Biostats Flashcards
1
Q
What is power?
A
Probability of detecting a difference between groups when a difference exists
Power = 1-beta
2
Q
What affects power?
A
sample size
Beta error
1-tailed > 2-tailed test
standard deviation (the smaller SD, higher the power)
3
Q
What can contribute to obtaining a type 1 error?
A
Type 1 error = rejecting a true null hypothesis
Reasons: underpowered study, confounding factors, selection bias, multiple testing of same hypothesis
4
Q
How can one control for confounding factors?
A
Randomization Multi-variate regression analysis Double blinding Matching cases (same sex/age/etc) Stratification
5
Q
Name 5 different types of bias possible in a study?
A
Selection bias Recall bias Confounding bias Information bias Publication bias Lead-time bias Attrition bias Measurement bias
