confidence intervals and hypothesis tests

Question 1

small samples n<30

degrees of freedom and critical values and SE

Answer 1

less precise sd and se so use student t distribution

as n gets smaller df get less close and critical values an SE increase(less accurate)

Question 2

student t test assumptions

Answer 2

independent observations, sampled from normal distribution

Question 3

equation for t test

Answer 3

sample mean +- t(5%, n-1) x SE

Question 4

confidence intervals with 2 samples basic and assumptions

Answer 4

if pop mean is same u1-u2 is 0

assumes independence, normal distributed, equal variability

Question 5

SE=

Answer 5

SE= (square root) sp^2(1/n1 + 1/n2) where sp^2 is estimate of common variance

Question 6

hypothesis tests basics

h0 h1 p

Answer 6

h0 no effect or difference
h1 some effect or difference
p value probability that the data or data more extreme would occur assuming h0 is true

Question 7

p< or p>

Answer 7

p<0.05 reject h0

p>0.05 fail to reject h0

Question 8

hypothesis test disadvantages and single tailed

Answer 8

much difference if p= 0.049 or 0.051?
5% false
interpretation issues
doesn’t say where the deviation from h0 is clinically or scientifically important
1 tailed doesn’t detect change in the other direction of that ur interested in (eg if new treatment is worse)

Question 9

1 sample, 2sample, paired t test assumptions

Answer 9

independent observations, sampled from normal distribution, population variability equal in both groups *

Question 10

ci advantage

Answer 10

usually better as more info on precision and magnitude of effect

Question 11

statistical and clinical significance differences

Answer 11

statistical looks for high precision whereas clinical looks more for change and magnitude of change