In: Math
29) After elections were held, it was desired to estimate the
proportion of voters who regretted that
they did not vote. How many voters must be sampled in order to
estimate the true proportion to
within 2% (e.g., + 0.02) at the 90% confidence level? Assume that
we believe this proportion lies
close to 30%.
A) n = 2017
B) n = 1421
C) n = 2401
D) Cannot determine because no estimate of p or q exists in this
problem.
E) n = 1692
For the given binomial sample size and null-hypothesized value
of p0, determine whether the sample size is large
enough to use the central limit theorem to conduct a test of the
null hypothesis Ho: p = p0.
n = 700, p0 = 0.01
A) No
B) Yes
Let p be the true proportion of voters who regretted that they did not vote
Let
be the sample proportion of voters in a sample of size n who
regretted that they did not vote
The hypothesized value of p is
Using this we can calculate the standard error of proportion as
The significance level for 90% confidence interval is
The critical value of z for 90% confidence interval is
This can be written as
From the standard normal tables we get for z=1.64, P(Z<1.64) = 0.5+0.4495=0.9495 and for z=1.65 we get P(Z<1.65) = 0.5+0.4505=0.9505. Using linear interpolation we can get
.
The 90% confidence interval is
But we want the true proportion to be within 2% (e.g., +- 0.02) at the 90% confidence level
That is we want a 90% confidence interval as
we get
The number of voters must be sampled in order to estimate the true proportion to within 2% (e.g., +- 0.02) at the 90% confidence level is n=1421
ans:B: n=1421
Using the central limit theorem, to approximate binomial
distribution with a normal distribution, n has to be large enough
that
and
have to be at least 5
For n=700 and
and
Since both are greater than 5, we can use central limit theorem and approximate the sampling distribution using a normal distribution
ans: Yes