Means Testing: The Procedures
Before proceeding to conduct a hypothesis test between two sample means, we must determine certain characteristics and parameters about the samples, some of which you will recognize from previous lessons.
Question no. 1: Are the samples related in some way?
If a relation exists between the two variables, the samples are said to be dependent; the data of sample 1 could be paired with data from sample 2. In cases like these, we would be interested in the differences (d ) between the paired values. If the populations are normally distributed and a relationship exists, we would conduct a hypothesis test of two dependent means involving the Student t-distribution.
If no relation exists, then the variables are independent. Other questions must be asked to determine the testing procedures to be used on independent variables.