Warne 1st edition Full Test Bank

Test Bank

Notes

The questions in this test bank are grouped together by chapter. Occasionally in multiple choice questions, a question can give students hints about the correct answer of another question. This is called cluing, and it can interfere with gaining an accurate view of the student’s knowledge. Efforts have been made to eliminate cluing within each chapter. However, cluing may be present for questions in different chapters. When selecting questions for a test, I recommend that instructors ensure that there is no cluing for items across chapters.

I wrote test questions with two assumptions: (1) students will not have any formulas memorized, and (2) students will not have access to any statistical tables. Instructors who require their students to memorize formulas or who will provide the statistical tables to students may find it advisable to adapt multiple choice questions into short answer questions. Some instructors may also wish to write hand calculation problems if that is an important learning outcome for their students. Additionally, some questions can be adapted to test students’ knowledge of similar concepts in other chapters.

Every chapter has at least 20 multiple choice questions and 5 short answer questions in the test bank. Short answer questions are always the last five listed for that chapter. Instructors who desire more short answer questions can adapt multiple choice questions into short answer questions. Instructors who desire calculation questions can adapt the Guided Practice examples in the textbook or end-of-chapter questions for this purpose. If an instructor does this, please ensure that students are provided with the relevant tables. Instructors who require students to memorize formulas should communicate this information clearly to students; other instructors should provide the relevant formulas with their test materials.

Correct answers are marked in bold.

Chapter 1

The family of research methods that uses statistics to analyze numerical data is called
1. Qualitative methods
2. Statistical analysis
3. Quantitative methods
4. Descriptive statistics
Which of the following is not a reason why students in the social sciences need to learn statistics?
1. Evaluate the conclusions of researchers
2. To help the social sciences obtain the same respect as the physical sciences
3. Communicate findings to others
4. Interpret research to create practical, real-world results
Why do students and practitioners in the social sciences need to know how to separate good research from bad research?
1. To know which articles should receive media attention and awards.
2. To properly select articles to discuss with professors or teachers.
3. The quality of published research varies greatly.
4. All of the above.
How do research questions and research hypotheses differ?
1. Research questions are not scientific, while research hypotheses are scientific.
2. Research hypotheses are appropriate for qualitative methods, while research questions are appropriate for qualitative methods.
3. Research questions are part of confirmatory research, while research hypotheses are speculations about the world.
4. Research hypotheses are expected beliefs the researcher has; research questions tend to be more exploratory.
To be scientific, a research hypothesis must be
1. Sophisticated
2. Falsifiable
3. Plausible
4. Interesting
In statistics, what is a population?
1. Every person, event, or object that a researcher could wish to study.
2. The particular people who are part of a study.
3. All of the people who live in a country where the researchers conducts their study.
4. None of the above.
Why do researchers rarely have population data?
1. It is unethical to force people to participate in a research study.
2. Constraints of money and time may make it unfeasible to gather data from every population member.
3. Many populations are too large.
4. All of the above.
A characteristic that is the same for all sample or populations members is a
1. Variable
2. Invariant characteristic
3. Constant
4. Consistent variable
A dependent variable is the ________________ in the study.
1. Constant
2. Hypothesized cause
3. Operationalization
4. Outcome variable
An independent variable is the ________________ in the study.
1. Constant
2. Hypothesized cause
3. Operationalization
4. Outcome variable
In a _________________ study, a researcher does not manipulate or control the independent variable.
1. Experimental
2. Correlational
3. Analytical
4. Laboratory
Which type of study has more applicability to real-life situations?
1. Experimental
2. Correlational
3. Analytical
4. Laboratory
In a _________________ study, a researcher manipulates or controls the independent variable.
1. Experimental
2. Correlational
3. Analytical
4. Laboratory
What is a common criticism of experimental studies?
1. The studies are difficult to understand.
2. They are too artificial.
3. The data from experimental research is often not easy to interpret.
4. All of the above.
The purpose of __________________ statistics is to describe the data at hand.
1. Descriptive
2. Inferential
3. Analytical
4. Extrapolated
What is a model?
1. A simplification of a complex reality
2. The summary of a study’s findings
3. An understandable explanation of a study’s design
4. All of the above
The purpose of __________________ statistics is to estimate information about the population on the basis of sample data.
1. Descriptive
2. Inferential
3. Analytical
4. Extrapolated
Why do social scientists often create models?
1. To simulate a scenario and test whether a scientist’s theory is supported by data.
2. Because researchers often wish to create a study that can serve as an example of future studies on the same topic.
3. Reality is sometimes too messy and complicated to comprehend.
4. All of the above.
Which of the following is a type of model?
1. Statistical models
2. Theoretical models
3. Visual models
4. All of the above
Which type of model is described in numerical terms or as an equation?
1. Statistical models
2. Theoretical models
3. Visual models
4. All of the above
A bar graph, line graph, and chart are examples of
1. Statistical models
2. Theoretical models
3. Visual models
4. All of the above
Theoretical models take the form of
1. Verbal descriptions
2. Numerical equations
3. Graphs or charts
4. None of the above
A theory is
1. A scientific law which has been shown to display validity across many studies that have been conducted throughout the world.
2. A logically consistent system of principles that posit causal explanations for a wide range of phenomena in many contexts.
3. A foundational idea that permits the design, execution, and interpretation of research within a scientific field.
4. All of the above.
Tony conducted a study where he collected data on two variables: the number of hours a person works and their job satisfaction. He believes that people with more satisfying jobs will choose to work more hours. This is an example of a
1. Experimental study
2. Analytical study
3. Correlational study
4. Laboratory study
Tony conducted a study where he collected data on two variables: the number of hours a person works and their job satisfaction. He believes that people with more satisfying jobs will choose to work more hours. Which variable is the dependent variable?
1. Job satisfaction
2. Number of hours worked
3. There is not enough information to choose a dependent variable.
4. Both are dependent variables.
If every model is wrong to some degree, explain how researchers judge a model.
1. Models are judged by whether they are useful.
Write a research hypothesis.
1. Answers will vary, but the response should be a testable or falsifiable belief about a study’s outcome. It should not be in the form of a question.
Why do researchers never say that their hypothesis or theory has been proven to be true?
1. Because this would require a scientist to make every possible observation of a phenomenon. Otherwise, there could be a scenario or situation where the theory is disproven.
If a population consists of college students in the social sciences, what would a constant among these individuals be?
1. Answers will very, but student responses should be a characteristic of individuals that is the same for all population members (e.g., status as a college student, species that people belong to).
Explain why demographic variables (e.g., race, sex) are considered independent variables.
1. These variables cannot be outcomes—and therefore cannot be dependent variables. It makes much more theoretical sense that these could be theorized causes.

