Questions: Final Review

Question 1

When certain social, economic, or demographic groups don’t cooperate with the census process and escape measurement, the measures obtained from the census will have _____ than the measures obtained from a proper representative sample

A. Lower Systematic Error
B. Higher Systematic Error
C. Lower Random Error
D. Higher Random Error

Answer 1

B. Higher Systematic Error

Systematic Error (Bias): Occurs when the error-producing factor is stable and operates in a constant direction. Bias is the worst type of error because it cannot be quantified or even identified. Sampling procedures are designed to avoid it.

Question 2

For any given population, a representative sample will collect ____ than a census (on the same population)

A. A Smaller Number of Responses
B. A Larger Number of Responses
C. The Same Number of Responses
D. There is no general answer to this question

Answer 2

A. A Smaller Number of Responses

Census: The act of measuring the relevant characteristics of ALL the individuals in a population. It’s expensive, time-consuming, technically demanding, and intrusive. No census is perfect: When a census misses a number of individuals, it tends to do so systematically (as opposed to randomly), which introduces serious bias in the results.

Question 3

For a research project that aims at studying the number of hours US children spend daily playing video games, the target population may be defined as:

A. Video Games
B. Households
C. US Residents
D. Children

Answer 3

D. Children

Target Population: The aggregate of all the individuals whose characteristics are the focus of the marketing research project.

Question 4

A complete list of population members is called:

A. Mailing List
B. Target Population
C. Sampling Frame
D. Sample

Answer 4

C. Sampling Frame

Sampling Frame: Includes the whole population to be sampled. It consists of a list of all members of the population to be sampled or a rule identifying them.

Question 5

A probability sample is:

A. Biased
B. Representative
C. Large
D. Exploratory

Answer 5

B. Representative

Probability Sample: Used when representativeness is important, for example in a descriptive survey. The selection of sampled individuals must be random, and the chance of picking a particular sampling unit in the population must be quantifiable.

Question 6

Which of the following is a sampling frame?

A. The definition of the target market or target population
B. The calculation of sample
C. A rule for identifying all population members
D. The project framework

Answer 6

C. A rule for identifying all population members

Sampling Frame: Includes the whole population to be sampled. It consists of a list of all members of the population to be sampled or a rule identifying them.

Question 7

A sample is a _____ when all members of the population have a known chance of being sampled

A. Non-Probability Sample
B. Cluster Sample
C. Census
D. Probability Sample

Answer 7

D. Probability Sample

Probability Sample: Used when representativeness is important, for example in a descriptive survey. The selection of sampled individuals must be random, and the chance of picking a particular sampling unit in the population must be quantifiable.

Question 8

Non-probability samples are useful when:

A. The research project is descriptive, for example a representative survey
B. The research project is exploratory, for example a pilot survey
C. They are never useful
D. They are always useful, as long as they are large enough

Answer 8

B. The research project is exploratory, for example a pilot survey

Non-Probability Samples: Are not representative, but they may be useful for exploratory purposes.

Question 9

Simple Random Sampling (SRS) relies on:

A. Small Sample Size
B. Simple Sample Structure
C. Random Numbers
D. Random Sampling Costs

Answer 9

C. Random Numbers

Simple Random Sample: Random numbers (ex: lottery); hard to do since you need a sample frame in the form of a list.

Question 10

Generally, a cluster sample is ____ than a SRS

A. More Expensive
B. Less Expensive
C. More Accurate
D. More Representative

Answer 10

B. Less Expensive

Cluster Sample: The population is divided into subsets (clusters), and a sample of clusters is created. The process may be repeated with a different level of clusters, until we have a few clusters that can be specified in terms of the original sampling units. At this point, the final sample is drawn by SRS. Cluster sampling is a way to reduce sampling costs, but it has comparatively high sampling error (less precision, or low statistical efficiency).

Question 11

Generally, a stratified sample is ____ than a SRS

A. More Expensive
B. Less Expensive
C. More Accurate
D. A and C

Answer 11

D. More Expensive AND More Accurate

Stratified Sample: The population is divided into strata (such as segments), and each stratum is sampled separately. This increases precision (high statistical efficiency), but can be costly: the more strata, the higher the cost.

