## Preview Extract

Chapter 10 Determining the Size of a Sample

1) Which of the following is NOT a component needed for SSI's Formula for determining

how many telephone numbers are needed?

A) incidence rate

B) completion rate

C) working phone rate

D) completed interviews required

E) number of qualified interviewers

Answer: E

2) Which of the following would be defined as the percentage of respondents that qualify for

a survey based on criteria such as age, income, and race?

A) demographic incidence

B) geographic incidence

C) demographic occurrence

D) product incidence

E) customer data

Answer: A

3) When all other factors are held constant, as we increase the level of accuracy, the sample

size and the cost of a marketing research survey are best characterized by which of the

following?

A) The sample size will increase but the cost will decrease.

B) The sample size will decrease but the cost will increase.

C) The sample size and the cost of the survey will increase.

D) The sample size will remain the same but the cost will increase.

E) The sample size will decrease but the cost will remain the same.

Answer: C

4) Unfortunately, many managers falsely believe that sample size is:

A) related to proper data analysis.

B) related to the representativeness of the sample.

C) determined by computer programs.

D) an irrelevant "statistical" technicality issue.

E) related to the level of accuracy desired.

Answer: B

5) Which of the following statements is most accurate regarding the relationship between

sample size and the sample representativeness?

A) There is a high relationship between the size of a sample and the representativeness of the

population from which it is drawn.

B) There is no relation between the size of a sample and the representativeness of the

population from which it is drawn.

C) You cannot have a representative sample unless the sample size is equal to or exceeds 10

percent of the population.

D) Sample size determines representativeness but only if the sample plan is a probability

sampling plan.

E) Sample size determines representativeness but only if the sample plan is a nonprobability

sampling plan.

Answer: B

6) The sample size determines:

A) representativeness.

B) accuracy.

C) representativeness and accuracy.

D) the population statistic value.

E) the mean generated from the sample statistic.

Answer: B

7) Sample accuracy refers to:

A) the extent to which the sample is validated.

B) the extent to which the sample statistics differ from the true population values the statistics

represent.

C) the extent to which the population statistics differ from the representativeness of the

sample.

D) a statistical concept that can be assessed only theoretically.

E) how close the sample statistics match the predetermined values expected by management.

Answer: B

8) In a study of the U.S. workforce population, a research company interviews 5,000 persons

on a street corner in New York. If they decide to increase the sample size to 10,000, they

would:

A) increase the sample size and the representativeness of the sample.

B) increase both sample accuracy and representativeness of the sample.

C) increase only the representativeness.

D) decrease the sample accuracy.

E) increase the sample size, but the sample would still not be representative of the population.

Answer: E

9) One of the reasons why a marketing practitioner should have a basic understanding of

sample size determination is because:

A) many practitioners have a false belief that sample size doesn't determine a sample's

representativeness.

B) it helps managers to manage their resources better.

C) a marketing manager should understand that a sample's representativeness is not related to

its accuracy.

D) the size of the sample is never a major cost factor.

E) managers always have a "small sample bias"; they believe small samples are more

accurate.

Answer: B

10) Which of the following is NOT one of the axioms of sample size and accuracy?

A) The only perfectly accurate sample is a census.

B) A probability sample will always have some inaccuracy (sample error).

C) Increasing sample size increases the sample's representativeness.

D) A probability sample size can be a very tiny percentage of the population size and still be

very accurate.

E) The size of a probability sample depends on the client's desired accuracy balanced against

the cost of data collection.

Answer: C

11) Which of the following is the most correct method of determining sample size?

A) percentage of population approach

B) all that can be afforded approach

C) all that time will permit approach

D) using the "100" for local study; "1,000" for national study approach

E) the confidence interval approach

Answer: E

12) Sources of error that come from sources other than the sample selection method and

sample size are referred to as:

A) serious mistakes.

B) nonsampling errors.

C) errors caused by competitors.