Chapter 2

According to Stevens (1946, p. 677) measurement is:
1. The application of scientific methods to obtain data
2. The use of scientific instruments to measure objects
3. The scientific assessment of objects
4. The assignment of numbers to objects
Data at the highest level of measurement:
1. Has all qualities shared by the lower levels of measurement.
2. Possesses an absolute zero point.
3. Can be doubled without causing problems.
4. All of the above.
The process of defining a variable in a way that allows a researcher to collect numerical data about is called
1. Operationalization
2. Objectivism
3. Reductionism
4. Quantitative defining
Data at Stevens’s highest level of measurement:
1. Has all qualities shared by the lower levels of measurement.
2. Possesses an absolute zero point.
3. Can be doubled without causing problems.
4. All of the above.
Which option lists the four levels of data in ascending order (i.e., from lowest to highest)?
1. Nominal, ordinal, ratio, interval
2. Ratio, interval, ordinal, nominal
3. Nominal, ordinal, interval, ratio
4. Ratio, ordinal, interval, nominal
According to Stevens (1946, p. 677) measurement is:
1. The application of scientific methods to obtain data
2. The use of carefully designed scientific instruments to measure objects
3. The scientific assessment of objects using formal methods
4. The assignment of numbers to objects or events according to rules
Nominal data must be
1. Able to preserve the ranking of subjects
2. Mutually exclusive and exhaustive
3. Different for different group members within the same group
4. All of the above
In nominal data, the numbers assigned to the categories
1. Must be assigned in order so that larger groups receive larger numbers
2. Are arbitrary
3. Can never be negative
4. Must reflect the rank order of the subjects
Which of the following is not an acceptable mathematical function for nominal data?
1. Counting
2. Classification
3. Calculating proportions
4. Calculating averages
In addition to the characteristics of nominal data, ordinal data must also have
1. Rank order in the numbers
2. Proportional representation of group members in the population
3. Absolute zero
4. Proportions calculated from scores
Why is it always better to collect data at the highest level possible?
1. Higher levels of data are easier to collect than lower levels of data.
2. Lower levels of data require more preparation of the data before the statistical analysis can begin.
3. Higher levels of data can always be converted down to lower levels.
4. All of the above.
Which mathematical procedures are acceptable for interval-level data, but not ordinal or ratio data?
1. Calculating proportions
2. Dividing to form averages
3. Ranking subjects
4. Dividing to form ratios
What is the property that interval data have that ordinal data do not?
1. Equal spacing between scale points
2. Absolute zero
3. Consistent data application rules
4. Arbitrary numbers assigned to categories
An absolute zero
1. Is not present unless it is possible for a person in the sample to obtain a score of zero.
2. Is required to calculate averages.
3. Indicates the total absence of the quality being measured.
4. All of the above.
The number of movies that a person has seen in the past month is what type of data?
1. Interval
2. Nominal
3. Ratio
4. Ordinal
Why is it that the Celsius temperature scale not a ratio-level scale?
1. Negative numbers are possible (and meaningful) in the Celsius scale
2. Because it lacks an absolute zero point
3. The ratios that are formed with its numbers do not represent true ratios between temperatures of the amount of heat
4. All of the above
The level of data of ___________________________ is ambiguous.
1. Rating scales
2. Mental health variables
3. Reaction time
4. Group-level variables (e.g., a nation’s average education level)
A(n) _________________ variable permits a wide range of scores that form a constant scale with no gaps at any point along the scale.
1. Ratio
2. Interval
3. Dependent
4. Continuous
A(n) _________________ variable has a limited number of possible values and do not form a constant, uninterrupted scale of scores.
1. Operationalized
2. Discrete
3. Ordinal
4. Limited
A sociologist asks her subjects their religious affiliation. What type of variable would this be?
1. Continuous
2. Interval
3. Nominal
4. Ordinal
A researcher collects data on the length of individuals’ commute to their job. What type of variable is this?
1. Ordinal
2. Ratio
3. Discrete
4. Nominal
Individuals learning a second language were labeled as “not proficient” (group 1), “basic proficiency” (group 2) “high proficiency” (group 3), and “fully proficient” (group 4). What type of data is this?
1. Ordinal
2. Nominal
3. Continuous
4. Ratio
The Fahrenheit temperature scale is an example of a(n) ________________ variable.
1. Ordinal
2. Independent
3. Interval
4. Discrete
What is the minimum level of data required to calculate proportions?
1. Ratio
2. Ordinal
3. Nominal
4. Interval
What level of data can be used to rank order scores?
1. Ratio
2. Ordinal
3. Interval
4. All of the above
Give an example of an operationalization.
1. Answers will vary, but the response should be a method of defining a construct in a way that allows numerical data to be collected about it.
What does it mean that categories in the data must be “mutually exclusive and exhaustive”?
1. The categories are non-overlapping (mutually exclusive) and every sample member belongs to a category (exhaustive)
Explain why test scores can be ordinal-, interval-, or ratio-level data, depending on the interpretation.
1. If the score is the number of questions correctly, then the variable is ratio-level data. If the score is interpreted as the amount of a trait that the subject possesses, then the variable is interval-level data. If it is possible that there are not equal spaces or intervals between scores on a test, then the variable would be ordinal-level data.
Explain what reductionism is and why it is a shortcoming of quantitative research.
1. Reductionism is a philosophy that redefines variables so that they are a shallower version of the construct of interest. It is a problem of quantitative research because it means researchers don’t really study their constructs, but rather the operationalizations of the constructs.
Why is it acceptable (and sometimes necessary) to create a “miscellaneous” or “other” category for data?
1. Because categories must be exhaustive, which means that every sample member must belong to a category. Creating this “miscellaneous” category ensures that individuals who would otherwise not have a category can belong to one.

