Overview (T-TEST command)

T-TEST compares sample means by calculating Student’s t and displays the two-tailed probability of the difference between the means. Statistics are available for one-sample (tested against a specified value), independent samples (different groups of cases), or paired samples (different variables). Other procedures that compare group means are ANOVA, ONEWAY, UNIANOVA, GLM, and MANOVA (GLM and MANOVA are available in SPSS® Statistics Standard Edition or the Advanced Statistics Option).

Options

Statistics. There are no optional statistics. All statistics available are displayed by default.

Basic Specification

The basic specification depends on whether you want a one-sample test, an independent-samples test or a paired-samples test. For all types of tests, T-TEST displays Student’s t, degrees of freedom, and two-tailed probabilities, as well as the mean, standard deviation, standard error, and count for each group or variable.

  • To request a one-sample test, use the TESTVAL and VARIABLES subcommands. The output includes a One-Sample Statistics table showing univariate statistics and a One-Sample Test table showing the test value, the difference between the sample mean and the test value, and the two-tailed probability level.
  • To request an independent-samples test, use the GROUPS and VARIABLES subcommands. The output includes a Group Statistics table showing summary statistics by group for each dependent variable and an Independent-Samples Test table showing both pooled- and separate-variance estimates, along with the F value used to test homogeneity of variance and its probability. The two-tailed probability is displayed for the t value.
  • To request a paired-samples test, use the PAIRS subcommand. The output includes a Paired Statistics table showing univariate statistics by pairs, a Paired Samples Correlations table showing correlation coefficients and two-tailed probability level for a test of the coefficient for each pair, and a Paired Samples Test table showing the paired differences between the means and two-tailed probability levels for a test of the differences.

Subcommand Order

Subcommands can be named in any order.

Operations

  • If a variable specified on GROUPS is a long string, only the short-string portion is used to identify groups in the analysis.
  • Probability levels are two-tailed. To obtain the one-tailed probability, divide the two-tailed probability by 2.

Limitations

  • Maximum of one TESTVAL and one VARIABLES subcommand per one-sample t test.
  • Maximum of one GROUPS and one VARIABLES subcommand per independent-samples t test.