Quasi-Experimental Methods

Fadda, R.

Writing - Original Draft Preparation

2018-01-01

---

Abstract

A quasi-experiment is a research method in which the experimenter uses preexisting differences between participants rather than random assignment to allocate the participants to different conditions of the study. An example of a quasi-experiment is a comparison between the behavior of children at high risk and at low risk of autism spectrum disorder. The first condition corresponds to children who have an older brother or sister who has received a clinical diagnosis of an autism spectrum disorder from expert clinicians. The second condition corresponds to children who do not have an older brother or sister with autism spectrum disorder. For objective reasons, the researcher cannot assign the two conditions to all the participants. Sometimes, researchers cannot randomly assign the participants to the experimental conditions for ethical reasons. For example, let us assume a researcher is aiming to evaluate the effects of malnutrition on child development. To achieve this aim, one group of children will receive the recommended amount of food per day, while the other group will not receive enough food. The researcher then measures the children’s growth curves in the two groups. However, in this case, the researcher must face the dilemma of deciding how to select the group that will receive the recommended amount of food versus the malnourished group. Naturally, no researcher has the right to take on such a responsibility. Therefore, the best way to solve this ethical dilemma is to apply a quasi-experimental design in which naturally fooddeprived children are compared with children living in different conditions. In other cases, researchers cannot assign participants to an experimental condition for practical reasons. For example, if we want to study the effects of supplementary water consumption at school on children’s cognitive performance, it is difficult to assign pupils within the same class to different conditions. In such a case, it is more practical to consider the whole class a unit, to administer supplementary water to a group of classes, and then to compare the cognitive performance of those classes with those that did not receive supplementary water during the school day. As a consequence of the fact that it is impossible to assign participants randomly to the conditions of a study, any differences in the dependent variables between the groups may be due not just to manipulation of the independent variables but also to several other factors that differ between the groups. For example, if we identify differences between males and females in relation to disgust sensitivity, these differences may be due to neurobiological factors, for example, the activation of different brain circuits in response to disgust elicitors, differences in educational history, or cultural factors. Thus, further research is needed to explore alternative explanations. Usually, these alternative explanations are not determined by logic; rather, researchers consider the most relevant ones according to the field of study. A focus of the study of development is how behavior changes over time. Chronological age is the variable most used to operationalize changes in relation to time. Because experimenters cannot assign chronological age randomly to their participants, quasi-experimental methods offer a unique opportunity to investigate changes in the course of human development. Researchers mainly use quasi-experimental design to explore the following: (a) differences between preselected groups at the same chronological age, (b) comparisons between groups characterized by different chronological ages, and (c) differences in the same group of participants at different chronological ages. This entry provides a description of three main quasi-experimental designs that explore these differences, namely, nonequivalent control group design, cross-sectional design, and longitudinal design. Also considered are the implications of each design for the interpretation of the results as well as threats to internalvalidity and ways to address them. The entry concludes with a summary of best practices to enhance the utility of quasi-experimental designs for studying behavior in the life span of human development.