Several limitations of this study could provide alternative
explanations for the results. The study was not blinded, leaving
the possibility for investigator bias. Given that the same
investigators analyzed workstations and provided the training,
the intervention group scores were vulnerable to inflation.
The study also is vulnerable to participant bias, in that
participants may have behaved differently because they knew
they were being studied and, in particular, what was being
evaluated. In other words, members of the education group
may have been more likely to implement changes and attend
to their positioning, accounting for some of their improvement.
Educational interventions, therefore, may not be as
effective as the results of this study might indicate. Finally,
the participants in each cell may have spoken with members
of other cells outside of work about the interventions they
were receiving, resulting in some misclassification. This
possibility could account for the improvements seen in the
education group.