Factorial Designs: Interpreting Ambiguous Situations

Dylan was the picture of cool as he sat facing the interviewer, one arm casually thrown over the back of the chair and his long legs stretched out in front of him. Despite his demeanor, he felt on edge. The interviewer had suggested that it might be helpful to have a friend present during the interview and had asked Dylan to choose someone from his class. A research assistant had been sent to bring his friend to the interview. As Dylan looked up, the assistant appeared in the doorway and said that his friend wasn’t coming. “Some friend,” thought Dylan, reacting with anger to what he perceived as an intentional slight.

Not all adolescents would react as Dylan did.

Some would simply assume their friends would have wanted to come if they could but weren’t able to leave the class. What might account for differences among adolescents in the way ambiguous social situations such as this are perceived? And how might such differences affect them and their friendships?

Attachment theory (Bowlby, 1969, 1980) proposes that individuals form working models of relationships, in the form of expectancies, based on their earliest experiences of acceptance or rejection. These expectancies, in the case of rejection, promote a readiness to perceive the actions of others negatively, whereas expectancies of acceptance promote more charitable interpretations. Geraldine Downey, Amy Lebolt, Claudia Rincon, and Antonio Freitas (1998) were interested in whether adolescents who defensively expect to be rejected are more likely than others to be upset by ambiguous situations such as the one described above and whether their reactions actually create problems with friends.

On the basis of a rejection sensitivity questionnaire, the researchers distinguished two groups of early adolescents (mean age = 12.2 years): those who were high in expectations of rejection and those who were low in rejection expectations. Adolescents from these two groups were randomly assigned to either an experimental or a control condition that differed in how ambiguous the rejection was. Those in the experimental condition were told their friend would not join them but were given no explanation for the friend’s refusal. Those in the control condition were also told their friend would not join them but were informed this was because the teacher would not allow their friend to leave. After hearing that their friend would not join them, all adolescents completed measure indicating how much distress they felt.

These investigators used a factorial design that included two variables: rejection sensitivity (high or low) and ambiguity of the rejection (experimental or control). In a factorial design, two or more independent variables, or factors, are completely crossed so that each level of one variable is combined with each level of all the other variables. Factorial designs provide information about the effect of each independent variable alone, called a main effect, and information about the effect of a variable when another variable is present, called an interaction.

An interaction exists when the effect of a variable changes when a second variable is present. We might find, for instance, that neither type of adolescent would experience much distress under the low ambiguity condition (control) but that those with high rejection expectations would experience more distress than those with low expectations when the reason for the refusal was ambiguous (experimental). The existence of an interaction means that we must qualify what we say about a variable. Are adolescents with high rejection expectations more likely than those with low expectations to be distressed by a friend’s refusal to join them? It depends. If they are given no reason for the refusal, then they are. But if they are told why, then they are not.

Downey and her associates found the interaction they had expected (see Figure 1 in the source paper). Adolescents with high rejection expectations experienced more distress only in the ambiguous situation, in which they presumably interpreted the refusal as intentionally hostile. A longitudinal follow-up one year later revealed that differences in rejection sensitivity predicted the amount of conflict adolescents experienced with peers and the extent to which they had difficulties with their teachers. Thus, adolescents who expected to be rejected were more likely to perceive situations as instances of rejection and to react defensively and angrily, leading to interpersonal difficulties with both their friends and their teachers.



G. Downey, A. Lebolt, C. Rincon, & A. L. Freitas. (1998). Rejection sensitivity and children’s interpersonal difficulties. Child Development, 69: 1074–1091.