D) errors caused by clients.

E) errors caused by statisticians.

Answer: B

13) Which of the following is true regarding probability samples?

A) They are as perfect as a census and contain no errors caused by competitors.

B) They will always contain some inaccuracy (sample error).

C) They contain serious mistakes, but can be adjusted by statistical weighting procedures.

D) They are particularly susceptible to nonsampling errors.

E) When they approach large values, say 1,000, they are equivalent to a census.

Answer: B

14) Which of the following is true regarding a probability sample?

A) The larger the sample size, the more likely the sample is representative.

B) The larger the sample size, the more accurate it is (less sample error).

C) The more representative the sample, the larger the sample size.

D) They only become accurate when they are larger than 1,000.

E) The larger the sample size, the more room there is for inaccuracy.

Answer: B

15) If we were to graph the relationship of sample size (x axis) to sample accuracy (y axis),

we would notice that:

A) there is a linear relationship between size and accuracy.

B) accuracy increases quickly, up to about sample size 500, and then accuracy levels off with

relatively small gains made even when sample size is increased to as much as 2,000.

C) accuracy increases quickly, up to about sample size 5,000, and then accuracy levels off

with relatively small gains made even when sample size is increased to as much as 200,000.

D) accuracy increases quickly, up to about sample size 50, and then accuracy levels off with

relatively small gains made even when sample size is increased to as much as 100.

E) We cannot graph sample size and accuracy because we must also include the level of

confidence and the variability within the sample data.

Answer: B

16) If you were to graph sample accuracy and sample size, which of the following

generalizations would be most accurate?

A) Accuracy is very low, even with small sample sizes of 50 or below.

B) Accuracy constantly increases as sample size increases; a sample of 2,000 is four times

more accurate than a sample of 500.

C) With increases in sample size, sample accuracy decreases.

D) We cannot graph sample size and accuracy because we must also include the level of

confidence and the variability within the sample data.

E) Accuracy increases rapidly when sample size increases up to about 500 and then levels

off.

Answer: E

17) If we know the level of confidence (1.96 for 95 percent), variability estimates, and the

size of a sample, there is a formula that allows us to determine:

A) the costs of the sample.

B) the size of the sample.

C) the representativeness of the sample.

D) p or q.

E) the accuracy (sample error).

Answer: E

18) Which of the following is the best definition of variability?

A) It is the amount of dispersion in a data set containing interval or nominal data.

B) It is the difference between scores in the present sample and scores in a previous sample.

C) It is the amount of dissimilarity in ordinal data.

D) It is the amount of dissimilarity (or similarity) in respondents' answers to a particular

question.

E) It is the amount of responses in one respondent's answers to a particular survey question.

Answer: D

19) Which of the following is true with regard to variability?

A) p is always less than q

B) q is always less than p

C) q=100%-p

D) p and q are always equal

E) p+q=1.96

Answer: C

20) Consider that we have nominal data (responses are categorical) and the responses are

"Yes" or "No" to the question: "The next time you order pizza, will you use Domino's?"

Which of the following sets of responses shows the MOST variability?

A) 90 percent say "Yes" and 10 percent say "No"

B) 80 percent say "Yes" and 20 percent say "No"

C) 70 percent say "Yes" and 30 percent say "No"

D) 60 percent say "Yes" and 40 percent say "No"

E) 55 percent say "Yes" and 45 percent say "No"

Answer: E

21) Which of the following is the theory that allows us to say that if we conducted a survey

1,000 times, and we were to plot the answers to our survey, the plot would appear as a normal

curve?

A) the central limit theorem

B) the normal curve theory

C) the normal limit theorem

D) the confidence interval theorem

E) the variability coefficient theory

Answer: A

22) 95 percent of the observations under the normal curve fall within ________ times the

sample error.

