Explanations for Improvement in Both Experimental and Control Groups

Abstract

A true experimental design with random assignment to groups protects against false causal inferences that could be made when both the treatment and control groups change because of factors such as testing effects, reactivity, contamination, maturation, history, and other measurement effects. The occurrence of these phenomena, however, provides interesting information about factors affecting health care attitudes, knowledge, and behavior change, which can interfere with a nursing study's ability to demonstrate an experimental effect. In this article, we discuss these design threats, illustrate them with examples from recent health research, and suggest strategies for decreasing them in clinical nursing studies.