Question 12

A Quota Sample is also:

A. Stratified Sample
B. Simple Random Sample
C. Large Sample
D. Non-Probability Sample

Answer 12

D. Non-Probability Sample

Non-Probability Samples: Are not representative, but they may be useful for exploratory purposes.

Quota Sample: Selection of sampled individuals is based on both the goal of including certain subgroups, and the judgment of the interviewer.

Question 13

A statistic calculated from a probabilistic sample can be expected to differ from the “true” population value because of:

A. Systematic Error
B. Random Error
C. Poor Execution
D. Wrong Sampling Technique

Answer 13

B. Random Error

Random Error (Sampling Error): Occurs because a sample does not include all population members and as a result may underestimate or overestimate the true average measure of a characteristic in the population. Its randomness follows probabilistic processes that allow us to quantify its incidence. Sampling procedures are designed to maintain it within acceptable bounds.

Question 14

What is probably true of a non-probability sample?

A. It’s Small
B. It’s Large
C. It’s Expensive
D. It’s Biased

Answer 14

D. It’s Biased

Non-Probability Samples: Are not representative, but they may be useful for exploratory purposes.

Question 15

Suppose you know that Gender and Income are equally related to the measures you want to study. For your project, it’s especially important that you reduce sampling error and increase the accuracy of your sampling estimates. You have some financial flexibility, but not infinite resources, you can:

A. Use stratified sampling, with Gender as the stratification variable
B. Use stratified sampling, with Income as the stratification variable
C. Use cluster sampling, with counties as the first level clusters
D. Use cluster sampling, with Income as a clustering variable

Answer 15

A. Use stratified sampling, with Gender as the stratification variable

Stratified Sample: The population is divided into strata (such as segments), and each stratum is sampled separately. This increases precision (high statistical efficiency), but can be costly: the more strata, the higher the cost.

Question 16

You need to build a representative sample of your target market but don’t have a list of all population members. You then decide to divide the market by ZIP codes, and select a sample of ZIP codes. Subsequently, you get a list of city blocks doe each of the sampled ZIP codes, build a sample of city blocks. Finally, you will develop a list of individuals who live in this sample of city blocks, and build a final sample of individuals. This is an example of:

A. Simple Random Sampling
B. Judgement Sampling
C. Stratified Sampling
D. Cluster Sampling

Answer 16

D. Cluster Sampling

Cluster Sample: The population is divided into subsets (clusters), and a sample of clusters is created. The process may be repeated with a different level of clusters, until we have a few clusters that can be specified in terms of the original sampling units. At this point, the final sample is drawn by SRS. Cluster sampling is a way to reduce sampling costs, but it has comparatively high sampling error (less precision, or low statistical efficiency).

Question 17

Sample representativeness depends on:

A. Random selection from the target population
B. Adequate inclusion of all different groups, based on gender, income, etc.
C. Large sample size
D. Elimination of sampling error

Answer 17

A. Random selection from the target population

Sample representativeness depends on random selection from the target population.
The target population is the aggregate of all the individuals whose characteristics are the focus of the marketing research project. For example, when we study a target market, all the individuals in that target market are the target population. These individuals are called sampling units.

Question 18

A sample result is representative if:

A. The sample is large
B. The sample is precise
C. The sample is probabilistic
D. The sample is based on stratification

Answer 18

C. The sample is probabilistic

Samples that avoid bias are called probabilistic samples, and are representative. Sample representativeness depends on random selection from the target population.

Question 19

The population mean is a _____ that we try to estimate with the corresponding sample _____.

A. Interval, Number
B. Statistic, Parameter
C. Parameter, Statistic
D. Calculation, Deviation

Answer 19

C. Parameter, Statistics

Population Mean (Arithmetic Average): Parameter that we try to estimate with corresponding sample statistics.

Question 20

If you wan to estimate the percentage of brand X users in the total market, you will calculate the _____ of brand X users in the sample.

