Scribd
Upload
88 views

May 2021 Examination Diet School of Mathematics & Statistics MT4614

This document provides instructions for an examination on the module Design of Experiments. It outlines the exam duration, instructions, and questions. The exam consists of 5 questions related to experimental design topics, including blocking, randomization, analysis of variance, and constructing experimental designs. Students are asked to analyze data, construct design plans, and explain statistical concepts.

Uploaded by

Tev Wallace
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
88 views

May 2021 Examination Diet School of Mathematics & Statistics MT4614

This document provides instructions for an examination on the module Design of Experiments. It outlines the exam duration, instructions, and questions. The exam consists of 5 questions related to experimental design topics, including blocking, randomization, analysis of variance, and constructing experimental designs. Students are asked to analyze data, construct design plans, and explain statistical concepts.

Uploaded by

Tev Wallace
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 6

MAY 2021 EXAMINATION DIET

SCHOOL OF MATHEMATICS & STATISTICS


MODULE CODE: MT4614
MODULE TITLE: Design of Experiments
EXAM DURATION: 2 hours
EXAM INSTRUCTIONS: Attempt ALL questions.
The number in square brackets shows the
maximum marks obtainable for that
question or part-question.
Your answers should contain the full
working required to justify your solutions.
The sum of the marks on this paper is 100.
INSTRUCTIONS FOR ONLINE EXAMS:

Each page of your solution must have the page number, module code, and your student
ID number at the top of the page. You must make sure all pages of your solutions are
clearly legible.

MT4614 May 2021, Page 1 of 6


1. A marine ecologist has sent you an email requesting your advice on the design
of an experiment. Part of her email is as follows.

I am planning an experiment to find out whether different species


of detrivore eat different amounts of giant kelp on sandy beaches. To
measure this exactly, it will be done in the lab. On the morning that
each run of the experiment is set up, I will collect fresh kelp from the
beach, as well as twelve detrivores of the designated species. In the lab,
the detrivores will be placed into a single glass jar containing sand. A
fixed amount of kelp will be cut into small pieces and placed in the glass
jar as soon as the detrivores have all burrowed into the sand. They are
nocturnal, so I will leave them to eat overnight. In the morning, I will
find out how much they have eaten.
How many runs do you think that I should do for each species? I
will get some undergraduates to help me with this messy work, and I
want to know how many of them I need.

Make notes on three important things to discuss with her during your meeting.
For each of these, explain the purpose of discussing it. [20]

2. An experiment was conducted on a single field at a British agricultural research


station to compare three different types of fertilizer for their effect on oats. As it
was expected that oat yields would differ from one part of the field to another,
the field was divided into six blocks, each containing three plots. Each type of
fertilizer was used on one plot per block, and there was complete randomization
within each block. At harvest time, the yield of oats on each plot was recorded
in tons per acre (this was before the introduction of metric units in the UK).

(a) Explain why this is an instance of an orthogonal block design. [2]

(b) Explain why this method of randomization is different from the method that
would have been used in a completely randomized design. [2]

MT4614 May 2021, Page 2 of 6


(c) Give the details of one linear model that might be appropriate for the re-
sponses in this experiment. Define all notation that you use in this. [4]

The sum y of all the yields was 334.6, and the sum y 2 of the squares
P P
(d)
of all the yields was 6488.78. For each block, the sum of the yields was as
follows.

Block 1 Block 2 Block 3 Block 4 Block 5 Block 6


64.9 47.0 60.8 43.2 62.5 56.2

For each type of fertilizer, the sum of the yields was as follows.

Type 1 Type 2 Type 3


92.8 116.2 125.6

(i) Produce the analysis-of-variance table for these data. [11]

(ii) Give the R code to produce this analysis-of-variance table, under the
model given in (c). You may assume that the yields are in a vector
called yield and that fertilizer and block have been declared as
factors. [3]

(iii) Test the null hypothesis that there are no differences between types of
fertilizer. [2]

(iv) Estimate the difference between the average yield with Type 2 fertilizer
and the average yield with Type 1 fertilizer (when used on this variety
of oats). [1]

(v) Give the standard error of the above difference. [2]

(e) Suppose that the experimenter had given you the data without explaining
that blocks had been used. Briefly explain what, if anything, would have
been different about your answers to parts (i), (iii), (iv) and (v). [3]

MT4614 May 2021, Page 3 of 6


3. A commercial bakery is planning an experiment to compare four different recipes
for making the batter for chocolates cakes, in combination with five different
baking temperatures. In two consecutive weeks, it will put aside the first part of
the morning on Monday, Tuesday, Wednesday and Thursday to do this. Within
each week, the recipes will be randomized to days. On each day, enough batter
will be made, from a single recipe, to prepare five cakes, which will then be baked
in different ovens, one at each of the specifed temperatures. After the cakes
have been baked and cooled down, each one will be subject to a standard test to
measure how crumbly it is.

(a) What are the observational units in this experiment? How many are there? [2]

(b) What are the treatments in this experiment? How many are there? [2]

(c) Identify the relevant factors on the treatments, and give their numbers of
levels. Draw the Hasse diagram for the factors on the treatments. [6]

(d) Identify any relevant non-treatment factors on the observational units, and
give their numbers of levels. Draw the Hasse diagram for these factors. [6]

(e) Give the appropriate skeleton analysis-of-variance table, showing strata, sources
and degrees of freedom. [8]

MT4614 May 2021, Page 4 of 6


4. In 1977, the Statistics Department of the University of Edinburgh moved from
its previous location (in a terrace house near the centre of town) to the new
Science campus at King’s Buildings. Imagine that the Head of Department (an
expert on Design of Experiments) decides to conduct an experiment to find the
most efficient way for staff to travel from their homes to the new building. He
enrols nine members of staff (labelled S1 to S9) who live within Edinburgh. He
will ask each of them to test three different methods of travel, using each one for
the five working days of one whole week. The methods are walking, cycling, and
public transport. The experiment will last for three weeks. For the ten journeys
each week, the staff member will record the journey time in minutes. (Before
data analysis, these times will be normalized by dividing by the journey distance,
measured in some appropriate units.)

(a) Identify the treatments, experimental units, and observational units in this
experiment. How many of each are there? [6]

(b) Construct the design and randomize it, presenting the final plan in a form
suitable for the head of department. [14]

5. Someone in the university of St Andrews tried to plan an experiment to compare


the effectiveness of various sorts of daily exercise. She proposed the following.

Walk on East Sands for half an hour every morning


Walk up and down The Scores for half an hour every morning
Walk along Lade Braes for half an hour every morning
Walk on East Sands for half an hour every afternoon
Walk up and down The Scores for half an hour every afternoon
Walk along Lade Braes for half an hour every afternoon
No exercise

Explain to her how to partition the treatment degrees of freedom in a meaningful


way. [6]

MT4614 May 2021, Page 5 of 6


END OF PAPER

MT4614 May 2021, Page 6 of 6

You might also like