Statistical Implications of Large Databases
Since there are only one or two
large datasets of EMS patients in any state, there
are important statistical implications. As more questions
are asked, it is increasingly likely that a result
will be positive based upon chance alone and not a
real difference. Since there is not a second dataset
with which to validate the results, it becomes impossible
to tell which positive results are meaningful and
which are statistical flukes. When EMS researchers
conduct studies involving large state-based data sets,
they will need to validate those studies by repeating
them in other states or at a different time.
|