HW 8 Chapter 12
I already turned in Chapter 11 homework last week.
a) We need to use the ratio estimate.
b) Here, x=354- total AIDS in the sample, y=807- total number of beds in the
sample while Y = 2501- total number of beds
a) This is a two-cluster design. We need to use:
In this formula, C1 is composed of the cost of traveling to each sample cluster for the
purpose of listing the sampling units, the cost of listing these listing units, the c
a) We have to use the following formula:
Here, N = 20*40 = 800, P = 0.2
n = = 654.69 or 655
b) We need to list all the 800 patients and view the records of 655 patients. Assuming we
need 2.5 hours for listing all patients and 1
a) In the sample of stratum h, we give value 1 to the person having one or
more missing teeth and give value 0 to the person not having missing
teeth. Set this variable whose value is either 1 or 0 to be . Then the
Solution: Since we totally have 120 homes and want to select a 1 in 3 sample,
we choose 2,5,8,11,14,17,20,23,26,29,32,35.if 2 starts the sequence. Then
two homes with lead hazards are selected.
2/40 = 0.05
Survey Sampling HW2
Conclusion: We use stata program to do this problem. According to the results
above, we can see that the proportion of all workers in the plant having a fvc
less than 70% of that expected on the
(a) In this experiment, they were interested in whether the suspected agent increased the
time until tumor onset in experimental animals. (1) The time metric is in days or weeks
(2). Because the tumor is detectable only by micr
(a) In the poisson model, it is not necessary to include the offset because we assume n is fixed here.
Then the distribution of proportion is the same as the distribution of counts.
The summary of model:
glm(formula = tot
(a) Verify that the 18th obs has the highest Cooks distance and deviance residual and dfbeta. Remove the 18 th
obs and check the model.
> index = c(which.max(cooksd),which.max(dev),which.max(dfbeta)
18 18 18
(a) Plot the data, fitted curve along with the 95% point-wise confidence band.
> pred.lin9 = predict(pw.lin9.fit,se.fit=T,interval="prediction",type="response")
In predict.lm(pw.lin9.fit, se.fit = T, interval =
(a) P=1, y = 37.48 + 0.016*x
P=2, y = 8.564 + 0.156*x 0.0001*x^2
P=3, y = 22.69 + 0.01*x + 0.0002*x^2 0.0000001*x^3
P=4, y = 23.37 0.002*x + 0.0002*x^2 0.0000002*x^3 + 0.00000000002*x^4
The following graph shows the data with t
(a) Give graphical and numerical summaries of MAXFT variable.
Min. 1st Qu. Median Mean 3rd Qu. Max.
13.00 46.00 52.00 51.96 59.00 84.00
Min. 1st Qu. Median M
HW9 Chapter 13
We need to use the formula of .
Here, . Use stata to compute the mean and standard deviation.
So the required number of households to estimate the total number of persons to within 20% is
So the required n