Ch 9.3 Sallis Flashcards
What is the paired samples t-test?
In the paired samples t-test, the observations in one sample are directly related to the observations in the other sample, for example, the amount of sick leave from one year to the next for the same group of people at the same company. The sample is the same, measured at two time periods. The paired samples t-test is also called the repeated measures t-test or the dependent samples t-test.
The paired samples t-test is for testing mean differences in one population measured for the same thing at two time points. Imagine testing whether the labor costs for a company were equal for September one year μ1 compared to September the next year μ2. The hypotheses are:
H0: μ1 = μ2
H1: μ1 =/ μ2 (not equal to)
What is the t-test?
T-tests determine whether there is a statistically significant difference between the means of two groups or between the mean of one group and a specified test value. Sometimes they are referred to as the Student’s t-test, which is the pseudonym for the person who developed them.
What is the nonparametric Wilcoxon Signed Ranks Test?
If, for example, the continuous variables in the paired samples t-test were not normally distributed, then the Wilcoxon signed ranks test would be appropriate. It tests for median differences between paired samples or repeated measures by converting observations to ranks.
What is the independent samples t-test?
The independent samples t-test determines whether there is a statistically significant difference between the means of two unrelated samples. The samples are assumed to be mutually exclusive, meaning that no case is present in both groups. A typical example is comparing gender (assuming two categories) for a continuous variable, like the amount of sick leave for men and women.
What is the nonparametric Mann-Whitney U Test?
The nonparametric alternative to the independent samples t-test is the Mann– Whitney U test. In the previous example, we tested for differences in the continuous variable, revenue, for two hotel types, boutique and chain. We checked the distribu- tion of the Revenue variable, and found a skewness of 2.080, and a kurtosis of 4.924, which together indicate an abnormal distribution. The accepted cutoff indicating normality is an absolute value of 1. In this circumstance, it is advisable to test for group differences using a statistical method that is not based on having a normal distribution.
