You are on page 1of 5

8cmp06te_SP2.

qxd 3/13/06 2:25 PM Page 71

Answers

Investigation 2

ACE 4. The population being studied is students who


Assignment Choices attend the particular school. The sampling
method is to somehow choose 26 students, one
Problem 2.1 with a name beginning with A, one with B, etc.
Core 1–4 5. Systematic selection; Systematic selection
Other Connections 18–20 from class lists would probably give a fairly
Problem 2.2
predictive sample, provided the lists include
the names of all students in the school.
Core 5–9, 21, 25
However, some homerooms will be under-
Other Connections 22–24, Extensions 41–44 and
represented, and students with names such
unassigned choices from previous problems
as “Aardvart” are probably excluded.
Problem 2.3 6. Random selection; Based on the assumption

ACE ANSWERS
Core 10–13, 26 that 350 students attend the assembly, selection
Other Applications 14, Connections 27–33 and by choosing red and white beans could
unassigned choices from previous problems produce a representative sample because
every student has an equal chance of being
Problem 2.4 in the sample.
Core 15–17
7. Convenience Sampling; Selection by

2
Other Connections 34–40, Extensions 45, and
surveying all members of particular classes
unassigned choices from previous problems
would probably not give a representative
Adapted For suggestions about adapting
sample because all students do not have an
Exercises 5–8 and other ACE exercises, see the equally likely chance of being surveyed. It
CMP Special Needs Handbook. may be that students in these classes are
Connecting to Prior Units 18, 19, 22–24: Data
assigned more homework on average and
About Us; 19, 26: Comparing and Scaling; 20: Bits thus spend more time doing homework.
and Pieces III; 22–24: Data Distributions; 28–40: 8. Voluntary response; This voluntary-response
How Likely Is It? method would not give a representative
sample because it depends on people’s
choosing to complete and return the surveys.
Applications Not everyone will complete a survey, and
those who do may not place it in the box at
1. The population being studied is the teenagers
the end of the day; those who complete
who read the magazine. The sampling method
surveys may have some special reason for
is to ask readers to voluntarily write in to the
responding.
magazine.
9. Voluntary response; Possible explanation: The
2. The population being studied is middle school
results of this survey may not describe the
students. The eighth-grade students keep
opinions of all the show’s listeners because
records of the time they spend on the phone
only those listeners that have the time and
each day for a week.
want to call in with their opinion will be
3. The population being studied is middle school included. Some may try to call in and may not
students. Forty cards are chosen from a box get through, while others may not listen to the
containing cards with the names of all the radio program at the time that the opinion
students in the school. question is asked because they are at work.

Investigation 2 Choosing a Sample From Population 71


8cmp06te_SP2.qxd 3/13/06 2:25 PM Page 72

10–13. In each case, the company needs a device to b. Take the percent of students that said yes
make the choice of sample independent of any in the sample and multiply it by 700, the
influence other than chance. These could be: population of the school.
• A random number generator (i.e. RANDINT 16. 6 hours
(1, 5000)) will produce random integers between 17. Possible answer: 30 students is better because
1 and 5,000. The toy company could select 30 if you choose too small a sample, for example
numbers at random and test these video games 5 students, you could end up picking
as they came off of the assembly line. (We need
2 students with outlier data values. These
to track the number for each game as it is
outliers affect the data more since you only
produced.)
have 5 in your sample. With more students in
• A spinner coded 0–9 can be used. (i.e. spinning your sample, you are more likely to get a
the spinner six times with results 0, 5, 4, 1, 1, 7, representative sample.
means that the music company should test the
54,117th CD produced. Or the company could
subdivide the 200,000 CD’s by artist and test a Connections
random number for each artist).
18. D
• Pieces of paper numbered 0–9 placed in a hat
can also be used. (i.e. choosing 4 numbers at 19. a. One jar makes approximately
random with replacement from the hat allows 15 sandwiches so 100 jars would make
you to construct numbers from 0 to 9,999) The 1,500 sandwiches.
fireworks company needs numbers from 0 to b. From 5 to 18 years is 14 years; 100 4 14 is
1,500, so if the number chosen is greater than approximately 7 jars per year.
1,500, then it is disregarded and a new number
is chosen. c. There are 15 sandwiches per jar; there are
105 sandwiches per year; assuming 52
In each case, a sample size of at least 30 should weeks in a year, a child eats about 2 peanut
be chosen at random. butter sandwiches a week.
14. a. The first-come, first-served strategy may
20. J
produce some biased samples because the
students might be listed in some sort of 21. a. The more time has passed since the
order. We want to guarantee that every previous eruption, the greater the height of
student has an equal chance of being the eruption.
included in a sample. b. We can’t make a prediction about other
b. Again, this strategy does not give every geysers from just one situation. We would
student a chance of being included in the need data from other geysers in order to
sample, and the students might be listed in analyze whether this behavior is typical for
some order—perhaps with odd-numbered all geysers.
students differing from even-numbered 22. Median for grade 6, 20 minutes; Median for
students in some way. grade 8, 35 minutes
15. a. Possible sampling method: Find a random 23. a. Possible answer: The range of time spent on
sample of 50 students by taking a homework for students in grade 6 (0 to
numbered student list and using slips of 85 minutes) and grade 8 (0 to 80 minutes) is
paper to choose the sample. For example, if about the same. Where there is a gap in the
there are 700 students in the school place time for 6th graders between 50 and
the numbers 0, 1, 2, . . . , 9 in a bucket. Pick 70 minutes, there are no gaps in the times
one slip from the bucket for the hundreds for 8th graders. Time spent on homework
digit, replace and pick another for the tens, for students in grade 6 is clustered between
replace that, and pick another for the ones 0 and 20 minutes; time for students in grade
digit. If a number picked is greater than 8 is clustered between 20 and 45 minutes.
700, disregard and repeat the process.

