# 统计作业代写_STA 4164 Fall Assignment 3 ## STA 4164 Fall Assignment 3

### Instruction:

• Do not copy and paste all SAS output, use part of the output that is related to answering the question. Present the relevant output and write your answer below it.
• Late assignments will not be accepted for grading.
• If α is not specified, you could assume that it’s is 5% for any test and confidence intervals.
• Interpret coefficient, test, confidence interval in the context of problem.
• Round final results to 2 decimal places. To avoid rounding mistake, keep your intermediate calculation with at least 4 decimal places. ### Problem 1(50pt)  统计作业代写

For this problem, we will use the data of exercise # 19, chapter 14 in the textbook.Use SAS to fit a regression of ’fev’ on ’age’, ’height’, ’weight’.Then answer the following questions

a. Examine a plot of the jackknife residuals versus the predicted values. Are any regression assumption violations apparent? If so, suggest possible remedies.
b. Use PROC UNIVARIATE to examine numerical descriptive statistics, histograms, and normal probability plots of jackknife residuals. Is the normality assumption violated? Perform a suitable statistical test for normality assumption. Remember to state the hypothesis and your conclusion.
c. Perform outlier diagnostics. First, we will consider outliers that manifest on the predictor space using leverage statistics. Use PROC REG in combination with PROC SORT, and PROC PRINT to print out the top 5 observation with highest leverage. Which observation if any is suspected as outlier?
d. Now we will investigate outliers that manifest on the response direction. Repeat the analysis in (c) using jackknife residuals.
e. Lastly, we will check influential observations. Repeat the analysis in (c) using Cook’s distance.

### Problem 2(20pt)

We will use the data and output given in exercise # 4, chapter 16. To read the data into SAS, use the following codes: data prob2; input CITY Y X1 :comma5. X2 X3; datalines; {copy data here} run;

a. Manually perform backward elimination to choose the best model. Then use PROC REG with appropriate option to verify the result.
b. Use SAS to perform stepwise approach with option of ’slentry = 0.3’. Write the model achieved at each step. Perform hypothesis tests corresponding with each step. You can use SAS output with given p-value to conduct the tests.

### Problem 3(20pt)  统计作业代写

For this problem, we will use data in exercises #10, chapter 17.
a. Do the data provide sufficient evidence to suggest that values for mean dissolved oxygen content differ significantly among the four locations? Manually construct the appropriate ANOVA table and perform the relevant hypothesis test to answer that question. Verify the result using SAS.
b. Estimate the effect ˆαi for location ’i’ in fixed-effect one-way ANOVA using output from (a).
c. Use regression approach to answer question (a). Write the estimated regression function and conduct appropriate hypothesis test. Remember to state the hypothesis and give your conclusion. Hint: You need to create 3 dummy variables Z1, Z2, Z3 (with location 4 as reference group). 