A) ±1.64

B) ±1.96

C) ±2.58

D) ±95

E) ±1.95

Answer: B

23) Sample size is related to the size of the confidence interval in that:

A) the larger the sample size, the larger the confidence interval.

B) the larger the sample size, the more normal the confidence interval.

C) the smaller the sample size, the smaller the confidence interval.

D) the smaller the sample size, the more uniform the confidence interval.

E) the larger the sample size, the smaller the confidence interval.

Answer: E

24) The only time the population size is important in the calculation of sample size is:

A) always; all formulas include N, a count of the total population.

B) when the population is very, very large.

C) when the population is not normal.

D) when the level of accuracy needs to be less than ±5 percent.

E) when the population is small, relative to the sample size.

Answer: E

25) National opinion polls tend to use sample sizes ranging from:

A) 10 to 100

B) 1,000 to 1,200

C) 50,000 to 100,000

D) 1 million to 5 million

E) 10 million to 15 million

Answer: B

26) What three factors are needed to calculate sample size?

A) variability, accuracy, and confidence level

B) variability, accuracy, and population size

C) accuracy, confidence level, and population size

D) accuracy, population size, and costs

E) variance, standard deviation, and dispersion

Answer: A

27) If we are using the formula for calculating the sample size for estimating a percentage,

the formula will contain:

A) s.

B) %.

C) Z3.

D) pq.

E) N.

Answer: D

28) In sample size formulas, the symbol "e" stands for:

A) estimated parameter or desired confidence level.

B) acceptable population equation.

C) estimated statistic or desired variability.

D) estimated confidence interval.

E) acceptable sample error.

Answer: E

29) Level of confidence in sample size formulae is normally set at:

A) 95 percent or 96 percent.

B) 5 percent or 10 percent.

C) 95 percent or 99 percent.

D) 1 percent or 5 percent.

E) None of the above; level of confidence is determined by the formula.

Answer: C

30) If you were conducting a telephone survey of households using random digit dialing

numbers and you determined that you needed a sample size of 1,100, which of the following

would be most accurate?

A) You would need to obtain exactly 1,100 telephone numbers to call.

B) You would need to obtain far less than 1,100 telephone numbers because you know that

many will not answer or cooperate anyway.

C) You will need some multiple of the 1,100 numbers in order to ensure you account for

factors such as numbers that are for business phones, ineligible households (incidence rate),

and those numbers dialed whose owners refuse to participate.

D) You will need only a few extra numbers to allow for those who have moved away.

E) You start off with 1,100; there is no way to determine approximately how many numbers

you will actually need.

Answer: C

31) If we assume the "highest" amount of variability when estimating pq, then pq are:

A) 50, 50.

B) 1, 99.

C) 0, 10.

D) 1, 5.

E) 0, 5.

Answer: A

32) In trying to estimate the variability in the population in order to determine pq, which of

the following represents viable alternatives?

A) Use a random number generator to provide two random values for p and q.

B) Use the most conservative approach, p = 10, q = 90.

C) Use 90/10 or find a former study and calculate the variance or conduct a pilot study.

D) Use 1/5 or find a former study and calculate the variance or conduct a pilot study.

E) Use 50/50 or find a former study and calculate the variance or conduct a pilot study.

Answer: E

33) Which of the following statements best illustrates the concept of the "level of confidence"

or z value chosen in sample size formulae?

A) By setting z at 1.96 it means that the manager could expect, if she conducted the survey

many, many times, the value of p would fall within the sample error range 95 percent of the

time.

B) By setting z at 1.96 it means that the manager could expect, if she conducted the survey

many, many times, the value of p would fall within the sample error range 5 percent of the

time.

C) By setting z at 2.58 it means that the manager could expect, if she conducted the survey

many, many times, the value of p would fall within the sample error range 95 percent of the

time.

D) By setting z at 1.96 it means that the manager could expect, if she conducted the survey

many, many times, the value of p would fall within the sample error range 99 percent of the

time.