Chapter 3

In a frequency table, all of the frequency values will add up to
1. 0
2. 10
3. 100
4. The sample size
The width of a bar in a histogram is called the
1. Response frequency
2. Cumulative frequency
3. Interval
4. Ordinal width
The height of a bar in a histogram corresponds to the
1. Frequency of responses in that score range
2. The cumulative frequency of responses in the score range
3. The ordinal nature of histogram data
4. None of the above
What does a gap in the bars in a histogram represent?
1. The discrete nature of the data in the histogram
2. The presence of a score of zero, which indicates the variable is ratio-level data
3. A platykurtic distribution of scores
4. The absence of scores in the gap’s range
Which of the following is a characteristic of a normal distribution?
1. Unimodal
2. Symmetrical
3. Mean = Median = Mode
4. All of the above
The tails of a normal distribution
1. Extend to from -∞ to +∞
2. Contribute to the distribution’s platykurtic shape
3. Are skewed
4. All of the above
Belgian mathematician Lambert Adolphe Jacques Quetelet was the first scientist to observe that a human characteristic was normally distributed. What variable did Quetelet measure?
1. Intelligence test scores
2. Height
3. Reaction time
4. Chest circumference
What is the best method of determining whether a variable is normally distributed?
1. Examine the data in a frequency table
2. Create a histogram of the data
3. Research how the variable is reported in previous studies
4. All of the above
Skewness measures the degree to which a distribution is
1. Leptokurtic
2. Made of interval-level data
3. Continuous
4. Non-symmetrical
A histogram in which all the bars are the same height is called a
1. Uniform distribution
2. Positively skewed distribution
3. Bimodal distribution
4. Leptokurtic distribution
An unskewed distribution has a skewness value that is
1. Negative
2. Zero
3. Positive
4. Undetermined
Another term for a positively skewed distribution is
1. Skewed left
2. Skewed right
3. Upwardly skewed
4. Platykurtic
The region of a distribution between the peak and the tail is called the
1. Torso
2. Midregion
3. Midsection
4. Shoulder
What is the term for a distribution that has a taller peak, taller tails, and shorter shoulders than a normal distribution?
1. Mesokurtic
2. Platykurtic
3. Leptokurtic
4. Allokurtic
A normal distribution is
1. Mesokurtic
2. Platykurtic
3. Leptokurtic
4. Allokurtic
A platykurtic distribution generally
1. Has a shape similar to a normal distribution
2. Is more peaked and skinnier than a normal distribution
3. Is flatter than a normal distribution
4. Can vary in shape, according to the level of skewness in the distribution
A leptokurtic distribution generally
1. Has a shape similar to a normal distribution
2. Is more peaked and skinnier than a normal distribution
3. Is flatter than a normal distribution
4. Can vary in shape, according to the level of skewness in the distribution
In the histogram, if the peak is on the right side of the distribution, the data are:
1. Positively skewed
2. Negatively skewed
3. Has skewness = 0.
4. It is not possible to tell with this histogram.
A leptokurtic distribution will have a kurtosis value that is
1. Negative
2. Zero
3. Positive
4. Undetermined
Most distributions in the social sciences are
1. Unimodal
2. Bimodal
3. Trimodal
4. Platykurtic
Visually, a histogram is most like a
1. Infographic
2. Pie chart
3. Box plot
4. Frequency polygon
The center line in a box plot represents the
1. Minimum
2. Maximum
3. Median
4. Interquartile range
The rectangle in a box plot represents the
1. Minimum
2. Maximum
3. Median
4. Interquartile range
The advantage of a box plot over a histogram is that it can
1. Show whether a distribution is bimodal
2. Display distributions for multiple variables at once
3. Display the kurtosis of a variable
4. All of the above
A stem-and-leaf plot
1. Displays every score in the dataset
2. Does not show frequently appearing scores in the dataset
3. Is currently a more popular visual model than it was in the past
4. All of the above
Which type of visual model is best for showing trends over time?
1. Frequency polygons
2. Histograms
3. Line graphs
4. Bar graphs
In a scatterplot, each dot represents a
1. Variable
2. Sample member
3. Sample statistic
4. Population parameter
Which kind of data is appropriate for a frequency table?
1. Ordinal
2. Nominal
3. Ratio
4. All of the above
If a distribution has a skewness value of -1.8 and a kurtosis value of -1.4, what would be an appropriate label for the distribution?
1. Negatively skewed and mesokurtic
2. Negatively skewed and platykurtic
3. Unskewed and platykurtic
4. Positively skewed and platykurtic
If a distribution has a skewness value of +0.2 and a kurtosis value of +4.3, what would be an appropriate label for the distribution?
1. Unskewed and leptokurtic
2. Mesokurtic and positively skewed
3. Platykurtic and unskewed
4. Leptokurtic and negatively skewed
If a distribution has a skewness value of -1.4 and a kurtosis value of +0.3, what would be an appropriate label for the distribution?
1. Platykurtic and unskewed
2. Unskewed and leptokurtic
3. Unskewed and mesokurtic
4. Negatively skewed and mesokurtic
If a distribution has a skewness value of +2.3 and a kurtosis value of -2.1, what would be an appropriate label for the distribution?
1. Positively skewed and mesokurtic
2. Negatively skewed and leptokurtic
3. Positively skewed and platykurtic
4. Leptokurtic and negatively skewed
If a distribution has a skewness value of -9.6 and a kurtosis value of +4.1, what would be an appropriate label for the distribution?
1. Negatively skewed and leptokurtic
2. Positively skewed and leptokurtic
3. Negatively skewed and mesokurtic
4. Positively skewed and mesokurtic
What would be an appropriate label for the normal distribution?
1. Mesokurtic, unskewed, and unimodal
2. Leptokurtic, negatively skewed, and mesokurtic
3. Platykurtic, unskewed, and unimodal
4. Bimodal, positively skewed, and mesokurtic
Draw a distribution that is platykurtic, negatively skewed, and unimodal.
1. The distribution must be flat in shape, has a longer tail on the left, and only a single peak.
Draw a distribution that is mesokurtic, positively skewed, and bimodal.
1. The distribution must be as tall as a normal distribution, have a longer tail on the right, and two peaks.
What is the cumulative frequency in a frequency table?
1. Cumulative frequency is the number of sample members in a particular row in the frequency table, plus the number of individuals in all higher rows.
A perfectly mesokurtic distribution will have a kurtosis value of:
1. Zero
Explain the differences between a bar graph and a histogram.
1. A bar graph is for nominal/discrete data and always has gaps between bars, while a histogram is designed for continuous (or interval-data or ratio-) data and gaps show score intervals where there are no scores in the data.

