In: Statistics and Probability
Response |
No Response |
|
Treatment A |
45 |
120 |
Treatment B |
95 |
70 |
Ho: p1 - p2 = 0
Ha: p1 - p2 < 0
sample #1 ----->
first sample size, n1=
165
number of successes, sample 1 = x1=
45
proportion success of sample 1 , p̂1=
x1/n1= 0.2727
sample #2 ----->
second sample size, n2 =
165
number of successes, sample 2 = x2 =
95
proportion success of sample 1 , p̂ 2= x2/n2 =
0.5758
difference in sample proportions, p̂1 - p̂2 =
0.2727 - 0.5758 =
-0.3030
pooled proportion , p = (x1+x2)/(n1+n2)=
0.4242
std error ,SE = =SQRT(p*(1-p)*(1/n1+
1/n2)= 0.05441
Z-statistic = (p̂1 - p̂2)/SE = ( -0.303
/ 0.0544 ) = -5.5691
z-critical value , Z* =
-1.6449 [Excel function =NORMSINV(α)
p-value = 0.000 [Excel
function =NORMSDIST(z)
decision : p-value<α,Reject null hypothesis
there is enough evidence that Treatment B has the better response rate