ECO220Y: Homework 14

Note: Use the statistical tables with our course aid sheet, with approximation when necessary. Do not use software.

Required Exercises: Chapter 13: 1, 3, 5, 9, 11, 13, 19, 41, 49, 55

Required Problems:

(1) Recall Karlan and List (2007) “Does Price Matter in Charitable Giving? Evidence from a Large‐Scale Natural Field
Experiment.” This Stata summary of donation amount (in dollars) focuses on people in the control group (those offered
no match) and who gave money.

. summarize amount if (gave==1 & control==1), detail

amount
-------------------------------------------------------------
Percentiles Smallest
1% 3 2
5% 10 2
10% 10 3 Obs 298
25% 20 5 Sum of Wgt. 298

50% 25 Mean 45.54027

Largest Std. Dev. 41.37982
75% 55 150
90% 100 160 Variance 1712.29
95% 125 250 Skewness 1.906669
99% 160 300 Kurtosis 8.734306

(a) Does the “Nearly Normal Condition” hold? Explain why or why not. Should we be concerned? Can we
proceed and use the statistical formulae in Chapter 13?

(b) Compute the 80% Confidence Interval Estimator of the mean. Interpret it. Next, how does it compare with
the percentiles from the STATA summary above? Why?

(2) In Ontario, to meet regulatory standards, slot machines must have a minimum payout of 85%: for every dollar
played, it must pay out a minimum of 85 cents on average. The OLG brochure “Your Guide to Playing Slots” explains:

EXCERPT: In Ontario, the payout for slot machines has to be set at 85 per cent or
higher. This figure is set and controlled by the Alcohol and Gaming Commission of
Ontario (AGCO), the independent body that regulates specified OLG activities.
Eighty‐five per cent is the minimum payout rate and most machines are preset
for higher levels. Keep in mind, this doesn’t mean that for every $10 you put into
a slot machine that you will get $8.50 returned to you. The payout of 85 per cent
is based on the long‐term return of the machine, which could include hundreds
of thousands or even millions of spins.

(a) If a casino owner must prove a slot machine complies with standard, how to write the hypothesis test? Use
formal notation. In plain English, explain what Type I and Type II errors are in this context. For 𝑛 5,000
plays of the machine, what is the rejection region for a 5% significance level? What does it mean?

(b) If the AGCO must prove a slot machine does not meet the standard, how to write the hypothesis test? Use
formal notation. In plain English, explain what Type I and Type II errors are in this context. For 𝑛 5,000
plays of the machine, what is the rejection region for a 5% significance level? What does it mean?

Page 1 of 2
 (c) If you are most concerned about protecting players, which way should the hypotheses be set up? If you are
most concerned about protecting the profits of the casinos, which way should the hypotheses be set up?

(d) For Part (a), how to limit the chance of a Type I error? Would casinos appreciate this suggestion?

(e) For Part (a), how to limit the chance of a Type II error? Would casinos appreciate this suggestion?

(f) For Part (a), other things equal, which has a higher chance of a Type II error: a machine that is only
marginally more generous the minimum payout or a machine that is much more generous? Why?

(3) A researcher is auditing how efficiently the capacity of a ski lift is utilized during peak times when there is a line of
people waiting. Each lift has space for four people. For a random sample of 250 arriving lifts, the researcher records the
number of people who board it. The variable pct_occ records the percent of spaces occupied.

pct_occ | Freq. Percent Cum.

------------+-----------------------------------
0 | 10 4.00 4.00
25 | 21 8.40 12.40
50 | 122 48.80 61.20
75 | 23 9.20 70.40
100 | 74 29.60 100.00
------------+-----------------------------------
Total | 250 100.00

(a) In testing whether more than one fourth of lifts are fully utilized (i.e. all four spaces are occupied), what are
the hypotheses in formal notation? What is the value of the test statistic?

(b) It testing whether the average occupancy is above 50 percent, what are the hypotheses in formal notation?
What is the value of the test statistic?

(c) In testing whether less than 1 in 10 lifts are either empty or have only 1 occupant, what are the hypotheses
in formal notation? What is the value of the test statistic?

(d) If the researcher repeated this at a different ski resort and wished to compare whether the original ski resort
has a higher incidence of fully occupied lifts, what are the hypotheses in formal notation? From our course
aid sheets, what would be the correct formula to use to compute the test statistic?

(e) Continuing with Part (d), if they wished to compare whether the original ski resort has a higher average
occupancy, what are the hypotheses in formal notation?

(4) Test 𝐻 : 𝜇 0 versus 𝐻 : 𝜇 0. For all parts, using the appropriate table, not software.

(a) For n = 19, the mean is ‐0.421 and the s.d. is 1.244. What is the P‐value? Are these results statistically
significant at 𝛼 of 0.10, 0.05, 0.01, or 0.001?

(b) For n = 30, the mean is 0.321 and the s.d. is 0.501. What is the P‐value? Are these results statistically
significant at 𝛼 of 0.10, 0.05, 0.01, or 0.001?

(c) For n = 1,900, the mean is 0.053 and the s.d. is 1.351. What is the P‐value? Are these results statistically
significant at 𝛼 of 0.10, 0.05, 0.01, or 0.001?

Page 2 of 2