A. Size
B. Confidence
C. Average Consumption
D. Proportion

Answer 20

D. Proportion

The sample proportion is the statistic for the population proportion. If you want to estimate the % of brand-X users in the total market, you will calculate the proportion of brand X users in the sample.

Question 21

The symbol X stands for:

A. Sample Mean
B. Population Mean
C. Sample Standard Deviation
D. Population Standard Deviation

Answer 21

A. Sample Mean

The sample mean approximates the population mean according to a “normal probability distribution” (bell curve, symmetrical around the center, the center being the population mean) This fact is used to set a desired level of confidence

Question 22

The symbol μ stands for:

A. Sample Mean
B. Population Mean
C. Sample Standard Deviation
D. Population Standard Deviation

Answer 22

B. Population Mean

The population mean (arithmetic average) is a parameter that we try to estimate with corresponding sample statistics.

Question 23

The symbol σ stands for:

A. Sample Mean
B. Population Mean
C. Sample Standard Deviation
D. Population Standard Deviation

Answer 23

D. Population Standard Deviation

The sample standard deviation is the statistic for the population standard deviation.

Question 24

The symbol s stands for:

A. Sample Mean
B. Population Mean
C. Sample Standard Deviation
D. Population Standard Deviation

Answer 24

C. Sample Standard Deviation

The sample standard deviation is the statistic for the population standard deviation

Question 25

The symbol X stands for:

A. Random variable X, i.e. any individual measure for this variable
B. Sample Mean
C. Population Mean
D. B and C

Answer 25

A. Random variable X, i.e. any individual measure for this variable

Question 26

The sample mean has:

A. Only one possible value, equal to the population mean
B. Only two possible values
C. A range of possible values
D. No value

Answer 26

C. A range of possible values

The sample mean approximates the population mean according to a “normal probability distribution” (bell curve, symmetrical around the center, the center being the population mean) This fact is used to set a desired level of confidence

Question 27

Which of the following is true:

A. Population mean has a sampling distribution
B. Sample mean has a sampling distribution
C. Population mean is usually known
D. Sample mean is usually unknown

Answer 27

B. Sample mean has a sampling distribution

The sample mean approximates the population mean according to a “normal probability distribution” (bell curve, symmetrical around the center, the center being the population mean) This fact is used to set a desired level of confidence

Question 28

The standard error of the sample mean is an indication of:

A. Whether you should use a sample
B. Whether you need to use a sample
C. How variable the individual measures are in the population
D. How variable the possible values of the mean at a given sample size

Answer 28

D. How variable the possible values of the mean at a given sample size

The standard error of the mean gives a measure of the variability of the sample mean. The standard error of the mean represents an average measure of sampling error, i.e. of the “average” difference between all possible sample means and the population mean.

Question 29

To calculate the sample size you need to have the following inputs

A. Desired precision of the estimate
B. Desired confidence of the estimate
C. A measure of variability of the estimate
D. All of the above

Answer 29

D. All of the above

Question 30

The Confidence-Interval approach to sample size calculation is based on:

A. Population mean = sample mean +/- margin of error
B. Margin of error = (z)*(standard error of the sample mean)
C. z = score that represents a certain level of confidence
D. All of the above

Answer 30

D. All of the above

Confidence Level: The probability that the confidence interval includes the true population value. The higher the better (typically levels 99%, 95%, 90%). Represented by the value for the variable Z in the confidence interval

Question 31

When we use a z=1.96 (approximately 2), we incorporate a 95% confidence level in the estimated confidence interval. That means that we expect the following:

A. The sample mean is equal to 95% the population mean
B. The population mean is correct 95% of the times
C. There is 95% chance that the estimated confidence interval contains the population mean
D. There is 5% chance that the sample mean and the population mean have different values

Answer 31

C. There is 95% chance that the estimated confidence interval contains the population mean

Confidence Level: The probability that the confidence interval includes the true population value. The higher the better (typically levels 99%, 95%, 90%). Represented by the value for the variable Z in the confidence interval

Question 32

Everything else being equal, a larger margin of error makes for:

A. Better estimate
B. Worse estimate
C. Smaller sample
D. More precise estimate

Answer 32

B. Worse estimate

Precision/Accuracy – maximum size of margin of error; managerially decided (smaller margin of error = higher precision)

Question 33

What is a normal probability distribution?

