Exam 12: The T-Test

Nonnormality is sometimes overcome through variable transformation.

A combination of visual inspection and statistical testing should always be used to determine normality.

All t-tests first test for equality of means and then test for equality of variances.

The signed rank test is an independent samples test that examines differences of mean ranks to evaluate whether two samples come from the same population.

A paired t-test violates the assumption of homogeneity.

Recoding continuous variables as categorical variables is discouraged because it results in a loss of information.

Analysts should always examine the robustness of their findings.

One-tailed tests are used most often, unless compelling a priori knowledge exists or it is known that one group cannot have a larger mean than the other.

The term robust is used, generally, to describe the extent to which test conclusions are unaffected by departures from test assumptions.

The Mann-Whitney and Wilcoxon tests are equivalent.

When both variables are categorical, the t-test should be used for hypothesis testing.

The Kolmogorov-Smirnov test is a test of equal variance.

The term parametric statistics refers to tests that make assumptions about the distribution of data.

The paired samples t-test tests the null hypothesis that the mean difference between the before and after test scores is zero.

The t-test has four test assumptions.

The critical values of the t-test are provided by Student's t-test distribution.

When problems of nonnormality cannot be resolved adequately, analysts should consider nonparametric alternatives to the t-test.

A test for the equality of variances is the Levene's test.

The one-sample t-test tests whether the mean of a single variable is different from a prespecified value (norm).

The Mann-Whitney and Wilcoxon tests assign ranks to the testing variable and test whether the sums of ranks differ between the two categories.