E) By setting z at 2.58 it means that the manager could expect, if she conducted the survey

many, many times, the value of p would fall within the sample error range 1 percent of the

time.

Answer: A

34) If we are using the formula for calculating the sample size for estimating a mean, the

formula will contain:

A) s.

B) %.

C) Z3.

D) pq.

E) None of the above; there is no sample size formula for estimating a mean.

Answer: A

35) If we are using the sample size formula to be used when estimating a mean and we are

trying to determine s, we are trying to indicate:

A) the pq population mean.

B) the variability in the population represented by the standard deviation.

C) the sample size.

D) the allowable error represented by the standard deviation.

E) None of the above; the formula does not have an s.

Answer: B

36) If we are using the sample size formula to be used when estimating a mean and we are

trying to determine e, we are trying to indicate:

A) the allowable error.

B) the proportion of s to S in the sample.

C) the range of possible z scores in the population.

D) the percentage of the respondents who will answer p.

E) the percentage of the respondents who will answer q.

Answer: A

37) When attempting to balance the sample size with the cost of data collection, the textbook

illustrated that it is helpful to:

A) use an Excel spreadsheet.

B) rely on financial leveraging.

C) use a table that depicts data collection cost and sample error for different sample sizes.

D) use a table that depicts data collection cost and the costs of computing different sample

sizes.

E) use a table that depicts data collection cost contrasted to the cost of not doing research.

Answer: C

38) When determining sample size, the conventional approach would NOT be:

A) an average of the sample sizes of similar studies.

B) the modal sample size of previous surveys.

C) 5 percent of the entire population.

D) the sample size of a competitor's survey that the company somehow discovered.

E) the sample sizes normally reported in published reports.

Answer: C

39) The basic difference between an arbitrary and a conventional sample size determination

is that:

A) the conventional approach has no defensible logic, whereas the arbitrary approach appears

to have faulty logic.

B) the conventional approach doesn't appear to be logical, whereas the arbitrary approach has

defensible logic.

C) the arbitrary approach has no defensible logic, whereas the conventional approach is

logical.

D) the arbitrary approach has no defensible logic, whereas the conventional approach appears

logical but is faulty.

E) the arbitrary approach, though arbitrary, has sound logic, and the conventional approach

has conventional support.

Answer: D

40) A sample that has been determined by using a cost basis approach would be when:

A) the manager has discussed the statistical analysis with the research project director.

B) the manager has discussed the competitor's marketing results with the research project

director.

C) the manager has discussed the previous marketing study with the research project director.

D) the manager has discussed the budget with the research project director and they have

decided to spend "all they can afford" on the project.

E) none of the above

Answer: D

41) Which of the following samples have been determined by using the statistical analysis

approach?

A) the sample size needed to properly analyze subgroups

B) 1,000 respondents

C) between 1,000 and 1,200 respondents

D) 200 respondents because each interview is $30

E) a former study's sample, which generated favorable statistics

Answer: A

42) A finite multiplier is used when the population, relative to the sample size, is:

A) small.

B) large.

C) finite.

D) variable.

E) infinite.

Answer: A

43) By applying the finite multiplier it is possible to:

A) reduce the population and achieve the same accuracy level.

B) reduce the sample size or achieve the appropriate accuracy level.

C) reduce the sample size and achieve the same accuracy level.

D) reduce the accuracy level and maintain the same sample size.

E) None of the above; there is no "finite multiplier."

Answer: C

44) You are going to use a purposive sample (a nonprobability sample) to collect data from

fast-food restaurant customers. Which of the following concepts would be applicable to

determining sample size?

A) confidence intervals

B) sample size formula for estimating a percentage

C) sample size formula for estimating a mean

D) estimating a population value within a stated percent of allowable error

E) None of the above would be appropriate; sample size concepts and formulas are only

applicable when probability sampling plans have been used.