A. Random distribution that is symmetrical around the mean (and bell-shaped)
B. Type of confidence interval
C. Statement that is normally true
D. Normal distribution that uses probabilities on the x-axis

Answer 33

A. Random distribution that is symmetrical around the mean (and bell-shaped)

The sample mean approximates the population mean according to a “normal probability distribution” (bell curve, symmetrical around the center, the center being the population mean)
This fact is used to set a desired level of confidence.

Question 34

Everything else being equal, a confidence interval with a 90% confidence level is:

A. Used only for stratified sampling
B. It doesn’t exist
C. Worse that a confidence interval with 95% confidence level
D. Better than a confidence interval with a 95% confidence level

Answer 34

C. Worse that a confidence interval with 95% confidence level

Confidence Level: The probability that the confidence interval includes the true population value. The higher the better (typically levels 99%, 95%, 90%). Represented by the value for the variable Z in the confidence interval

Question 35

You calculated the needed sample size based on an educated guess for the standard deviation. What happens then if the true standard deviation is greater than your guess?

A. Larger sample mean than expected
B. Larger margin of error than expected
C. Smaller sample mean than expected
D. Smaller margin of error than expected

Answer 35

B. Larger margin of error than expected

Question 36

The estimate obtained from a probabilistic sample of 1,000 respondents is _____ than one from a sample of 500 respondents

A. More Numerous
B. More Difficult
C. More Accurate
D. More Certain

Answer 36

C. More Accurate

Samples that avoid bias are called probabilistic samples, and are representative.

Question 37

A probabilistic sample of 1,000 respondents is _____ than one of 500 respondents

A. More Certain
B. More Confident
C. More Elaborate
D. More Expensive

Answer 37

D. More Expensive

Samples that avoid bias are called probabilistic samples, and are representative.

Question 38

How do you guess the standard deviation in the population when you need to calculate the needed sample size?

A. Look at published sources
B. Ask for expert input
C. Look at prior estimates
D. All of the above

Answer 38

D. All of the above

Question 39

The _____ is NOT a measure of central tendency (or concentration)

A. Median
B. Standard Deviation
C. Mean
D. Mode

Answer 39

B. Standard Deviation

Measures of Central Tendency: Mean, Median, Mode

Question 40

The _____ can only be calculated from interval or ratio data

A. Mean
B. Mode
C. Median
D. Proportion

Answer 40

A. Mean

Mean: arithmetic mean; used to evaluate interval or ratio data, NOT categorical or ordinal

Question 41

The ____ is used to summarize ordinal data

A. Sample Mean
B. Population Mean
C. Sample Median
D. Population Mode

Answer 41

C. Sample Median

Question 42

A one-way table can be used to efficiently summarize responses to:

A. A single interval variable
B. Two interval variables
C. A single categorical variable
D. Two categorical variables

Answer 42

C. A single categorical variable

Tabulation is best used for categorical or ordinal variables. Its use and interpretation becomes difficult and inefficient for interval and ratio data. A frequency distribution in tabular form is also known as a one-way table: it lists all the values, counts, etc. for a single variable.

Question 43

In a two way table, all the row percents sum up to 100%

A. Across the entire table
B. For each column
C. For each row
D. In a two way table, percent figures don’t sum up to 100%

Answer 43

C. For each row

Question 44

How can you summarize variability in a measure?

A. Mean or median
B. Standard deviation and range
C. Mean and mode
D. A proportion

Answer 44

B. Standard deviation and range

Question 45

What is a statistically significant difference?

