. This is the two-tailed significance probability. For a one-tailed test, halve this probability.

For two-sample *t* tests, the TTEST procedure displays
all the items in the following list.
You need to decide whether equal or unequal variances are
appropriate for your data.

- Under the assumption of unequal variances, the TTEST
procedure displays results using Satterthwaite's method.
If the COCHRAN option is specified, the results for the
Cochran and Cox approximation are also displayed.
- -
- t Value, an approximate
*t*statistic for testing the null hypothesis that the means of the two groups are equal - -
- DF, the approximate degrees of freedom
- -
- Pr > |t|, the probability of a greater
absolute value of
*t*under the null hypothesis. This is the two-tailed significance probability. For a one-tailed test, halve this probability.

- Under the assumption of equal variances, the TTEST procedure displays
results obtained by pooling the group variances.
- -
- t Value, the
*t*statistic for testing the null hypothesis that the means of the two groups are equal - -
- DF, the degrees of freedom
- -
- Pr > |t|, the probability of a greater
absolute value of
*t*under the null hypothesis. This is the two-tailed significance probability. For a one-tailed test, halve this probability.

- PROC TTEST then gives the results of the test of equality of variances:
- -
- the
*F'*(folded) statistic (see the "The Folded Form F Statistic" section) - -
- Num DF and Den DF, the numerator and denominator degrees of freedom in each group
- -
- Pr > F, the probability of a greater
*F'*value. This is the two-tailed significance probability.

Chapter Contents |
Previous |
Next |
Top |

Copyright © 1999 by SAS Institute Inc., Cary, NC, USA. All rights reserved.