Mortality

In educational research, mortality (also referred to as experimental mortality or attrition) is a metaphorical term that is used to describe the loss of participants from a study prior to completion. Mortality is among one of eight common threats to internal validity. Threats to validity can be troublesome for research, as these threats limit the conclusions that can be drawn from a study.

Threats to internal validity inhibit researchers’ confidence in reporting that a relationship exists between an independent variable and a dependent variable. To make valid conclusions about the results obtained from a research study, there must be sufficient evidence to substantiate the claim. Mortality threatens this assumption because it compromises the quality and quantity of data garnered from a study.

Mortality is particularly problematic for longitudinal research, as there is an increased potential for reasons a participant may drop out prior to completion (e.g., geographic move, apathy, changes in availability). Studies that employ rigorous or demanding conditions are more susceptible to mortality. For example, studies that require extensive time commitments, are physically or psychologically demanding, or place other stressors on participants may be more likely to experience higher rates of mortality than studies with less demanding conditions.

It is reasonable to assume that mortality is likely to occur across both experimental and nonexperimental research to some degree. Mortality rates become a concern and a threat to internal validity when mortality rates are significantly different between the study’s groups. However, mortality is exclusively a problem not only when differential loss occurs within a study but also when substantially high rates of dropout occur across all study participants. When either of these issues occurs within a study, the results can be dramatically impacted, making it more difficult to conclude that the outcomes obtained were the result of the treatment condition rather than mortality rates.

The underlying problem with differential loss and high mortality rates within a study is that participants who drop out of a study prior to completion, for whatever reason, are characteristically different from participants who complete the study. Differential loss and high mortality rates within a study can lead to relevant biases between groups that may inflate, obscure, or confuse the effects of interest being studied. Additionally, in experimental research, when mortality is systematically related to the study’s design (e.g., treatment conditions are too demanding), it is unclear whether unintentional outcomes were produced by the research design rather than the manipulation of the independent variables.

Although there is no panacea for completely eliminating mortality, the best approach for dealing with this threat to internal validity is to employ randomization whenever possible. Using random assignment presumes that participants who are susceptible to dropout will be equally distributed across both groups.

See also Generalizability; Internal Validity; Random Assignment; Validity; Validity Generalization