A. Large difference
B. Small difference
C. Difference that is probably due to sampling error
D. Difference that is probably not due to sampling error

Answer 45

D. Difference that is probably not due to sampling error

If a correlation coefficient has a p-level equal or lower than the significance level, it is statistically significant

Question 46

By design, in hypothesis testing, rejecting the null hypothesis by mistake is _____ rejecting the research hypothesis by mistake

A. More important than
B. Less important than
C. As important as
D. Both are not important

Answer 46

A. More important than

The null hypothesis and the research (or alternative) hypothesis/ H0 (the null hypothesis) represents the status quo (the conservative choice), and is typically more protected than the alternative hypothesis from a possible mistaken rejection

Question 47

You are testing whether the new product performs better than a similar existing product in terms of sales. How do you formulate the null hypothesis?

A. H0: Sales for the new product are higher than sales for the old product
B. H0: Sales for the new product are at best equal to sales for the old product
C. H0: Sales for the new product are very good
D. H0: Sales for the new product are not very good

Answer 47

B. H0: Sales for the new product are at best equal to sales for the old product

The null hypothesis and the research (or alternative) hypothesis/ H0 (the null hypothesis) represents the status quo (the conservative choice), and is typically more protected than the alternative hypothesis from a possible mistaken rejection

Question 48

The level of significance in a conservative hypothesis test would be set at:

A. 0.5
B. 0.005
C. 0.1
D. 0.01

Answer 48

D. 0.01

In marketing, we generally choose a significant level of 0.05. If the risk from a failure of new formulation was major a conservative level of 0.01 would be used

Question 49

You test a new commercial to evaluate its performance in terms of increasing brand liking in the target market. Your Null hypothesis is:

A. H0: Brand liking changed significantly after airing the ad
B. H0: Brand liking did not change after airing the ad
C. H0: Brand liking increased after airing the ad
D. H0: Brand liking did not increase after airing the ad

Answer 49

D. H0: Brand liking did not increase after airing the ad

The null hypothesis and the research (or alternative) hypothesis/ H0 (the null hypothesis) represents the status quo (the conservative choice), and is typically more protected than the alternative hypothesis from a possible mistaken rejection

Question 50

In a cross table, you have area of residence (A, B) on the rows and sunscreen usage status (light, medium, heavy) on the columns. In a random sample, you find that only 20% of A residents are heavy users, and 80% of A residents are light/medium users. For region B residents, you have that 30% are heavy users, while 70% of B residents are light/ medium users. What is the first question a researcher would ask about these figures?

A. Is the difference between 30% and 20% managerially important?
B. Is the difference between 30% and 20% statistically significant?
C. Is there a difference between light and medium users?
D. Is the difference between 20% and 80% meaningful?

Answer 50

B. Is the difference between 30% and 20% statistically significant?

Question 51

In a cross table, you have area of residence (A, B) on the rows and sunscreen usage status (light, medium, heavy) on the columns. In a random sample, you find that only 20% of A residents are heavy users, and 80% of A residents are light/medium users. For region B residents, you have that 30% are heavy users, while 70% of B residents are light/medium users. You speculate that:

A. Medium users and heavy users are different
B. Heavy users have a higher chance of being residents of B than light/med. users
C. Residents of B have a higher chance of being heavy users than residents of A
D. Residents of B have a higher chance of being heavy users instead of light/med. Users

Answer 51

C. Residents of B have a higher chance of being heavy users than residents of A

Question 52

In a cross table, you have area of residence (A, B) on the rows and sunscreen usage status (light, medium, heavy) on the columns. In a random sample, you find that only 20% of A residents are heavy users, and 80% of A residents are light/medium users. For region B residents, you have that 30% are heavy users, while 70% of B residents are light/medium users. If you propose that area of residence and usage status are related, and test this idea with the data above, the corresponding null hypothesis is:

A. H0: Usage status does not differ by area of residence
B. H0: Usage status differs by area of residence
C. H0 : Heavy users are more frequent than light/medium users
D. H0: Heavy users are less frequent than light/medium users

Answer 52

A. H0: Usage status does not differ by area of residence

Question 53

The significance level (alpha) in hypothesis testing represents the chance of:

A. Rejecting the research hypothesis when it is true
B. Accepting the research hypothesis when it is false
C. Rejecting the null hypothesis when it is true
D. Two of the above

