Chapter 9: Student t-test

Chapter 9: Student t-test New Vocabulary: Degrees of Freedom for t-tests: a value that equals the total number of subjects included in both of the comparison groups minus 2 (df = total # of subjects - 2) Dependent Samples: paired or related groups of the same sample at a different time Independent Samples: samples that do not have a relationship with one another Student t-test: a test used when you are looking for a difference in the mean value of an interval-level or a ratio-level variable The Student t-Test -​ The student t-test is appropriate only when you are looking for a difference in the mean value of an outcome variable at the interval or ratio level -​ You must have at least 2 samples -​ Samples must be random and independent Statistical Significance -​ You must decide on your alpha, the power, and whether or not you are conducting a one-tailed or two-tailed test Degrees of Freedom for Student t-Tests -​ df = (total sample size of both groups) - 2 -​ There are two t-test values and two p-values and the one you report depends on whether the variances of the two groups are the same -​ If the variances are the same, the Levene’s test F-value will be insignificant, and you should report the t- and p-values assuming equal variances -​ If the variances are not the same, the Levene’s test F-value will be significant, and you should report the t- and p-values where equal variances are not assumed -​ Example: -​ Levene's F-statistic has a p-value > alpha, (1.493 > 0.357) meaning that we fail to reject the null and that the variances are equal among the two groups. To determine a statistically significant difference we must then look at the t- and p-values assuming equal variances. Summary -​ Null = no difference in group means -​ The student t-test is used to determine whether there is a difference in an interval/ratio level outcome/dependent variable in the two sample groups -​ Dependent and independent samples (if they have a relationship together) -​ In order to know what t- and p-value to report you must use Levene’s test for equality variances to find which values you should report