Hypothesis testing: introduction

Question 1

do null and alternat hypotheses relate to populations or samples?

Answer 1

POPULATIONS

Question 2

define null hyp

Answer 2

results due to chance

Question 3

define alt hyp

Answer 3

results due to treatment

Question 4

are null/alt hyps usually two sided?

Answer 4

yes

Question 5

type I error

• laymans terms
• in terms of null
• in terms of alt

Answer 5

false positive: saying there’s a relationship when there is none.

rejecting the null when the null is true

retaining the alt when the alt is false

Question 6

type II error

• laymans terms
• in terms of null
• in terms of alt

Answer 6

false negative: saying there’s no relationship when there is

retaining the null when null is false

rejecting the alt when the alt is true

Question 7

significance level relates to

Answer 7

type I error (prob of rejecting null when null is true)

Question 8

power relates to

Answer 8

prob of rejecting null when null is false

Question 9

p value: define

Answer 9

the probability that the null is true

Question 10

define: confidence interval (words)

Answer 10

interval around the sample statistic which we are __% sure contains the population parameter

Question 11

define: confidence interval (equation)

Answer 11

= sample statistic +- margin of error

margin of error = t multiplier X SE

Question 12

what p value relates to

• 90% CI
• 95% CI
• 99% CI

Answer 12

0. 1
1. 05
2. 01

Question 13

for infinity df, what’s t multiplier for 95% CI

Answer 13

1.96

Question 14

when it’s not df, what’s t multiplier for 95% CI?

Answer 14

need to look at table