Description of a Study

So far, we have examined actual studies in which all interactions turned out to be nonsignificant. Now, we want to consider what researchers typically do when a first-order interaction null hypothesis is rejected. To accomplish this task, let's take a look at a recent investigation that focused on college students, dating status, and "temptation." In the study we want to consider, 98 heterosexual undergraduate students (49 males and 49 females) initially filled out a questionnaire describing their current dating status. Based on their responses, each student was classified into one of two groups: "exclusive daters" and "nonexclusive daters." (To be in the first group, a student had to be in a dating relationship--or have just come from one--that had lasted for at least three months with neither party dating anyone else). About half of the full group fell into each of these dating categories, and there were about as many males as females in each group.

A week later, each of the 98 students was asked to assume the role of one of the three characters involved in a brief story. In this story, the student was described as eating dinner with a good friend at a restaurant one night. The story went on to say that an attractive, opposite-sex stranger first begins to flirt and then sends the server over with a message indicating his (or her) interest. A11 98 students were given the same story to read, except for one important little detail. In the story given to a random half of the students, the attractive stranger was said to be flirting with "you"; in the story given to the remaining students, it was "your friend" instead.

Immediately after reading one or the other of the two stories, each student was asked to fill out a second questionnaire. Three of the items in this questionnaire focused on the respondent's anticipated satisfaction with the attractive stranger (called the"target") as a romantic partner. A 1-to-5 Likert scale was used by the students when responding to the questionnaire items, with these scores on the three key items added together to form a composite that signified "romantic interest in the target."

In this study, a three-way ANOVA was used to analyze the students' composite scores that indicated their anticipated romantic interest in the story's attractive stranger. Two of the independent variables in this ANOVA were assigned factors because they dealt with characteristics of the 98 students who served as subjects. These factors were Gender (male, female) and Dating Status (exclusive, nonexclusive). The third independent variable was an active factor, for the researchers randomly assigned subjects to this factor's levels. This factor was called Threat/Opportunity, and its levels were labeled High and Low. Students were considered to be in the "high threat/opportunity" condition of this experiment if they were given the version of the story in which they were the recipient of the attractive stranger's flirtatious behavior and message; in contrast, students were considered to be in the "low threat/opportunity" condition if their story indicated that the attractive stranger was after their friend.

In Excerpt 15.12 [not shown here], we see that the three-way ANOVA yielded a two-way interaction between Dating Status andThreat/Opportunity. As you will note, the researchers indicate that the significant main effect of Dating status was "qualified by" the significant two-way interaction between Dating Status and Threat/Opportunity. In saying this, they were indicating that the significant main effect should not be interpreted directly. Instead, they focused on the significant interaction they obtained and probed it by conducting tests of simple main effects.

Copyright © 2012

Schuyler W. Huck
All rights reserved.

| Book Info | Author Info |

Site URL:

Top | Site Map
Site Design: John W. Taylor V