Answer 53

D. Two of the above

Question 54

If sampling data shows a difference between segments, this difference may be managerially important if:

A. It’s predictable
B. It’s not predictable
C. It’s arithmetically significant
D. It’s statistically significant

Answer 54

D. It’s statistically significant

Question 55

A test of hypothesis with a significance level equal to 0.05 has _____ than if it uses a significance level of 0.01

A. A higher chance of committing a type-I error
B. A lower chance of committing a type-I error
C. The same chance of committing a type-I error
D. A higher chance of committing a type-III error

Answer 55

A. A higher chance of committing a type-I error

We do so by controlling the probability of Type I error. (Type I error occurs when Ha is accepted by mistake.) Because the test is conservative (it protects the null hypothesis H0) we need to always specify and declare the probability of Type I error, but often we are not required to declare the probability of Type II error (Type II error occurs when Ha is rejected by mistake).

Question 56

The purpose of hypothesis testing is to assess:

A. Consistency between null hypothesis and research hypothesis
B. Consistency between different research hypotheses
C. Reliability of sampling data as a representation of the population
D. Consistency between a research hypothesis and sampling data

Answer 56

D. Consistency between a research hypothesis and sampling data

Question 57

Hypothesis testing is an example of:

A. Cross tabulation
B. Statistical Inference
C. Sampling Error
D. Statistical Calculation

Answer 57

B. Statistical Inference

Statistical inference means that we make a guess about population parameters based on the corresponding sample statistic.

Question 58

The p level is the probability of obtaining your particular sample result if the _____.

A. Population is well known
B. Null hypothesis is true
C. Sample is representative
D. Alternative or research hypothesis is true

Answer 58

B. Null hypothesis is true

The consistency is represented by the question “what is the chance of getting this sample result if the null hypothesis is true?” This chance is called p level (or p value). If this chance is low, the consistency between null hypothesis and sample data is low, thus we call the results inconsistent with the null hypothesis, and the null hypothesis is rejected

Question 59

If the p-level for your sample result (sample statistic) is 0.03 and your significance level (alpha) is 0.05, you will:

A. Reject the alternative/research hypothesis
B. Reject the null hypothesis
C. Reject both hypotheses
D. Recalculate the sample statistic

Answer 59

B. Reject the null hypothesis

If P [-Level] is low, reject H0 (Null Hypothesis)

Question 60

If the p-level for your sample result (sample statistic) is 0.07 and your significance level (alpha) is 0.05, you will:

A. Reject the alternative/research hypothesis
B. Reject the null hypothesis
C. Reject both hypotheses
D. Recalculate the sample statistic

Answer 60

A. Reject the alternative/research hypothesis

Research Hypothesis: the association between the two variables is systematic.

Question 61

If the p-level for your sample result (sample statistic) is 0.03 and your significance level (alpha) is 0.01, you will:

A. Reject the alternative/research hypothesis
B. Reject the null hypothesis
C. Accept both hypotheses
D. Recalculate the sample statistic

Answer 61

A. Reject the alternative/research hypothesis

Research Hypothesis: The association between the two variables is systematic.

Question 62

If the p-level for your sample result (sample statistic) is 0.001 and your significance level (alpha) is 0.01, you will:

A. Reject the alternative/research hypothesis
B. Reject the null hypothesis
C. Accept both hypotheses
D. Recalculate the sample statistic

Answer 62

B. Reject the null hypothesis

Null Hypothesis: There is no systematic association between the tabulated variables.

Question 63

A correlation coefficient r = -0.3 for variables X and Y might suggest a _____ relationship between the two variables.

A. Small
B. Positive
C. Negative
D. Statistically significant

Answer 63

C. Negative

Question 64

The correlation coefficient r for variables X and Y comes up equal to 0.2. You:

A. Conclude that X and Y are directly (positively) correlated
B. Conclude that X and Y are inversely (negatively) correlated
C. Test for the statistical significance of r
D. Analyze the X-Y relationship further using linear regression

Answer 64

C. Test for the statistical significance of r

Question 65

The correlation coefficient r for variables X and Y comes up equal to 0. You conclude that:

A. X and Y are negatively correlated
B. You need to test the statistical significance of r
C. There is no relationship between X and Y
D. X and Y are not linearly related

Answer 65

D. X and Y are not linearly related

Question 66

A single correlation coefficient can be used to evaluate the relationship between ____ variables.

