Unit 4 REVIEW Name __________________

1. Andrew manages a comic book store and is trying to increase sales. He decides to run an experiment for
the next 12 Mondays. He randomly picks 6 of the Mondays on which he will say hello to each person as
they enter the store. On the other 6 Mondays, he will say hello to each person and offer them something
to drink. For each day, he will determine the percentage of people who make a purchase. The percentage
of people who made purchases each day and the treatment received are given below.

a. What are the treatments for this experiment?

Saying hello and Saying hello and offering a drink.
b. Find the value of mean hello – mean hello and drink
24-17=7
c. Describe how to conduct one run of a random
assignment to decide whether the different treatments
cause a difference in the percentage of people who
make a purchase.
1. Assume there is no difference between the two treatments.
2. Put all the percentages on cards, and mix together.
3. Choose six cards and call them “hello only”, the remaining six are the “hello and drink offer”
cards.
4. Find the mean of each new stack, and find the difference in means.
5. Continue steps 3 and 4 many times.
6. Record the difference in means found in a histogram.
7. Find the total of means greater than 7 and less than negative 7.
8. Take the total of means divided by the number of means found. If this percentage is less than
5% than the test is statistically significant.

d. The histogram below provides the results of 200 random assignments. The value shown in the
histogram is mean hello - mean hello and drink. Use the histogram to estimate the probability
that if the treatments made no difference, you could get a difference just by random assignment
that is at least as extreme as what occurred in the real experiment.

4
Probability __200 = 2% __

e. Is the difference in the percentage of people making purchases statistically significant? Why or
why not?

Yes this test is statistically significant, 𝑝 < 2%

 2. Daniel recently discovered a new route to school. He wondered if the travel times were different for the
two routes. He decided to do an experiment. He randomly assigned five days to the old route and five
days to the new route. The travel time (in minutes) for each day are given below as well as the summary
statistics for these two sets of data.

a. Describe how to use a randomization test to determine if the travel times were different.
b. Assume there is no difference between the two treatments (travel routes).
c. Put all the times of the routes on cards, and mix together.
d. Choose five cards and call them “old route”, the remaining six are the “new route” cards.
e. Find the mean of each new stack, and find the difference in means.
f. Continue steps 3 and 4 many times.
g. Record the difference in means found in a histogram.
h. Find the total of means greater than 1.6 and less than negative 1.6.
i. Take the total of means divided by the number of means found. If this percentage is less than
5% than the test is statistically significant.
b. Daniel performs 200 runs and
creates the randomization
distribution at right. The values
shown in the histogram are mean
old route – mean new route. Use the
randomization distribution to
estimate the probability that, if the
route made no difference, Daniel
would get a difference just by the
random assignment of days that is
as extreme as what occurred.
21
= 10.5%
200

c. What is your conclusion about whether there is statistically significant evidence that the travel
times are different for the two routes?
The two routes do not show there is any significance statistically. 𝑝 > 5%
 3. A local agency wants to build a playground in an abandoned field. But they must first determine if the
soil is contaminated. To measure the level of contamination, they divided the field into 100 different
same-sized areas and randomly chose 25 of them. Then they took some soil from each of the chosen
areas and found no elevated levels of any toxic substances.

a. What type of study is this?

This is a survey, it is random but there are no treatments.

b. Explain why it cannot be the other two types of statistical studies. Explain how randomization is
used in these studies.
This cannot be an observational study because there is randomization, when you have an
observational study there is NO randomization.
This cannot be an experiment because there are NO treatments. The soil is just tested for level of
contamination.

c. What can you conclude, if anything, from this study?

You can conclude that the levels of toxic substances were not elevated from the 25 areas that were
chose. You can generalize that the field does not have elevated toxic substances because the areas
testes were RANDOMLY chosen.

4. The scores of written tests for a job are normally distributed with a mean of 70 marks and a standard
deviation of 2 marks.

a. Sketch a graph of the distribution of test marks. Include a scale on the axis that shows the mean
and one, two and three standard deviations from the mean.

b. What percent of candidates scores more than 70 marks?

50%
c. What percent of the candidates scored more than 68 marks?
84%
d. What percent of the candidates scored between 64 and 76 marks?
99.7%
e. What percent of the candidates scored less than 71 marks?
69.15%
f. If the hiring committee wanted to be sure that at least 97.5% of the candidates scored at least 70
marks, where should the mean score be set and why? Assume the distribution remains normal
with standard deviation of 2 marks.
74 would be the new mean: because you want 70 to be 2 deviations below the mean
70−𝑥
you can use the equation: 2= 2
5. In 2006, PISA (Programme for International Students Assessment) conducted an international study of
science learning. The scores for U.S. students on the test of Living Systems were approximately normal
with mean 487 and standard deviation 117.
a. Mike, a U.S. student, got a score of 547 on this test.

i. How many standard deviations above the mean was Mike’s score?
.5128 above the mean

ii. What was his percentile?

59.15%

b. The standardized value for Clayton’s score on the Living Systems test was –1.52. What was
Clayton’s score on this test?
306.16

c. Lauren scored at the 90th percentile for U.S. students. What was Lauren’s score?
639.1