Homework 1: Review of Probability and Statistics
ECON20110/30370 Econometrics
Judith Guo
September 27, 2024
As you work through the following problems, you may find it helpful to refer to the mind-map on proba-
bility and statistics, available on Blackboard, or course materials from Advanced Stats.
1. Consider the joint probability mass function (pmf), fX,Y (x, y), of random variables X and Y as shown
in the table below:
y
0 1 2
0 0.1 0.0 0.2
x
1 0.3 0.1 0.3
Answer the following questions:
(a) Marginal and Conditional Distributions
i. Find the marginal pmfs of X and Y ;
ii. Compute the conditional pmf of Y given X = x for x ∈ {0, 1}.
(b) Independence
Are X and Y independent? Justify your answer clearly.
(c) Law of Iterated Expectations (LIE)
i. Calculate E(Y ) and E[Y | X = x] for each value of x.
ii. Using these, prove the LIE:
E(Y ) = ∑ E[Y | X = x] · P{X = x}
x∈{0,1}
(d) Conditional Expectation
Derive an expression for E[Y | X].
2. (a) Covariance and Linearity of Expectation
Using the linearity of expectation, show that:
cov(Z, aX + bY + c) = a cov(Z, X) + b cov(Z,Y ).
(b) Variance Formula
From the result in part (a), show that:
var(aX + bY + c) = a2 var(X) + 2ab cov(X,Y ) + b2 var(Y )
[Hint: Let Z = aX + bY + c and use the result from part (a).]
1
Judith Guo ECON 20110/30370
3. Best Constant Predictor
Suppose you are predicting the score of a game player, denoted as X. Demonstrate that E(X) is the ‘best
constant predictor’ for X, meaning that it is the value of c that minimises the mean squared prediction
error:
L(c) := E[(X − c)2 ]
[Hint: Start by expanding the squared term, apply the linearity of expectation, and take the first-order
condition to find the optimal value of c.]
4. Hypothesis Testing
Consider a clinical trial involving two groups of participants (100 individuals per group), all of whom
are suffering from the same health condition. One group (the treatment group) receives a new dietary
supplement, while the other group (the control group) is given a placebo. In the treatment group, 75
participants report improvements, while in the control group, only 65 participants show improvements.
(You may assume that participants were randomly selected from a larger population and randomly
assigned to the two groups, allowing us to treat them as independent groups for analysis.)
(a) Conducting a Hypothesis Test
Test the hypothesis that the new dietary supplement is ineffective versus the alternative that it is
beneficial. Conduct the test at both the 1% and 10% significance levels.
(b) p-value
Calculate the p-value associated with the test statistic. How does this p-value support the conclu-
sion you reached in part (a)?
(c) Type I and Type II Errors
What are Type I and Type II errors? Discuss what these errors mean in the context of this clinical
trial. Which type of error would be more critical to avoid in this scenario, and why?
(d) Sample Size
Suppose the number of participants in each group increased to 300, but the improvement rates
remained the same. What conclusions can you draw from this new scenario? Would this result be
surprising? Explain why or why not.