Chapter 4

Models of central tendency are used to
1. Move extreme scores more towards the center of the distribution
2. Describe where the middle of the histogram lies
3. Measure the degree to which the center of the sample reflects the center of the population
4. Display data in an understandable manner
Which measure of central tendency can be calculated for all levels of data?
1. Mean
2. Median
3. Mode
4. Trimmed mean
Which measure of central tendency is most influenced by extreme values?
1. Mean
2. Median
3. Mode
4. Trimmed mean
A drawback of the trimmed mean is that
1. It requires eliminating scores from the dataset.
2. The trimmed mean requires ordinal data to calculate.
3. It does not guarantee the removal of both high and low outliers.
4. All of the above.
The interquartile range is a model of variability that is most like which model of central tendency?
1. Mode
2. Median
3. Mean
4. Trimmed mean
Calculating the 10% trimmed mean requires using how much of data?
1. 10%
2. 80%
3. 90%
4. 110%
In statistics, n is the abbreviation for
1. Numbers
2. Sample size
3. Median
4. Trimmed mean
An outlier is a score that is unusual because it is
1. A score that needs to be removed from a dataset before analysis can occur
2. Much higher or much lower than the other scores in the dataset
3. Not a score that belongs to population members
4. A duplicate score that distorts the data
What is an advantage of the mode?
1. It can be used in later statistical calculations
2. It is not distorted by outliers
3. It corresponds to the center number in the distribution
4. All of the above
Which of the following samples would consist entirely of outliers?
1. Children who watch cartoons
2. People who catch influenza
3. Nobel Prize winners
4. None of the above
The median is defined as the _______________ of a distribution.
1. Average score
2. Standard deviation
3. Most frequent score
4. Middle score
The advantage of the sample mean is that it is
1. Easily understandable
2. The best estimate of the population mean
3. The average score in the sample
4. All of the above
The mean is defined as the score that
1. Exceeds half the scores in the dataset
2. The deviation scores of the sample sum to zero
3. Is the most common score in the dataset
4. None of the above
One disadvantage of the mean is that it
1. Maximizes the deviation scores
2. Cannot be used in later statistical calculations
3. May not be a score in the dataset
4. All of the above
What is the impact that outliers have on the mean?
1. They pull the mean closer to their scores
2. Outliers have no impact on the mean
3. They move the mean closer to the middle of the distribution
4. Outliers move the mean closer to the mode by making the distribution more skewed
A deviation score is defined as
1. The sum of all scores in a dataset, minus the median
2. The amount of sampling error in a score
3. The most common score in the dataset
4. A score minus the mean
What is a disadvantage of the trimmed mean?
1. It does not produce a realistic estimate of central tendency
2. It cannot eliminate the influence of outliers
3. It may not be appropriate to eliminate data from the dataset
4. All of the above
When models of central tendency differ, what is the appropriate course of action?
1. Find the average value of the different models to produce a more accurate estimate of central tendency
2. Use the mean as a model of central tendency because it is the most accurate model of central tendency
3. Re-calculate the central tendency values to determine which one was in error
4. Choose a preferred model and justify the decision
What is a statistical assumption?
1. A requirement that the data must meet in order for a statistical procedure to be used
2. An underlying belief about the nature and purpose of a statistical procedure
3. The assumed population parameter that the sample statistic is supposed to estimate
4. A formula used to calculate a statistic
What is the procedure for calculating the range of data?
1. Divide the mean by the highest score in the dataset
2. Subtract the lowest score from the highest score
3. Subtract 1 from the highest score
4. None of the above
The interquartile range is defined as
1. The range of the middle 50% of scores
2. The third quartile minus the first quartile
3. The 75th percentile minus the 25th percentile
4. All of the above
Another term for the median is the
1. Derived central tendency
2. Second quartile
3. Most central score
4. Distributional midpoint
Which model of variability will be produce the largest value?
1. Range
2. Interquartile range
3. Standard deviation
4. It depends on the dataset
Which formula is appropriate for calculating a standard deviation when a researcher has sample data and does not wish to generalize to the population?
1. Any of the above equations is appropriate
In the equation , the n – 1 is called a(n)
1. Compensatory correction
2. Underestimate correction
3. Bessel’s correction
4. Sampling adjustment
In the equation , why is it necessary to put n – 1 in the denominator?
1. Because having n in the denominator would be appropriate for population data
2. To convert the standard deviation to the variance
3. Using n in the denominator would result in a population standard deviation estimate that is too small
4. All of the above
Which level of measurement is most basic level required to calculate the mean?
1. Nominal
2. Ordinal
3. Interval
4. Ratio
Which level of measurement is most basic level required to calculate the median?
1. Nominal
2. Ordinal
3. Interval
4. Ratio
Which level of measurement is most basic level required to calculate the mode?
1. Nominal
2. Ordinal
3. Interval
4. Ratio
Which level of measurement is most basic level required to calculate the range?
1. Nominal
2. Ordinal
3. Interval
4. Ratio
Which level of measurement is most basic level required to calculate the interquartile range?
1. Nominal
2. Ordinal
3. Interval
4. Ratio
Which level of measurement is most basic level required to calculate the trimmed mean?
1. Nominal
2. Ordinal
3. Interval
4. Ratio
Which level of measurement is most basic level required to calculate the standard deviation?
1. Nominal
2. Ordinal
3. Interval
4. Ratio
Which level of measurement is most basic level required to calculate the variance?
1. Nominal
2. Ordinal
3. Interval
4. Ratio
What is the relationship between models of central tendency and models of variability?
1. Models of variability verify the results from models of central tendency
2. They provide complementary information needed to understand the distribution
3. Models of central tendency provide limits to the possible values of models of variability
4. All of the above
If the variance of a dataset is equal to zero, this indicates that
1. All the scores in the data are equal to one another
2. The standard deviation is also equal to zero
3. There is no variability in the data
4. All of the above
Explain how to calculate a 5% trimmed mean.
1. Remove the highest 5% and lowest 5% of scores and calculate the mean of the remaining of scores (i.e., the middle 90% of scores).
Name an advantage of keeping outliers in a sample.
1. Studying outliers can provide information that can help more typical individuals
2. Also acceptable: samples that consist of outliers may be interesting.
How do you calculate a median where there is an even number of scores?
1. Find the middle two scores, add them together, and divide by two.
2. Also acceptable: find the number halfway between the middle two scores.
What is the relationship between the standard deviation and the variance?
1. The standard deviation is the square root of the variance.
2. Also acceptable: the variance is equal to the standard deviation squared.
Why is it important to calculate models of both central tendency and variability?
1. Both types of models provide different, but complementary, information.