Answer: D

45) Sample size does not have to be "huge" in order to have reasonably accurate data. For

example, samples that are 400 or less may provide reasonably accurate information.

Answer: True

46) The sample size is always related to how representative the sample is of the population.

Answer: False

47) Instead of determining representativeness, the size of the sample affects the sample

accuracy of the results.

Answer: True

48) Sample size has a direct bearing on how accurate the sample's findings are relative to the

true values in the population.

Answer: True

49) Sample accuracy refers to how close a random sample's statistic is to another random

sample statistic drawn on the same population. If the two samples result in the same, or

nearly the same, data, then one has achieved sample accuracy.

Answer: False

50) Large sample size bias refers to a belief that sample size determines a sample's

representativeness.

Answer: True

51) The only perfectly accurate sample is a census.

Answer: True

52) The size of a probability sample depends on the client's desired accuracy (acceptable

sample error) balanced against the cost of data collection for that sample size.

Answer: True

53) The confidence interval approach to determining sample size applies the concepts of

accuracy, variability, and confidence interval to create a "correct" sample size.

Answer: True

54) Sampling error is the difference between the sample findings and the findings that the

client expected to have prior to the survey.

Answer: False

55) Nonsampling error pertains to all sources of error other than the sample selection method

and sample size.

Answer: True

56) A random sample must be perfectly accurate to be considered a very good representation

of the population.

Answer: False

57) A graph of the relationship between accuracy and size of the sample shows that accuracy

decreases as sample size increases.

Answer: False

58) A graph of the relationship between accuracy and size of the sample shows that accuracy

increases dramatically up until sample sizes near 500. Thereafter, it takes much larger

increases in sample size to gain increases in accuracy.

Answer: True

59) Sample error may only be determined after data have been collected. Even if we know

the level of confidence, an estimate of variability (p∗q) and the sample size, n, we cannot

compute the amount of sample error we can expect to experience.

Answer: False

60) Variability is defined as the amount of dissimilarity or similarity in respondents' answers

to a particular question.

Answer: True

61) In a "Yes/No" question, 50 percent "Yes" and 50 percent "No" shows less variability than

does 90 percent "Yes" and 10 percent "No."

Answer: False

62) The axiom, "You can take any finding in the survey, replicate the survey with the same

probability sample size, and you will be very likely to find the same finding within the ±

percent range of the original finding," is based on the idea of the confidence interval.

Answer: True

63) The basic concept of the central limit theorem is that all research findings are limited in

their application simply because they are based upon a sample and not a census.

Answer: False

64) The most significant factor in calculating sample size (n) is the size of the population (N).

For example, if you calculated that you needed a sample size of 300 in order to have an

accurate sample for your hometown, it would take at least 10 times this amount, or about

3,000, in order to have an accurate sample representing the entire United States.

Answer: False

65) A probability sample size can be a very tiny percent of the population size and still be

very accurate (have little sample error).

Answer: True

66) The only time that the population size is a consideration in sample size determination is

in the case of a "small population."

Answer: True

67) Using the confidence interval formula for calculating sample size, the amount of

variability believed to be in the population must be estimated.

Answer: True

68) Using the confidence interval formula for calculating sample size, the level of confidence

desired to the estimate of the population values must be determined and is represented in the

formula by the z value.

Answer: True

69) The desired accuracy should NOT be considered in order to calculate the proper sample

size. This is because accuracy cannot be determined until the data are collected and analyzed.

Answer: False

70) In sample size formulae, acceptable sample error is noted by "e."

Answer: True

71) It is almost always up to the researcher to educate the manager on what might be

acceptable or "standard" sample error.

Answer: True

72) In determining the level of confidence, any level of confidence is possible, but marketing

researchers typically use between 60 percent and 100 percent , depending on the importance

of the issue and how much the client has to spend on the research.

Answer: False

73) If you calculate sample size with an e of 3 percent, and then you decide to change e to 5

percent, the required sample size will go up.

