This week you have explored three different approaches to t tests. By this point, you know that each test has assumptions about the data and type of research questions it can answer. For this Assignme
Skill Builder: Hypothesis Testing for Independent Samples t-test
What You Need to Know
Before reading this skill builder, be sure to review the following concepts:
Type I error and type II error
Alpha
Sample
Population
One-tailed versus two-tailed tests
Outliers
Definitions of categorical variables vs. continuous variables
Random sampling
A hypothesis test is a way for a researcher to obtain empirical evidence in support or in opposition of a position or theory. In dissertation studies, for example, the research question gives rise to an alternative hypothesis, which the student typically wishes to support, and a null hypothesis, a statement of no effect, that the student typically wishes to test and reject. According to the logic of hypothesis testing, the initial focus is on the null hypothesis until the data can cause the researcher to reject it and switch to believing in the alternative.
One hypothesis test that is frequently performed is the t-test for two independent groups. If you wanted to compare a new drug to a placebo, for example, you could use a t-test for two independent groups. One group of patients would receive the drug, and another group would receive the placebo. You could then compare the two groups to see if one group showed more of an improvement in their health. This skill builder focuses on how to decide when to use the independent t-test and also on how to follow the steps of hypothesis testing when conducting a t-test. As you read the material below on the steps for hypothesis testing, you will note that performing the steps in order is important. Recall that hypothesis testing is about making an inference about the population. In the case of the independent samples t-test, researchers are trying to make an inference about whether two groups in the population are different from one another on a continuous dependent variable.
The statistical model for the independent groups t-test includes several requirements of the data. Most basically, if you are using an independent samples t-test, the following should be true:
The independent variable should be categorical with two levels or values (two groups).
The dependent variable should be continuous.
Sex would be an example of a categorical variable that has two levels: male and female. Many researchers are quite liberal in what they consider to be a continuous variable. If calculating a mean for the variable makes sense, most are willing to call the variable continuous. For example, in the social sciences, researchers often use Likert scales to measure variables, which often have 5 response options that range from 1 (strongly disagree) to 5 (strongly agree). Although a variable measured with a Likert scale is not technically continuous, it would make sense to report the average (mean) score across participants for this type of variable. Many researchers, therefore, would use a variable measured with a Likert scale as a dependent variable in a t-test analysis.
There are many examples of research scenarios in which an independent-samples t-test would be used to examine the hypotheses. Here are a few:
Comparing a drug treatment group to a control group on a continuous dependent measure (e.g., severity of symptoms) is common. The independent variable would be type of treatment and the two levels of the independent variable would be “drug treatment” and “control.”
Comparing two kinds of educational programs where a test score is the dependent measure.
Comparing two methods of assembling a complex part with time to completion as a dependent variable.
