Google

A hypothetical study predicting individuals' willingness to have unprotected sex based on their measured mood (positive versus negative) and self-esteem (high versus low) exemplifies a non-experimental factorial design. This study is conceptualized as a $$2 	imes 2$$ factorial design with mood and self-esteem serving as non-manipulated between-subjects factors. Because neither independent variable is actively manipulated by the researcher, the study cannot establish causality.

Example of a Non-Experimental Factorial Design

The classification of a factorial design as experimental or non-experimental depends entirely on the active manipulation of its independent variables. For example, a study that simply measures participants' pre-existing moods to predict sexual risk-taking is a non-experimental factorial design. In contrast, a similar study by MacDonald and Martineau (2002) is classified as a true experiment because the researchers actively manipulated the participants' moods, which allows for stronger causal inferences.

Comparison of Experimental and Non-Experimental Factorial Designs

Because non-experimental factorial designs rely entirely on measured rather than manipulated variables, researchers must be extremely cautious about inferring causality due to the directionality and third-variable problems. For example, in a non-experimental study linking participants' measured moods to their willingness to have unprotected sex, one cannot definitively claim that mood causes the risk-taking behavior. The observed effect on sexual willingness might instead be caused by an unmeasured third variable that happens to be correlated with the participants' moods.

Example of Causal Limitations in a Non-Experimental Factorial Design

In a 2 x 2 factorial study predicting behavior based on measured mood (positive vs. negative) and self-esteem (high vs. low), why can't the researchers establish a causal relationship?

Learn Before

Related