Answer: False

74) With the 99% level of confidence, the corresponding z value is 2.58.

Answer: True

75) It is "true" to say that, in practice, you can estimate variability in the population by using

one of three methods: (a) set pq to the most conservative amounts of 50 percent/50 percent,

(b) estimate pq based upon prior research studies, or (c) conduct a pilot study.

Answer: True

76) There is a separate sample size formula for estimating a percentage of the population than

for estimating a mean of the population.

Answer: True

77) Variability (standard deviation) of a population where a mean is being estimated may be

estimated by dividing the range by 6.

Answer: True

78) Researchers should follow to the sample size formula closely when calculating sample

size. No other consideration, such as the cost to the client, should be considered, as it will

surely lower the accuracy of the study.

Answer: False

79) The arbitrary approach to sample size determination takes the desired level of accuracy

into account.

Answer: False

80) The conventional approach to sample size determination follows some convention or

number believed somehow to be the "correct" sample size.

Answer: True

81) It is important that marketing researchers try to educate managers/clients on the

determinants of sample size as an ethical issue arises when the researcher benefits by using

larger samples.

Answer: True

82) Generally, a small population situation is one in which the sample exceeds 5 percent of

the total population size.

Answer: True

83) A finite multiplier is an adjustment factor that should be used when the sample size is

small, relative to the population size.

Answer: True

84) It is inappropriate to apply sample size formulae when an unknown bias is introduced by

a subjective sampling method (i.e., nonprobability).

Answer: True

85) When using nonprobability samples instead of probability samples, different formulas

must be applied to determine the appropriate sample size that precisely balances desired

accuracy, variability in the population, and level of confidence.

Answer: False

86) Information provided to clients cannot accurately represent likely outcomes and results

compared to alternative available methodologies.

Answer: False

87) An amusement park owner is considering a survey to determine customer preferences for

a new water ride. The owner and the researcher discuss the concept of the level of accuracy.

The owner accepts a level of accuracy of ±5 percent. Assuming that the survey finds that 70

percent of the survey respondents indicate they want the ride, what does having a level of

accuracy of ±5 percent accuracy actually mean?

A) that there will be a 5 percent chance that the owner will make the wrong decision as to

whether or not to build the new water ride

B) that the real percentage of the park's customers who prefer the new water ride falls

between 0 and 5 percent

C) that there is a 95 percent chance that the owner will make the right decision as to whether

or not to build the new water ride

D) that the real percentage of the park's customers who prefer the new water ride falls

between 65 and 75 percent

E) that there is a 5 percent chance, ±, that the owner will make the right decision

Answer: B

88) Jack McCombs is the owner of several Firehouse Subs sandwich shops in Anytown. He

has been spending $200,000 a year in various media in Anytown in an attempt to build

awareness of his stores. Instead of continuing to spend the same amount on advertising every

year, Jack wants an assessment as to what he has gained from the advertising he has paid for.

He is interested in knowing what percentage of the population in Anytown is aware of his

store's name, menu, and locations — three factors for which he has attempted to build

awareness. His advertising agency quoted him a price for a survey and told him they would

use a sample size of 150, and that they were assuming 35 percent of the respondents would be

aware of his advertising. Jack was reluctant to use the advertising agency to conduct the

survey. He felt like he needed another firm so as to avoid any conflict of interest. He found a

marketing researcher in town who was certified as a CPR by the Marketing Research

Association. The CPR asked Jack how accurate he wanted the results of the survey to be.

Jack said he wanted the percentage to be within ±3 percent of the real population percentage.

Assuming he wanted to be 95 percent confident of the accuracy of the results, the CPR used

the following formula to determine the accuracy of the survey recommended by the ad

agency: ± sample error percent = 1.96 ∗ the square root of pq/n. Using the formula, what

would be the sample error if Jack were to use the ad agency survey recommendation?

