In: Statistics and Probability
Never forget that even small effects can be statistically significant if the samples are large. To illustrate this fact, consider a sample of 128 small businesses. During a three-year period, 12 of the 87 headed by men and 7 of the 41 headed by women failed.
(a) Find the proportions of failures for businesses headed by
women and businesses headed by men. These sample proportions are
quite close to each other. Give the P-value for the test of the
hypothesis that the same proportion of women's and men's businesses
fail. (Use the two-sided alternative). What can we conclude (Use
α=0.05α=0.05)?
The P-value was so we conclude that
Choose a conclusion. The test showed strong evidence of a
significant difference. The test showed no significant
difference.
(b) Now suppose that the same sample proportion came from a
sample 30 times as large. That is, 210 out of 1230 businesses
headed by women and 360 out of 2610 businesses headed by men fail.
Verify that the proportions of failures are exactly the same as in
part (a). Repeat the test for the new data. What can we
conclude?
The P-value was _______ conclude that
Choose a conclusion:
The test showed strong evidence of a significant difference.
The test showed no significant difference.
(c) It is wise to use a confidence interval to estimate the size
of an effect rather than just giving a P-value. Give 95% confidence
intervals for the difference between proportions of men's and
women's businesses (men minus women) that fail for the settings of
both (a) and (b). (Be sure to check that the conditions are met. If
the conditions aren't met for one of the intervals, use the same
type of interval for both)
Interval for smaller samples: ___ to ____
Interval for larger samples: _________to _________
What is the effect of larger samples on the confidence
interval?
Choose an effect:
. The confidence interval is unchanged.
The confidence interval's margin of error is reduced.
The confidence interval's margin of error is increased.
a)
x= | 12 | 7 |
p̂=x/n= | 0.1224 | 0.1707 |
n = | 98 | 41 |
estimated prop. diff =p̂1-p̂2 = | -0.0483 | |
pooled prop p̂ =(x1+x2)/(n1+n2)= | 0.13669 | |
std error Se=√(p̂1*(1-p̂1)*(1/n1+1/n2) = | 0.0639 | |
test stat z=(p̂1-p̂2)/Se = | -0.76 | |
P value = | 0.4498 | (from excel:2*normsdist(-0.76) |
The test showed no significant difference.
b)
x= | 360 | 210 |
p̂=x/n= | 0.1379 | 0.1707 |
n = | 2610 | 1230 |
estimated prop. diff =p̂1-p̂2 = | -0.0328 | |
pooled prop p̂ =(x1+x2)/(n1+n2)= | 0.14844 | |
std error Se=√(p̂1*(1-p̂1)*(1/n1+1/n2) = | 0.0123 | |
test stat z=(p̂1-p̂2)/Se = | -2.67 | |
P value = | 0.0076 | (from excel:2*normsdist(-2.67) |
since p value <0.05
The test showed strong evidence of a significant difference.
c)
since sample success is less than 10 for one sample. adding 1 in success and 2 in sample size
for smaller sample:
x= | 13 | 8 |
p̂=x/n= | 0.1461 | 0.1860 |
n = | 89 | 43 |
estimated diff. in proportion=p̂1-p̂2= | -0.0400 | |
Se =√(p̂1*(1-p̂1)/n1+p̂2*(1-p̂2)/n2) = | 0.0702 | |
for 95 % CI value of z= | 1.960 | |
margin of error E=z*std error = | 0.137522 | |
lower bound=(p̂1-p̂2)-E= | -0.1775 | |
Upper bound=(p̂1-p̂2)+E= | 0.0975 | |
from above 95% confidence interval for difference in population proportion =(-0.1775 ,0.0975) |
Interval for larger samples:
x= | 361 | 211 |
p̂=x/n= | 0.1382 | 0.1713 |
n = | 2612 | 1232 |
estimated diff. in proportion=p̂1-p̂2= | -0.0331 | |
Se =√(p̂1*(1-p̂1)/n1+p̂2*(1-p̂2)/n2) = | 0.0127 | |
for 95 % CI value of z= | 1.960 | |
margin of error E=z*std error = | 0.024854 | |
lower bound=(p̂1-p̂2)-E= | -0.0579 | |
Upper bound=(p̂1-p̂2)+E= | -0.0082 | |
95% confidence interval for difference in population proportion =(-0.0579 , -0.0082) |
The confidence interval's margin of error is reduced.