Biases of study design
To participate in this randomized controlled clinical trial women had to agree to be randomly assigned to either the cesarean or to the vaginal birth group. This introduces a monumental bias into the study that cannot be overcome in the analysis section. If a woman is neutral about mode of delivery, she is not committed to a vaginal birth and may not be psychologically prepared for the demands thereof. This biases the study toward a Type I error (rejecting the null-hypothesis although it is true). A more appropriate design would be a cohort study where women self-select intervention. This is effectually a naturalistic experiment. While the cohort study is not randomized, the authors of the TBT found randomization did not afford the anticipated benefits since the trial was analyzed by intent to treat, but “women randomized to the planned vaginal birth group had a high rate of cesarean delivery, which would have reduced our ability of finding an association between cesarean delivery and adverse outcomes, if one existed” (Hannah, et al. 2002:1830)
| « Previous: Introduction | » Next: Recruitment |