Chapter 5

What is the purpose of a linear transformation?
1. To remove outliers from a sample
2. To display the data and compare it to a normal distribution
3. To convert a set of scores to a new scale
4. All of the above
Adding a constant to every score will have what impact on the mean?
1. It will decrease the mean by the same amount
2. It will increase the mean by the same amount
3. It will not change the mean
4. The impact of this linear transformation on the mean varies
Subtracting a constant from every score will have what impact on the mean?
1. It will decrease the mean by the same amount
2. It will increase the mean by the same amount
3. It will not change the mean
4. The impact of this linear transformation on the mean varies
Adding a constant to every score will have what impact on the median?
1. It will decrease the median by the same amount
2. It will increase the median by the same amount
3. It will not change the median
4. The impact of this linear transformation on the median varies
Subtracting a constant from every score will have what impact on the median?
1. It will decrease the median by the same amount
2. It will increase the median by the same amount
3. It will not change the median
4. The impact of this linear transformation on the median varies
Adding a constant to every score will have what impact on the trimmed mean?
1. It will decrease the trimmed mean by the same amount
2. It will increase the trimmed mean by the same amount
3. It will not change the trimmed mean
4. The impact of this linear transformation on the trimmed mean varies
Subtracting a constant from every score will have what impact on the trimmed mean?
1. It will decrease the trimmed mean by the same amount
2. It will increase the trimmed mean by the same amount
3. It will not change the trimmed mean
4. The impact of this linear transformation on the trimmed mean varies
Adding a constant to every score will have what impact on the range?
1. It will decrease the range by the same amount
2. It will increase the range by the same amount
3. It will not change the range
4. The impact of this linear transformation on the range varies
Subtracting a constant from every score will have what impact on the range?
1. It will decrease the range by the same amount
2. It will increase the range by the same amount
3. It will not change the range
4. The impact of this linear transformation on the range varies
Adding a constant to every score will have what impact on the interquartile range?
1. It will decrease the interquartile range by the same amount
2. It will increase the interquartile range by the same amount
3. It will not change the interquartile range
4. The impact of this linear transformation on the interquartile range varies
Subtracting a constant from every score will have what impact on the interquartile range?
1. It will decrease the interquartile range by the same amount
2. It will increase the interquartile range by the same amount
3. It will not change the interquartile range
4. The impact of this linear transformation on the interquartile range varies
Adding a constant to every score will have what impact on the standard deviation?
1. It will decrease the standard deviation by the same amount
2. It will increase the standard deviation by the same amount
3. It will not change the standard deviation range
4. The impact of this linear transformation on the standard deviation varies
Subtracting a constant from every score will have what impact on the interquartile range?
1. It will decrease the standard deviation by the same amount
2. It will increase the standard deviation by the same amount
3. It will not change the standard deviation
4. The impact of this linear transformation on the standard deviation varies
Adding a constant to every score will have what impact on the variance?
1. It will decrease the variance by the same amount
2. It will increase the variance by the same amount
3. It will not change the variance
4. The impact of this linear transformation on the variance varies
Subtracting a constant from every score will have what impact on the variance?
1. It will decrease the variance by the same amount
2. It will increase the variance range by the same amount
3. It will not change the variance
4. The impact of this linear transformation on the variance varies
Multiplying every score by a constant will change the ______________________ by multiplying the statistic by the same constant?
1. Median
2. Mean
3. Trimmed mean
4. All of the above
Dividing every score by a constant will change the ______________________ by dividing the statistic by the same constant?
1. Median
2. Mean
3. Trimmed mean
4. All of the above
Multiplying every score by a constant will change the ______________________ by multiplying the statistic by the same constant?
1. Interquartile range
2. Range
3. Standard deviation
4. All of the above
Dividing every score by a constant will change the ______________________ by dividing the statistic by the same constant?
1. Interquartile range
2. Range
3. Standard deviation
4. All of the above
If a sample member has an original score less than the mean, you can predict that the individual’s z-score will be
1. Negative
2. Zero
3. Positive
4. It is impossible to tell
A linear transformation changes the _______________ of a histogram
1. Shape
2. Skewness
3. Axis
4. Kurtosis
A set of z-scores has a mean of _________ and a standard deviation of _________.
1. 0, 1
2. 100, 10
3. 10, 1
4. 1, 0.1
Which z-score is closest to the mean?
1. -3.2
2. 1.5
3. 1.9
4. -1.8
Which z-score is furthest from the mean?
1. -3.2
2. 1.5
3. 1.9
4. -1.8
Which of the following is a benefit of z-scores?
1. A set of z-scores will always be normally distributed
2. They permit comparisons of scores across different scales
3. Calculating z-scores will convert ordinal data to interval data
4. All of the above
Another term for z-scores is
1. Normalized scores
2. Standardized scores
3. Distributional scores
4. Zero-centered scores
In a normal distribution, half of z-scores are below what value?
1. -1
2. 0
3. .50
4. 10
In a normal distribution, 2.28% of scores will be above what value?
1. -15
2. -2
3. 1
4. 2
If the mean of a set of scores is 13 and the standard deviation is 4, what is the original score for someone who had a z-score of +1.5?
1. 7
2. 9
3. 17
4. 19
If the mean of a set of scores is 9 and the standard deviation is 6, what is the original score for someone who had a z-score of -2.5?
1. -6
2. 6
3. 15
4. 24
When data are converted to z-scores, what happens to the shape of the distribution?
1. There are no changes in the shape of the histogram.
In a normal distribution, almost all (99.73%) of z-scores will be between which two values?
1. -3 and 3
Interpret a z-score of +1.6.
1. The person’s original score is 1.6 standard deviations above the mean of the dataset.
If every score in a dataset if multiplied by 2, explain how you can estimate the variance of the scores after this linear transformation.
1. The variance can be calculated by multiplying the standard deviation by 2 and squaring the result.
Explain why the z-score formula is a type of linear transformation.
1. The z-score formula subtracts the same number (in this case, the mean) from every score and then divides the result by the same number (which is the standard deviation). Both of these numbers are constants.

