Biostats Flashcards
1
Q
Loss to follow up bias?
A
Selection bias - more specifically attrition bias
2
Q
Misclassify data 2/2 individual differences in interpretation of data
A
Observer bias
3
Q
Factorial design study
A
Randomization looking at multiple goals
4
Q
Effect modification
A
- like confounding, except stratification shows correlation. Confounding does not
5
Q
Typically there is a trade off between sens and spec
A
as one goes up, other does down. OBVI
6
Q
Effects of increasing sample size on power and CI?
A
increase power, narrow CI
7
Q
If confidence interval includes null (or crosses 1.0), what can we say about p value?
A
greater than 0.05, statistically insig
8
Q
Attributable risk calculation?
A
(risk in exposed-risk in unexposed)/risk in exposed
9
Q
How will decreasing cutoff point for a test (so as to create more false positives) affect PPV?
A
decrease
