Introduction

This module examines validity and bias issues from the perspective of both the researcher and reader. This module also suggests questions that assess the validity of a research article.

Evaluating the validity and reliability of the information you find is a crucial step in Evidence-Based Practice. You may wonder why you should worry about articles published in reputable peer-reviewed research journals. Isn’t the purpose of the peer-reviewed process to weed out poorly designed studies? That’s true; however, weaknesses do occur in research. Some weaknesses are the result of bias, and some are the result of confounding variables.

Bias is any deviation of results or inferences from the truth because of the way(s) in which the study is conducted. Bias does not necessarily carry an allegation of prejudice, such as the investigators’ desire for particular results. Our job as critical readers is to determine whether the weakness is serious enough to warrant reinterpretation of the study's finding. We need to question whether the study measures what it intended to measure and whether the researchers have accounted for bias and confounding variables.

Validity

There are two types of validity: internal validity and external validity (see: applying research to clinical practice). Internal validity must be established before external validity. Internal validity tries to determine the connection between the independent variable and dependent variable. Internal validity asks, "Did the treatment cause the effect?" Use the questions below to help determine internal validity. Questions that examine internal validity:

• Is there a clear research question?
• Does the research method match the question?
• Is there a control group?
• Is the sample drawn from the population to which the researchers seek to generalize the results?
• Are randomization and blinding employed when appropriate?
• How is the analysis performed, and what statistical methods are used?
• In treatment studies, are all participants accounted for and analyzed according to intention to treat?
• Are the conclusions supported by the study findings?