Get ahead with a $300 test prep scholarship
| Enter to win by Tuesday 9/24
Terms in this set (15)
Three things to assess to determine validity of a study
Bias, Confounding, Random Error
was the study done without bias, confounding, or random error
is the study generalizable to all populations, not just the study population
A systematic error that results in an incorrect (invalid) estimate of the measure of association between exposure and outcome
Who primarily introduces bias?
Investigator or study participants. It is not intentional, rather, something overlooked
Bias away from the null
Spurious association; Assumes an association where one doesn't exists. OR or RR will be <1
Bias toward the null
Mask an association; OR or RR will move toward 1 when should be <1 or >1
List the studies bias can occur in:
All study types: experimental, cohort, and case-control
Where in the study does bias occur?
In the design and conduct of the study
Can bias be fixed in the analysis phase?
No, it can be evaluated for effect on outcome, but not fixed.
Study subjects don't represent target population.
What leads to selection bias?
Using procedures to select study subjects that lead to a different result than one would've gotten from the entire population actually targeted.
What studies does selection bias most likely occur in?
case-control or retrospective cohort because exposure and outcome already occurred
Can selection bias occur in prospective cohort and experimental studies?
Yes, from differential loss to follow-up because this impacts which subjects are selected for analysis.