CompTIA DataX DY0-001 (V1) Practice Question

The correct answer identifies differential attrition as the primary threat. In an RCT, randomization is used to create groups that are, on average, balanced in terms of both known and unknown confounding variables. This balance is what allows for causal inference. Differential attrition, which is the unequal loss of participants between groups, can break this balance. In this scenario, participants who experienced adverse side effects disproportionately left the treatment group. The remaining participants in the treatment group are therefore a self-selected subset who could tolerate the medication, and they are no longer comparable to the control group. This introduces selection bias, which is a major threat to the study's internal validity because the observed effect might be due to the differences in the remaining groups rather than the medication itself.

External validity refers to the generalizability of study results to a broader population, which is a different concept from internal validity. While it may be a concern, it is not the most direct threat described in the scenario. The initial randomization is designed to handle confounding variables; the problem here is that the integrity of the randomization was compromised after the study began. Finally, the issue is not primarily about a Type I error resulting from a reduced sample size; the core problem is the systematic bias introduced by the non-random dropout of participants.