Chapter 6

Probability is the branch of mathematics that is concerned with the
1. Result of a specific event that that occurs in nature
2. Likelihood of outcomes for an event
3. The combinations of independent and dependent variable values that occur in a study
4. All of the above
_______________ probability is calculated by observing trials and their outcomes.
1. Analytical
2. Theoretical
3. Empirical
4. Calculated
_______________ probability is estimated on the basis of theory.
1. Analytical
2. Theoretical
3. Empirical
4. Calculated
A probability of 1.00 indicates that
1. The outcome is certain to happen
2. The outcome is equally likely as any other outcome
3. The outcome’s likelihood is uncertain
4. The probability’s interpretation is uncertain
_______________ is calculated in the same way as relative frequency.
1. Theoretical probability
2. Joint probability
3. Empirical probability
4. None of the above
Independent trials are
1. Scores collected for an independent variable
2. Separate results for the different probability calculations
3. A causal impact on probability results
4. Trials that do not impact the outcome of any other trial
When calculating the likelihood that either of two possible outcomes could occur in a single trial, one should
1. Add together the probabilities for each outcome
2. Subtract the lesser probability from the greater probability
3. Multiply the probabilities for each outcome together
4. Use the probability of the more likely outcome
When calculating the likelihood that two outcomes will occur simultaneously, one should
1. Add together the probabilities for each outcome
2. Subtract the lesser probability from the greater probability
3. Multiply the probabilities for each outcome together
4. Use the probability of the more likely outcome
A probability distribution is
1. Normal distribution that has been converted to a joint probability
2. Histogram of probabilities for different outcomes
3. A uniform distribution of probabilities
4. The most likely histogram for a dataset.
Which of the following is an example of the gambler’s fallacy?
1. “My lucky rabbit’s foot will help me win this game of blackjack.”
2. “I got a good parking spot at work every day this week, and that streak will continue.”
3. “It has been a long time since I’ve caught a fish, so I’m due for one.”
4. “I have just as many good days as bad days, so a good day is possible tomorrow.”
Which of the following is an example of the hot hand fallacy?
1. “My lucky rabbit’s foot will help me win this game of blackjack.”
2. “I got a good parking spot at work every day this week, and that streak will continue.”
3. “It has been a long time since I’ve caught a fish, so I’m due for one.”
4. “I have just as many good days as bad days, so a good day is possible tomorrow.”
In a probability distribution, what does the height of a bar represent?
1. The type of probability that the researcher calculated
2. The most likely outcome for an individual
3. The probability of an outcome
4. The shape of the histogram that a variable produced
What happens to an empirical probability distribution as the number of trials increases?
1. The empirical probability distribution better resembles the theoretical probability distribution
2. The empirical probability becomes more spread out as the number of trials increases
3. The empirical probability distribution becomes more uniform
4. The empirical probability distribution becomes less variable as the number of trials increases
When an outcome is determined by the sum of independent trials, what shape is the theoretical probability distribution?
1. Uniform
2. Negatively skewed
3. Positively skewed
4. Normal
What is the process of drawing conclusions about populations based on sample data?
1. Generalization
2. Descriptive statistics
3. Extrapolation
4. Evidentiary procedure
A histogram of sample statistics (e.g., means) is called a
1. Statistic distribution
2. Sampling distribution
3. Standard error
4. Sampling frequency histogram
A distribution of an infinite number of sample means (n of each sample > 25) will be normally distributed if the original population distribution is
1. Normal
2. Uniform
3. Leptokurtic
4. Any shape
With a large number of sample means, the best estimate of the population mean is
1. The variance sample mean values
2. The standard deviation of sample mean values
3. The mean of sample mean values
4. The largest sample mean value
The standard deviation of a series of sample means is the
1. Population standard deviation
2. Standard error
3. Sampling error
4. Variance parameter
When calculating means from a large number of samples, __________________ will produce a smaller sampling error.
1. More consistent means
2. Less consistent means
3. Larger means
4. Smaller means
To reduce sampling error, a researcher should
1. Sample from a normally distributed population
2. Reduce the standard deviation of the sample means
3. Select samples that have fewer outliers
4. Have a larger sample size
Populations with high variability will produce _____________ sampling error than populations with low variability.
1. More
2. Equal
3. Less
4. More inconsistent
The ____________________ will always have a larger standard deviation than the distribution of means.
1. Individual sample histograms
2. Theoretical sampling distribution
3. Population distribution
4. Empirical sampling distribution
Explain the Central Limit Theorem.
1. The Central Limit Theorem states that a sampling distribution of means will be normally distributed if each sample will have an n of 25 or more.
Explain why the population standard deviation will always be smaller than the standard deviation of sample means.
1. Means are less variable than raw data because any outliers in a particular sample will usually have their influence lessened by the other scores in the sample. As a result, a collection of means will have less variability than raw data.
Why does the Central Limit Theorem permit researchers to use a theoretical probability distribution when generalizing, instead of needing to use an empirical probability distribution?
1. Because a distribution of an infinite number of sample means will be normally distributed (as long as n is 25 or more per sample) it is not necessary to create an empirical distribution because a researcher already knows the ultimate shape of the sampling distribution.
Explain what sampling error is.
1. Sampling error is the normal variation in statistics across samples.
List an example of when a joint probability would be appropriate to calculate.
1. Answers will vary, but student responses should refer to the probability that two specific outcomes will occur in succession or simultaneously.