A) ±3 percent

B) ±4 percent

C) ±5 percent

D) ±6 percent

E) ±7 percent

Answer: E

89) Political Research Associates has been hired to conduct a survey to determine the

percentage of voters. If the election were held today, who would vote for Candidate X for

president? Candidate X has not even announced her candidacy for president and there are no

previous surveys that would indicate voter preferences. Still, Political Research Associates

must estimate the variability in the population in order to determine the size of the sample

they need for their survey for Candidate X. Which of the following would be the wisest

choice for estimating variability?

A) ±5 percent

B) p = 1.96

C) q = 100 percent, p = 50 percent

D) p = 50 percent, q = 50 percent

E) n = 51 percent, q = 49 percent

Answer: D

90) Zoom-IT is a high-technology firm specializing in electronic consumer products. In

particular, the company has worked on expansion systems for LCD and plasma screens. The

company has developed a method for expanding the 2-inch screen on Apple iPod video

display units to 8 inches by increasing the weight of the iPod by only 3 ounces. The "pop out"

screen relies on Space Age plastics and an electronic expansion system that works similar to

camera shutters. The only drawback is that the add-on device costs an additional $350, more

than the cost of the iPod. Before going into production and marketing, Zoom-IT CEO Jane

Ellen Roberts, decides she wants some evidence that the device will sell. She commissions

the marketing research department to conduct a survey and the critical question asks for

likelihood to purchase the product on a 10-point scale. The researchers, attempting to

determine the ideal sample size, realize there is no former study on this issue on which they

could estimate the standard deviation in the population. They are considering the expense of a

small pilot test, but researchers James Hughes and Bennet Alford make a reasonable

recommendation to:

A) use 50/50.

B) use the standard deviation from a study they conducted on expanding plasma screens from

40 inches to 90 inches.

C) use a table of random numbers to generate a random standard deviation.

D) divide the 10 scale points by 6 because ±3 (or 6) standard deviations covers the range of

observations (10) and by dividing by 6 will yield a reasonable estimate of 1 standard

deviation which they may use for s.

E) None of the above; you do not need to estimate standard deviation in the sample size

formula for the mean.

Answer: D

91) A study is to be performed for a local restaurant, McGuire's, and Mr. McGuire wants to

know the awareness of the restaurant name as well as satisfaction with food, service, and

prices. Furthermore, being in the restaurant business for many years, Mr. McGuire is certain

that he has very different clientele depending on which meal and whether they are there

during the weekdays or weekends. Therefore, he wants to know answers to these issues by

subgroups such as those who have eaten lunch meals, dinner meals, weekday patrons, and

weekend patrons. Given these analysis goals, which sample size approach should be

considered?

A) arbitrary approach

B) conventional approach

C) statistical analysis approach

D) confidence percentage approach

E) the cost basis approach

Answer: C

92) Dot Miller owns a small chain of dive shops. She is interested in knowing how many

potential customers she will need to sample in several large cities within a three- to four-hour

drive of her shops located on the California coast. She wants to estimate the mean response to

a 10-point scale measuring the likelihood that they will subscribe to a dive package which, if

they take advantage of all the dives, represents a savings of 50 percent over taking the dives

individually. In estimating the standard deviation in the population for the formula for

calculating the sample size for estimating a mean, Dot can:

A) rely on the knowledge of other owners in the dive industry to share their knowledge of

surveying scuba divers.

B) use some prior knowledge about the population, undertake a pilot study or estimate the

range, and divide by 10.

C) use some prior knowledge about the population, undertake a pilot study or estimate the

range (7), and divide by 6.

D) undertake a pilot study using members of the dive population who live near the present

dive shops.

E) None of the above; if you do not have the actual standard deviation, you cannot use the

formula.

Answer: C

Test Bank for Marketing Research

Alvin C. Burns, Ronald F. Bush

9780133074673, 9780134895406, 9780134167404