Meta-analysis
This is where you pool similar studies to see if the combined
numbers can give you a conclusion and carry enough “weight”
to prove a point. Although this may be good in theory, a meta-analysis
may not give different studies different weight in their analysis.
That is to say that poor studies have just as much validity as
well-done studies. Obviously, this can be a problem. Although
useful information can sometimes be gleaned from these studies,
I approach them cautiously. If anything, conclusions reached by
this method should prompt a hypothesis that is then tested in
a prospective study if possible (in my ideal world).
Bias
Bias is the introduction of an internal or external influence
which changes the study results. A plethora of bias possibilities
exist, and you can name them for the area from which they come.
If a study is not blinded and the investigators have a tendency
to make the study subjects report better outcomes, then that could
be Investigator Bias.
If a drug company is sponsoring the study and has employees whose
pay depends on a good outcome participating as subjects, then
this could be Financial Bias.
If a country has banked decades worth of time and millions in
grants because it believes what it is studying will be earth changing,
then this could be Political Bias.
You get the idea….
