Attrition Bias

When data are collected over two or more points in time, it is common for some participants to drop out of the study prematurely. The attrition of the original sample can occur in longitudinal research as well as in experimental designs that include pretest, posttest, and follow-up data collection. In longitudinal research, which often lasts many years, some participants move between data points and cannot be located. Others, especially older persons, may die or become too incapacitated to continue participation in the study. In clinical treatment studies, there may be barriers to continued participation in the treatment program, such as drug relapse or lack of transportation.

Attrition of the original sample represents a potential threat of bias if those who drop out of the study are systematically different from those who remain in the study. The result is that the remaining sample becomes different from the original sample, resulting in what is [Page 58]known as attrition bias. However, if sample attrition over time is not systematic, meaning that there are no unique characteristics among those who drop out, then there is no attrition bias, even though the sample has decreased in size between waves of data collection. It is important, then, for researchers who collect multiple waves of data to check for attrition bias.

Attrition bias is one of the major threats to multiwave studies, and it can bias the sample in two ways. First, attrition bias can affect the external validity of the study. If some groups of people drop out of the study more frequently than others, the subsequent longitudinal sample no longer resembles the original sample in the study. As a result, the remaining sample is not generalizable to the original population that was sampled. For example, a longitudinal sample examining the grieving process of women following the death of a spouse may fail to retain those participants who have become too distraught to fill out the questionnaire. The nonparticipation of this group may bias the findings of the study toward a minimization of depressive symptomatology as a component of the grieving process. In other words, the composition of the sample changes to the point that the results are no longer generalizable to the original population of widows.

Second, systematic, as opposed to random, attrition can negatively affect the internal validity of the study by altering the correlations among the variables in the study. This problem occurs in longitudinal research because the subsamples that are dropping out of the study at a higher rate are underrepresented in the longitudinal sample, which may lead to correlations between variables that are different from the true correlations in the original sample. For example, the underrepresentation of widows with depressive symptomatology in the second or third wave of a study may alter the correlation between insomnia and length of time since the death of the spouse.

Selective attrition affects the internal validity of experimental research when there are differential dropout rates between the treatment and control groups. In a clinical trial of a depression treatment, if the participants in the treatment group drop out at a higher rate than do the participants of the control group, the results of the study will be biased toward showing artificially successful treatment effects, thus compromising the internal validity of the study. However, if the dropout rates are comparable, the threats to internal validity due to attrition are minimal.

...