72 Samples and Populations


8cmp06te_SP2.qxd 3/13/06 2:25 PM Page 73

b. Students in grade 6 seem to spend less time 1


28. 18; Since Annie is one person in a class of 18
studying on Monday nights than do 1
she has a 18 chance of being chosen on any
students in grade 8. Using the median,
1
typically a 6th grade student spends 20 day. So there is a 18 chance she will be picked
minutes while an 8th grade student spends on Monday.
35 minutes. At most, 25% of 6th grade 1
29. 18; Since Annie is one person in a class of
students spend no time on homework. Only 1
one 8th grade student spends no time on 18 she has a 18 chance of being chosen on any
homework. For both grade levels, few 1
day. So there is a 18 chance she will be picked
students spent more than 50 minutes on on Tuesday.
homework. 1
30. 324; To understand what the chance is of being
24. These data may not be representative of all
chosen on both days students might make a
school nights because students may have
chart (see below) that is 18 by 18, with each
more homework on some nights than on
row representing the chance for any one
others.
student to be chosen on Monday, and each
25. a. The mean is 23; the range is 5. column representing the chance for any one
b. Answers will vary. Possible answers: Add student to be chosen on Tuesday (like tossing
44, 45, and 45, which gives a new mean of a die twice, only this time an 18-sided die).

ACE ANSWERS
29.5 and a range of 25. Or add 0, 0, and Choosing Annie on both days would be
10,000, which gives a new mean of 1,016.1 represented by a single square in this grid.
and a range of 10,000. 1
The area model indicates a 324 chance that
c. Answers will vary. Possible answer: Add 5, Annie will be chosen BOTH days.
9, and 8, which gives a new mean of 18.3 12
and a range of 20. Or add 22, 22, and 22, 31. 18 or about 67%

2
which gives a new mean of 22.7 and a range 11
32. 17 or about 65%
of 5.
2 2
d. Answers will vary. Generally speaking, 33. Since 3 of the class are girls we would expect 3
students may find that the ranges change. of the chosen students to be girls, that is, 4 out
Adding values above the mean will shift of 6. But this is a very small sample so we
the mean higher; adding values below the would not be surprised if there were only
mean will shift the mean lower. 2 girls chosen.
26. H
34. a. 800 = 1 ; 799; 798
400 399 398
8 112 2
27. a. 25 = 350 (32%); 112 students,
8 112 400 399 398 63520800
since 25 = 350 b. 12.5%; 800 3 799 3 798 = 510081600  12.5%
b. 25%; 88 students, 25% 3 350  1 150 1
88
35. a. 2; P(6th grade girl) = 300 or 2
88 students (350)
1 150 150 100 1
c. Sample 1 predicts the greater fraction of b. 8; P(3 girls) = 300 3 300 3 200 = 8 or
students 1 1 1
3232 = 8
1
2
d. Answers will vary. It is quite possible to get
36. 150 students would prefer a healthy food
different results from different samples,
even when using a random sample. 37. 500 students; 150 + 150 + 100 + 100 = 500
students.
200 1
38. 800 = 4

