POLI 110-003
Instructor: Michael Weaver
February 27, 2020
Midterm Examination
Winter 2019, Term 2
Exam Length: 80 Minutes
20 points
Name
Student Number
TA Name
1
Types of Claims:
(1) For this question, examine the following two claims:
(i) “The best possible government for China is the current regime under Xi Jinping and the
Communist Party.”
(ii) “The Chinese government should have been more transparent with its citizens about what it
knew and was doing about the COVID-19 (coronavirus) epidemic to prevent widespread anger
toward the regime.”
Assume you accept claim (i) is true.
(a) What type of claim is (i)? Be as specific as possible. ( 12 point)
(b) What type of claim is (ii)? Be as specific as possible. ( 12 point)
(c) If we assume (i) is true, on the basis of that alone, can we accept (ii) is true? If not, give an example of
another claim (iii) that we would have to accept (in addition to (i)) in order to accept (ii) and indicate
what kind of claim (iii) is. (1 point)
(2) Consider the following quote:
“Make no mistake, the end of Teck Resources’ Frontier oilsands mine is Justin Trudeau’s fault. . .
The fault is clearly with the Trudeau government’s entirely spineless response to blockades across
the country. Would you invest a nickel in Canada right now when all anyone has to do to choke off
the economy is throw a few wooden pallets across a rail line or highway and claim to be defending
Indigenous rights or the environment?” - Lorne Gunter, Toronto Sun, 24 February, 2020.
(a) What kind of claim is in this quote? (identify whether it is empirical or normative and then which
sub-type it is.) ( 12 point)
(b) What is the evidence given for this claim? ( 12 point)
2
(c) Identify one way in which the evidence for this claim does not meet the criteria for scientific evidence
(identify one specific criterion it does not meet and explain why this is so). (1 point)
Concepts, Variables, and Measures
For questions 3 and 4, select (and circle) one of the following causal claims:
1. Protests in solidarity with the Wet’suwet’en nation have increased citizens’ disapproval of the Liberal
government.
2. Anti-immigrant rhetoric in the United States has increased political participation among young Latino
voters.
3. Losing a job with high social status and then only finding new employment in a job with low social
status makes a person more likely to vote for a radical/extreme political party.
(3) Using either the concept that is causing the effect or the concept that is affected by the
cause (please write the concept here):
(a) Please write down a variable that corresponds to this concept. (1 point)
(b) Propose a measure for the variable you gave in Q3.b (1 point)
3
(4) Give a hypothetical example of one way that using the variable you gave in Q.3a could lack
validity (your example does not need to be true). Be sure to explain why (in your hypothetical
example) this variable would specifically lack validity. (2 points)
(5) Identify the level of measurement for the following two variables:
(a) Ranking of the most-equal countries based on the proportion of people receiving government benefits.
(1 point)
(b) A country’s change in average annual temperature (in degrees Celsius) between 1900 and 2019. (1
point)
(6) Under the Russian Constitution, a person may only serve as President for a maximum of
two consecutive terms. Because of this rule, Vladimir Putin is not eligible to run for reelection
in 2024. You have been hired by Putin’s Chief of Staff to discover the fraction of eligible voters
who support eliminating this rule (so that Putin may run again).
(a) Describe a measure for this variable that does not suffer from sampling bias but still produces
measurement bias. Explain clearly why the procedure would generate measurement bias. (2 points)
(b) Describe a different measure for this variable that does suffer from sampling bias. Explain clearly
why the procedure would generate sampling bias. (2 points)
4
(7) The following chart shows the performance of four different measures (X1, X2, X3, X4) of
variable X. The x-axis shows the difference between the observed value of X and the true
value from each measure (XObserved −XTrue). The y axis shows how frequently this difference
occurs across 10,000 cases. The dark vertical line indicates where XTrue − XObserved = 0 (the
measure exactly gets the right value for the variable).
X1 X2 X3 X4
−10 −5 0 5 10 15 −10 −5 0 5 10 15 −10 −5 0 5 10 15 −10 −5 0 5 10 15
0
2000
4000
6000
True X − Observed X
Fr
eq
ue
nc
y
Different measures of X
(a) Which of these measures obviously suffer from systematic measurement error? (1 point)
(b) Which of these measures obviously suffer from random measurement error? (1 point)
(8) Fill in the blanks with words or phrases to make the sentence accurate and precise. Use
exactly 3 different words to fill in these blanks; some words must be used more than once: (2
points)
When it is costly or impossible to measure a variable for the — all of the
cases we are interested in — we measure the variable for a — a subset of
cases — and use the to make (a/an) about
the .
(9) Please label the following sentences as True or False.
(a) Public health researchers want to know how the legalization of abortion reduced risks to women’s
health. To do this, they needed to measure the fraction of women who received abortions in illegal
clinics before legalization. They did this using a random sample survey of Canadian women from 1968
(the year prior to the legalization of abortion in Canada) that asked respondents whether they had ever
received an abortion. Because abortion was illegal at the time of the survey, many women lied and said
“no”. This is an example of sampling bias. (1 point) TRUE | FALSE
5
(b) Public health researchers a University X want to study smoking and vaping habits among young women.
They take a random sample of 20 from the 10,000 women enrolled at their university. In reality, only
10% of women at the university smoke or vape at all. In an extremely unlikely chance occurrence (about
1 in a million) all of the 20 women in the sample smoke or vape. This is an example of sampling bias.
(1 point) TRUE | FALSE
6