Chapter 7

The z-test is
1. A test of equality of two sample means
2. A test of whether the sample and population standard deviations are equal
3. The only statistical test that does not require a null hypothesis
4. The simplest null hypothesis statistical significance test
The purpose of a z-test is to determine whether
1. The null hypothesis fits the data
2. The sample is typical of the population
3. The sample mean and population means are equal
4. All of the above
What is the null hypothesis of a z-test?
1. s_x = σ_x
2. = 0
If the null hypothesis is rejected, this indicates that the null hypothesis
1. Does not fit the data well
2. Was incorrect
3. Requires an additional sample for verification
4. All of the above
The α value indicates the
1. Location of the rejection region
2. Decision to reject the null hypothesis is correct
3. Size of the rejection region
4. Appropriate same size for a null hypothesis statistical significance test
The shape of a sampling distribution in a z-test is
1. Uniform
2. Normal
3. Platykurtic
4. Different, depending on the sample size
The groups in a z-test are the
1. Sample and population
2. Means and standard deviations
3. Higher scoring individuals and lower scoring individuals
4. Outliers and the other sample members
Why do most researchers test the null hypothesis if they do not really believe it?
1. The null hypothesis is the most logical hypothesis to test
2. Disproving the null hypothesis can provide support for a researcher’s theory
3. A researcher cannot prove their hypothesis to be true, but they can disprove a null hypothesis
4. All of the above
What are the limits to the values of α?
1. -1 to 1
2. 0 to 1
3. 1 to 10
4. 1 to 100
A one-tailed z-test is appropriate when
1. The data indicate that it would be easier to reject the null hypothesis with a one-tailed test than with a two-tailed test
2. The rejection region is smaller than usual
3. The researcher has no prior expectations of the results of the study
4. Prior research leads an expectation of the results of the study
A two-tailed z-test is appropriate when
1. The data indicate that it would be easier to reject the null hypothesis with a one-tailed test than with a two-tailed test
2. The rejection region is smaller than usual
3. The researcher has no prior expectations of the results of the study
4. Prior research leads an expectation of the results of the study
What is the alternative hypothesis for a two-tailed z-test?
The purpose of an alternative hypothesis is to
1. Provide a hypothesis for the researcher to believe if the null hypothesis is rejected
2. Give the null hypothesis a different form of expression
3. Find results that can support a research hypothesis
4. All of the above
Which alternative hypothesis is appropriate for a one-tailed z-test where the rejection region is on the left side of the sampling distribution?
Which alternative hypothesis is appropriate for a one-tailed z-test where the rejection region is on the right side of the sampling distribution?
The critical value is defined as the location on the sampling distribution where
1. The sample mean’s location in relationship to the population mean
2. The rejection region begins
3. The distribution becomes normal
4. None of the above
The z-observed value is obtained by
1. Using a formula to calculate it
2. Dividing the α in half
3. Calculating the effect size
4. Looking it up in a table
In the seventh step of a z-test, it is appropriate to reject the null hypothesis if
1. If the z-critical value is closer to the middle of the distribution than the z-critical value
2. The α value is less than the p-value
3. The z-observed value is beyond the z-critical value
4. The rejection region is larger than the sampling distribution
The z-observed formula most closely resembles which other formula?
1. The standard deviation formula
2. The standard error formula
3. The mean formula
4. The z-score formula
The effect size for a z-test is
1. A measure of how poorly the null hypothesis fits the data
2. Related to the z-score formula
3. A measure of the difference between sample and population means
4. All of the above
If the null hypothesis is a perfect model for the data, the effect size for a z-test will be
1. 100%
2. 0
3. -1
4. 10
In theory, the limits of Cohen’s d are
1. 0 and 1
2. -∞ to +∞
3. -3 and +3
4. 0 and 100
A Type I error occurs when
1. A researcher rejects the null hypothesis when the null hypothesis is actually true
2. A researcher makes a calculation error when conducting a null hypothesis statistical significance test
3. The null hypothesis is incorrectly defined by the researcher
4. A researcher retains the null hypothesis when the null hypothesis is actually false
A Type II error occurs when
1. A researcher rejects the null hypothesis when the null hypothesis is actually true
2. A researcher makes a calculation error when conducting a null hypothesis statistical significance test
3. The null hypothesis is incorrectly defined by the researcher
4. A researcher retains the null hypothesis when the null hypothesis is actually false
Which of the following is true of Type I and Type II errors?
1. It is impossible to commit both errors at the same time
2. It is never possible to know whether one has committed an error
3. The p-value is the probability of committing a Type I error if the null hypothesis were perfectly true
4. All of the above
On a sampling distribution, a p-value can be represented as
1. The area between the critical value and the middle of the distribution
2. The area between the observed value and the middle of the distribution
3. The area beyond the observed value
4. The area beyond the critical value
A p-value can never equal zero because
1. The p-value is calculated with only negative numbers
2. It would indicate that the sample could have never come from the population
3. A zero would indicate that the null hypothesis is a perfect model for the data—which never happens in real life
4. The limits of p-values are from 1 to 100.
The value of ______________ is always arbitrary.
1. Cohen’s d
2. The z-observed value
3. α
4. The null hypothesis
As sample sizes get larger, it becomes ________________ to reject the null hypothesis
1. Easier
2. More difficult
3. Less sensible
4. More intuitive
In statistics, a special case is
1. A specific application of the General Linear Model to data
2. The example that a statistical procedure’s creator used to demonstrate it
3. A situation where the assumptions of a statistical method break down and do not apply
4. A simplification of a more complex statistical method
What is the advantage to learning the General Linear Model as a statistics student?
1. The GLM provides a framework for understanding statistics
2. Each GLM procedure is just a modified version of previous procedures
3. It makes new statistical procedures easier to understand
4. All of the above
What are the three characteristics of General Linear Model (GLM) procedures?
1. (1) All GLM procedures apply weights to data during the analysis process. (2) All GLM procedures examine the relationships between independent and dependent variables. (3) All GLM procedures produce an effect size.
Why is a rejection region in the tails of a z-distribution?
1. Because this is where the least likely sample means would be if a null hypothesis were true. Therefore, we put the rejection region in the tails because sample means this range fit the null hypothesis very poorly and indicate that it should be rejected.
Provide an example where a one-tailed z-test would be appropriate and explain why the test is appropriate.
1. Responses will vary, but the student’s example indicates that previous theory or research leads them to expect a result of the study.
A z-table sometimes does not display a value that you need. What should you do if the table does not contain the exact value indicated by your data?
1. Find the value highest value in your table that does not exceed the exact desired value in the dataset.
2. Also acceptable: Use the Price Is Right rule.
What information does Cohen’s d provide that a z-test does not?
1. Cohen’s d quantifies the size of the difference between means, whereas a z-test merely describes whether the null hypothesis fits the data or not.