Investigation 2 Choosing a Sample From Population 73


8cmp06te_SP2.qxd 3/13/06 2:25 PM Page 74

500 43. Answers will vary. (Teaching Tip: It might be


39. Install 1 healthy food machine; 800 of all
interesting to invite a representative of one of
students surveyed chose a healthy food the local media to visit your class to describe
machine so this would be an option that how they conduct polls. This would make a
would probably satisfy all students. nice culminating activity for the unit.)
600
40. a. 800 or 75% probability 44. a. 1000 4 203,000,000  0.0005%
b. Can’t tell; we don’t know the breakdown b. Possible answer: The sample is probably
of male/female for the 600 who chose taken from a cross section of people rather
healthful snacks. You can’t tell this from the than just of people in one geographical
Principal’s poll. It may be that a preference location, occupation, political party, age
for healthful snacks is dependent on group, gender, or the like.
gender. (In fact Alyssa’s sample seems to 45. a. Answers will vary.
indicate this.)
b. Answers will vary.
c. Can’t tell; we don’t know the breakdown of
c. Answers will vary.
male/female for the 600 who chose healthy
food. d. Answers will vary. Students should use the
6 line plot with a sample size of 30 to predict.
d. Install 1 healthy food machine; 8 of all
e. Possible answer: Students will predict that
students surveyed chose fruit machine so
the company’s fixed percent of each color
this would be an option that would
will be very similar to the percent of each
probably satisfy all students. It seems that
color in the population. They may reason
he has a large majority supporting his
that the larger sample size will generally
decision. (He may want to check that the
produce the most accurate estimate.
proportion in favor is the same for both
(The M&M company’s fixed percent is
genders by checking a sample of males and
30% brown, 20% red, 20% yellow, 10%
females separately.)
green, 10% blue, and 10% orange.)

Extensions Possible Answers to


41. Possible answer: Pre-election polls can be Mathematical Reflections
inaccurate as a result of a poor sampling
method. When sensitive issues are being 1. Collecting data from a sample rather than
contested in an election, polled voters may from an entire population simplifies the task.
not accurately report their views, especially if If the sample is chosen wisely, you can tell a
those views are unpopular or are believed to lot about what is true for the population.
be unacceptable for voicing in public. (Note: 2. Convenience sampling is selecting a sample of
Pollsters now often ask people whether or not people who are easy to reach. The easiest
they are likely to actually vote. Also, it is sample to reach, though, might not be a good
known that polls in recent years about voting predictor of the entire population. Systematic
issues that involve race and gender have led sampling is a formulaic procedure for
to inaccuracies.) choosing from a population. It may be a fairer
42. Possible answer: A political party may poll method than convenience sampling, but one
people from areas in which the majority of must be careful that the systematic procedure
people are of their same party or hold views does not bias selection toward or away from
similar to those held by the candidates. They some sector of the population. Voluntary-
might frame questions in a way that response sampling produces data from people
encourages responses favorable to their who choose to respond. It may be easy to
candidate with the intent being to create a implement, but the data can be misleading
public impression that their candidate is “on a because you get reports from either those
roll” in the hopes that others will be who are atypically responsible or those who
persuaded to join that side. care strongly about an issue. Random

74 Samples and Populations


8cmp06te_SP2.qxd 3/13/06 2:25 PM Page 75

sampling is a system that gives every sample 4. Identify each individual with a code number,
of a population an equal chance of being letter or name; then select the sample by
selected. It can be harder to implement, but it generating random numbers (using spinners, a
may give the best results in terms of being calculator or computer, or number cubes). Or,
predictive of the population. draw names or numbers from a bowl without
Note: Probability tools can be applied to looking. Using a calculator or a computer to
distributions of random samples. generate random numbers is probably the
3. a. In a random sample, every sample of given most efficient method. Using spinners or
size in a population has the same chance of number cubes will work but is more time-
being selected. With the other three types— consuming. Drawing names from a bowl may
convenience, in which people who are be even more time-consuming because you
present are sampled; voluntary response, in must first write each value on a slip of paper.
which people choose to respond; and 5. Several samples from the same population
systematic in which some rule is used to should have similar distributions (as shown in
select the sample—everyone does not have box plots, stem plots, and line plots) and
an equally likely chance to be a part of the summary statistics (median, quartiles, and
sample. mean). However, similarity among small
b. Random sampling is less likely to produce samples will be less likely than among larger
a biased sample—a sample that is not a samples.

ACE ANSWERS
good predictor of the population. In the
long run, most random samples will
somewhat resemble the population(s) from
which they were drawn, though there is
always a small chance that a random
sample will be quite different from the

2
population.

Investigation 2 Choosing a Sample From Population 75

You might also like