Other limitations of this study include the fact that
although thesurve y was mailed to a random sample
within the community, the respondents were still
self-selected, thereby subjecting the sample to bias. It
is possible that the persons responding to our survey
were in better health and were impacted by other
unidentified factors than those persons who chose
not to respond. It is for this reason that we note our
findings may not be generalized to the population
at large.