Skip to main content icon/video/no-internet

Inferential statistics allows researchers to examine whether there is enough evidence in favor or against the claims about a sample that is drawn from a population. The statistical process used for supporting or rejecting claims on the basis of sample data is referred to as significance testing. This entry starts with the explanation of the concepts central to significance testing such as null and alternative hypothesis, and one-directional and two-directional tests. Second, this entry describes what statistical significance means and the steps researchers should follow in the process of significance testing. Finally, this entry outlines statistical errors unique to significance testing.

Null and Alternative Hypothesis

In order to conduct a significance test, researchers first state predictions about the population as a null and alternative hypothesis. The starting point for every significance test is stating a null hypothesis (H0) that simply means there are no effects or differences in the population. For instance, if a researcher wants to examine how gender affects communication anxiety, the null hypothesis would be stated as “H0: There is no difference between men and women and their level of communication anxiety.” In any significance test, researchers want to reject the null hypothesis, and find evidence for the alternative hypothesis. An alternative hypothesis (Ha) is the prediction that there is an effect or difference in the population that does not occur due to chance. The alternative hypothesis of the preceding null hypothesis would be stated as “Ha: There is a difference between men and women and their level of communication anxiety.”

Another example for null and alternative hypothesis generation can be constructed using nonverbal immediacy and attentiveness. Suppose that a communication researcher wants to examine the relationship between nonverbal immediacy of teachers and student attentiveness in the college classroom. The researcher would formulate the null hypothesis as “H0: There is no relationship between nonverbal immediacy and attentiveness.” The alternative hypothesis for which the researcher wants to find statistical evidence would be stated as “Ha: There is a relationship between nonverbal immediacy and attentiveness.”

One-Tailed and Two-Tailed Tests

One important consideration while formulating a null and alternative hypothesis is to determine the direction of the difference or relationship among variables. If a researcher uses a hypothesis or research question that not only tests for the presence of a difference or a relationship between variables, but also specifies the direction of this difference or relationship, a directional or one-tailed test would be used. For instance, a researcher would use a one-tailed probability as the direction of the relationship is specified if he or she wants to test the following hypothesis: “Females will not experience significantly more communication apprehension than males.” On the other hand, if a researcher uses a hypothesis or a research question that simply states a difference or relationship between variables without predicting the direction of this difference or relationship, nondirectional or two-tailed probability would be used. Using the preceding example, a researcher would use a two-tailed test to examine the following hypothesis: “Females and males differ in communication apprehension.”

...

  • Loading...
locked icon

Sign in to access this content

Get a 30 day FREE TRIAL

  • Watch videos from a variety of sources bringing classroom topics to life
  • Read modern, diverse business cases
  • Explore hundreds of books and reference titles

Sage Recommends

We found other relevant content for you on other Sage platforms.

Loading