Biostats Intro

Question 1

Statistical Inference

Answer 1

Don’t actually study pops, study samples and make inferences about pops

Question 2

Sampling/Random Error

Answer 2

Extent to which sample does not precisely represent the population

Question 3

Standard Error (what it is and eq)

Answer 3

Standard dev of the sample, a general measure of imprecision. SE = SD/sqrt(N)

Question 4

2 Components that Affect Size of P-Value

Answer 4

Sample size and effect size

Question 5

SD scores and percentages

Answer 5

68% b/w 1 SD, 95% b/w 2, 99% b/w 3

Question 6

Z Score (& 3 critical values, & eq)

Answer 6

Standardized way of saying how many SD a datum is from mean
z = 1.96 has exactly 95% of curve below it (p = 0.05)
2.58 has 99% below
3.29 has 99.9% below
z (or actually t, but kinda interchangeable) = M/SE

Question 7

Type 1 vs. 2 Error and most common reasons

Answer 7

Rejected null when true, bad study methods

Accepted null when false, more common and small sample size usually

Question 8

Power Analysis

Answer 8

Specify amounts of Type 1 and 2 error we’re willing to take, as well as size of effect to detect and variance we expect, to tell us the sample size we need to accomplish the above

Question 9

CI Definition

Answer 9

Range within which the true population effect lies assuming a certain degree of assurance. CI containing the null value suggests no statistically significant difference b/w the 2 (so even if RR = 1.5, if CI includes 1.0 it isn’t statistically significant)

Question 10

CI Equation

Answer 10

CI = M +/- CV*SE (CV is the z score critical value, so like 1.96 for 95% CI)