MAST30025 Linear Statistical Models
Linear Statistical Models
Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: THEend8_
MAST30025 Linear Statistical Models Assignment 3
Submission deadline: Friday May 24, 5pm
This assignment consists of 5 pages (including this page) with 5 questions and 48 total marks
Instructions to Students
Writing
This assignment is worth 7% of your total mark.
You may choose to either typeset your assignment in LATEX, or handwrite and scan it to
produce an electronic version.
You may use R for this assignment, including the lm function unless otherwise specified.
If you do, include your R commands and output.
Write your answers on A4 paper. Page 1 should only have your student number, the
subject code and the subject name. Write on one side of each sheet only. Each question
should be on a new page. The question number must be written at the top of each page.
Scanning and Submitting
Put the pages in question order and all the same way up. Use a scanning app to scan all
pages to PDF. Scan directly from above. Crop pages to A4.
Submit your scanned assignment as a single PDF file and carefully review the submission
in Gradescope. Scan again and resubmit if necessary.
Blank page
MAST30025
Linear Statistical Models Assignment 3
Page 2 of 5 Semester 1, 2024
Question 1 (7 marks)
Let X =
[
1 2 1
2 1 1
]
and let A = XTX.
(a) Calculate A.
(b) Find a conditional inverse Ac such that r(Ac) = 1, or show that no such conditional inverse
exists.
(c) Find a conditional inverse Ac such that r(Ac) = 2, or show that no such conditional inverse
exists.
(d) Find a conditional inverse Ac such that r(Ac) = 3, or show that no such conditional inverse
exists.
Question 2 (10 marks)
We are interested in examining the yield of tomato plants that have been grown with certain
types of fertiliser. A study is conducted and the following data obtained:
Fertiliser
1 2 3
43 34 58
45 37 54
47 38 53
46 35
48
We fit the model
yij = µ+ τi + εij ,
where µ is the overall mean and τi is the effect of using the ith fertiliser.
For this question, you may not use the lm or ginv functions in R.
(a) Find a conditional inverse for XTX, using the algorithm given in Theorem 6.2.
(b) Characterise all solutions to the normal equations.
(c) Is 4µ+ τ1 + τ2 + τ3 estimable?
(d) Calculate a 95% confidence interval for τ1 − τ2.
(e) Calculate the F -test statistic for the null hypothesis τ1 = τ2 = τ3, and state the degrees
of freedom for the test.
MAST30025
Linear Statistical Models Assignment 3
Page 3 of 5 Semester 1, 2024
Question 3 (8 marks)
Consider the two-factor model with interaction
yij = µ+ τi + βj + ξij .
Suppose that there are a and b levels of the factors respectively. Now consider the set of
equations
ξij − ξ1j − ξi1 + ξ11 = 0, i = 2, . . . , a, j = 2, . . . , b.
(a) Show that the equations are not redundant.
(b) Show that these equations are equivalent to the hypothesis of no interaction.
(c) Thereby calculate the rank of the hypothesis of no interaction.
(d) Show that the hypothesis is testable, provided there exists at least one sample from each
combination of factor levels.
Question 4 (16 marks)
Data was collected on the world record times (in seconds) for the one-mile run. For males, the
records are from the period 1861–1999, and for females, from the period 1967–1996. The data
is given in the mile.csv, available on the LMS.
(a) Plot the data, using different colours and/or symbols for male and female records. Without
drawing diagnostic plots, do you think that this data satisfies the assumptions of the linear
model? Why or why not?
(b) Test the hypothesis that there is no interaction between the two predictor variables. In-
terpret the result in the context of the study.
(c) Write down the final fitted models for the male and female records. Add lines correspond-
ing to these models to your plot from part (a).
(d) Calculate a point estimate for the year when the female world record will equal the male
world record. Do you expect this estimate to be accurate? Why or why not?
(e) Is the year when the female world record will equal the male world record an estimable
quantity? Is your answer consistent with part (d)?
(f) Calculate a 95% confidence interval for the amount by which the gap between the male
and female world records narrow every year.
(g) Test the hypothesis that the male world record decreases by 0.4 seconds each year.
MAST30025
Linear Statistical Models Assignment 3
Page 4 of 5 Semester 1, 2024
Question 5 (7 marks)
An apple orchard has 24 trees set aside for an experiment which aims to examine the effect
of mulching on tree growth. There are three mulching treatments: A. Control (no mulch); B.
Wood chips; C. Garden compost. The trees are in a 4 × 6 rectangle, labeled as shown in the
diagram below. The experimenter has the resources to maintain 12 plots, each consisting of 2
adjacent trees. All trees in the same plot must have the same treatment.
1 2 3 4 5 6
7 8 9 10 11 12
13 14 15 16 17 18
19 20 21 22 23 24
In this question, if you need randomisation, use R and reproduce your R commands and output.
(a) Construct an appropriate experimental design. Draw the treatment allocation in the style
of the diagram above, and write down the matrices in the corresponding linear model.
(b) Now suppose the ground slopes down from the top to the bottom of the diagram.