A. Two variables
B. Two or more variables
C. One dependent variable and many independent variables
D. Many dependent variables and one independent variable

Answer 66

A. Two variables

Question 67

A correlation coefficient can be used to evaluate the _____ relationship between variables.

A. Nonlinear
B. Multivariate
C. Linear
D. Circular

Answer 67

C. Linear

Question 68

The question: “Do suburban residents use coupons more or less than urban residents?” implies an analysis with how many variables?

A. One variable (area of residence, with two levels: suburban/urban)
B. Two variables (area of residence, and coupon use)
C. Three variables (area of residence, coupon, and use)
D. Four variables (urban resident, suburban resident, coupon, and use)

Answer 68

B. Two variables (area of residence, and coupon use)

Question 69

The question: “Do younger consumer use social media more often than other consumers?” implies which analysis?

A. An estimate of the proportion of young consumers among social-media users
B. An estimate of the proportion of social-media users among young consumers
C. An analysis of association between two variables: age and frequency of social-media usage
D. An analysis of association between frequency of social-media usage and frequency of usage of other media.

Answer 69

C. An analysis of association between two variables: age and frequency of social-media usage

Question 70

In the general representation of the simple regression model, X indicates:

A. Slope
B. Intercept
C. Dependent variable
D. Independent variable

Answer 70

D. Independent variable

Multiple regression represents the relationship between a dependent variable Y and a set of independent variables (X1,X2, X3,..) with a linear function.

Question 71

The coefficients a and b in a simple regression model represent:

A. Dependent and independent variables
B. Statistical significance
C. Confidence interval
D. Intercept and the slope

Answer 71

D. Intercept and the slope

The most important coefficient is b (slope), representing the change in Y that can be expected from a unit change in X (prediction). A is value of Y when X = 0 (Intercept)

Question 72

What is the meaning of the slope in a regression model?

A. Change in the dependent variable corresponding to a unit change in the independent variable
B. Average value for the dependent variable when the independent variable is equal to zero
C. Level of significance of the overall regression model
D. Variance of the dependent variable

Answer 72

A. Change in the dependent variable corresponding to a unit change in the independent variable

Question 73

Which of the following is the general expression of the multiple regression model?

A. Y = a+ b1X1+ b2X2 + b3X3
B. (Y)(X4)= a+ b2X2 + b3X3
C. b2X2 = b3X3
D. X1= b1Y+ b2X12 + b3X3

Answer 73

A. Y = a+ b1X1+ b2X2 + b3X3

Simple regression represents the relationship between the two variables X and Y with a linear function Y=a + bX + e

Question 74

In regression, if the slope coefficient has a p-level of 0.04 and your chosen significance level is 0.05, you conclude that statistically:

A. Not significantly different from zero
B. Significantly positive
C. Significantly negative
D. Significantly different from zero

Answer 74

D. Significantly different from zero

Question 75

In multiple regression analysis, the F test of significance for the whole model had a p-level (also indicated in SPSS as “sig”) p < 0.03. You also looked at the slope coefficients for the individual independent variables. The coefficient for variable X1 was equal to 0.7, with p-level (or “sig”) p

Answer 75

C. The model is significant but the coefficient for X1 is not significant.

Question 76

In multiple regression, the test of significance for each coefficient must be preceded by:

A. A test of significance for the model as a whole
B. Nothing: the significance of each coefficient is the output of the analysis
C. A calculation of the confidence interval for each coefficient
D. A graph of the cloud of points

Answer 76

A. A test of significance for the model as a whole

Question 77

Hypothesis testing can be used to:

A. Analyze differences among potential segments.
B. Compare a market-test result to a standard for prediction purposes.
C. Quantify average market responses based on a sample response.
D. All of the above.

Answer 77

D. All of the above.