Mod 7

Question 1

What are the parameters of the sampling distribution of a difference in two population proportions?

Answer 1

center/mean: p1 - p2
spread/standard deviation: square root of p1(1-p1)/n1 + p2(1-p2)/n2

Question 2

What are the conditions under which the shape is approximately normal?

Answer 2

n1p1, n1-p1, n2p2, n2-p2 are at least 10

Question 3

When do you implement a large-sample confidence interval for a difference in population proportions?

Answer 3

• Collected data in reasonable ways
• Sample size in which there are 10 successes and 10 failures in each of the two groups
• n1p1, n1-p1, n2p2, n2-p2 are at least 10

Question 4

What is the formula for confidence interval of a two proportions?

Answer 4

p1 - p2 +/- z* (standard error of p1^ - p2^)

Question 5

What is the interpretation of confidence level?

Answer 5

if you took many samples, we expect or are confident that 95% of them will capture the true proportion/true difference between population groups

common question: Does the difference include 0? Interested in whether the confidence interval includes zero or nondifference is a plausible value

• If confidence interval misses zero, then there is a difference between groups

Question 6

When do you implement a large-sample hypothesis test for the difference in proportions?

Answer 6

np0 and n(1-p0) ≥ 10
- use numbers of successes and failures in samples to check sample size requirement
- At least 10 successes and failures in sample

Question 7

What is the HMCCC method of a large sample hypothesis test for difference in proportions?

Answer 7

Hypotheses in terms of p 1 and p 2 method
H0: p1 = p2 or p1 - p2 = 0
Ha: p1 > p2, p1 < p2, p1 ≠ p2

Method: difference in population proportions

Check that conditions are met:
At least 10 successes and failures in sample

Compute test statistic (here we use a combined estimate of p -p c in the standard error) and p-value

Communicate conclusions:

Compare P-value to alpha
P-value less than alpha = reject the null
P-value greater than alpha = fail to reject the null

State conclusion (reject/fail to reject)
Give conclusion in context of problem