prgmSIGTEST

Type 1: reject Ho, Ho true(α)
Type 2: failtorej Ho, Ho false(β)
α and β inv related

Power: reject Ho, Ho false: 1-β
if type 2 more serious, larger alpha
Increase power: larger sample, larger α, larger departure from Ho

Proportion cond:
Random
10% if w/o replacement (N >= 10n)
Large Counts: n*Po>=10, n(1-Po)>=10
z=(p^-Po)/sqrt((Po(1-Po))/n)

Mean Cond:
Rand, 10%
Large Counts:n>=30, or parent dist normal, or graph
t=(x̅ -μo)/(Sx/sqrt(n))
write out df

Steps:
We want to test at α=_ lvl
Ho,Ha, where p or μ is true _ of _
Conditions
Calc
Since p or μ is _ than α, we _
We do(not) have sufficent evidence to conclude _