Chapter 8

When conducting a one-sample t-test, why is it necessary to make some adjustments to the z-test procedures?
1. Because the one-sample t-test is not part of the General Linear Model
2. The population distribution shape is not known
3. Because σ is not known
4. All of the above
The groups in a one-sample t-test are the
1. Sample and population
2. Means and standard deviations
3. Higher scoring individuals and lower scoring individuals
4. Outliers and the other sample members
What is the null hypothesis of a one-sample t-test?
1. s_x = σ_x
2. = 0
The default α value in a one-sample t-test is
1. 0
2. .01
3. .05
4. .50
What is the alternative hypothesis for a two-tailed one-sample t-test?
Which alternative hypothesis is appropriate for a one-tailed one-sample t-test where the rejection region is on the left side of the sampling distribution?
Which alternative hypothesis is appropriate for a one-tailed one-sample t-test where the rejection region is on the right side of the sampling distribution?
HARKing stands for
1. Hypothesizing after results are known
2. Histogram with high and random kurtosis
3. Hypotheses that alter recent knowledge
4. Histogram and ratio data kurtosis
Which of the following is an example of HARKing?
1. A researcher using the statistical results to alter their beliefs about the phenomenon
2. Searching for an explanation of a study’s results
3. Changing a two-tailed test into a one-tailed test in order for to make it easier to reject the null hypothesis
4. Completing a study and then using those results to plan a new study
What are degrees of freedom?
1. The differences between the sample standard deviation value and the population standard deviation value
2. The number of data points that can vary, given the statistical results for the dataset
3. The difference between the t-observed and the t-critical values
4. The difference between the results of a z-test and the results of a one-sample test for the same data
In the seventh step of a one-sample t-test, it is appropriate to retain the null hypothesis if
1. If the z-critical value is closer to the middle of the distribution than the z-critical value
2. The α value is greater than the p-value
3. The t-observed value is beyond the t-critical value
4. The rejection region is larger than the sampling distribution
A Cohen’s d value of ______________ is approximately the smallest effect size that would be noticeable to an attentive observer in daily life.
1. 0.00
2. 0.20
3. 0.50
4. 1.00
In a one-sample t-test, Cohen’s d quantifies the
1. Number of degrees of freedom in the sample
2. Importance of a finding
3. Difference between two group means
4. Statistical significance of the one-sample t-test
Comparing effect sizes allows researchers to
1. Ascertain which variables have the largest group differences
2. Determine the most effective treatment for a condition
3. Help understand the relationship between group differences on different variables
4. All of the above
Under what condition does a one-sample t-test simplify to a z-test?
1. When a variable is normally distributed
2. When the degrees of freedom is equal to infinity
3. When the sample is representative of the population
4. All of the above
When the p-value for a one-sample t-test is less than the α value,
1. The t-observed value is closer to zero than the t-critical value is
2. The null hypothesis should be rejected
3. A Type II error has occurred
4. The null hypothesis should be retained
If using t-table in Appendix A2, each of the following is a possible p-value range, except
1. p < .05
2. p > .05
3. .01 > p > .05
4. p < .01
If the results of a one-sample t-test are statistically significant, this indicates that
1. The statistical results are important
2. The null hypothesis was rejected
3. The results can be translated into a treatment to help people
4. There is no difference between the sample and population means
What is practical significance?
1. A result that are almost statistically significant
2. A result that indicates that fewer sample members need treatment than before
3. A group difference that is practically useful in the real world
4. All of the above
What is clinical significance?
1. A result that are almost statistically significant
2. A result that indicates that fewer sample members need treatment than before
3. A group difference that is practically useful in the real world
4. All of the above
Confidence intervals describe
1. A range of plausible values for a population parameter
2. The degree of confidence that a researcher has in their quality of their sample
3. The values that cannot be population parameters
4. The degree to which future samples will replicate the findings
A point estimate of a population parameter is
1. Inside the confidence interval
2. The best estimate that a sample produces for a population parameter
3. Usually not precisely correct
4. All of the above
The minimum value of the confidence interval is called the
1. Bottom
2. CI minimum
3. Lower limit
4. Lower threshold
The maximum value of the confidence interval is called the
1. Top
2. CI maximum
3. Upper limit
4. Higher threshold
A narrow confidence interval indicates
1. That there is too much sampling error in the population
2. That the point estimate is likely trustworthy
3. The one-sample t-test cannot produce interpretable results
4. The sample size is small
A 95% confidence interval indicates that
1. There is a 95% chance that a study’s results would replicate in later samples
2. We would be 95% sure that the population parameter is inside the confidence interval
3. 95% of later sample means would fall within the confidence interval
4. 95% of confidence intervals from repeated samples would include the population parameter
A confidence interval indicates the level of _________________ in the sample.
1. Sampling error
2. Effect size uncertainty
3. Mean error
4. Capture percentage
Confidence intervals can also be used to
1. Verify the results of the Central Limit Theorem
2. Conduct a null hypothesis statistical significance test
3. Determine the shape of the population distribution
4. Avoid the possibility of a Type I error
In addition to the population mean, the one-sample t-test can examine whether there is a difference between the sample mean and
1. The sample standard deviation
2. The population standard deviation
3. Any number the researcher chooses
4. All of the above
If the null hypothesis in a one-sample t-test is rejected when α = .01, what will happen when α is changed to .05?
1. The null hypothesis will be retained
2. The null hypothesis will be rejected
3. The one-sample t-test will be converted into a z-test
4. It is impossible to tell
Describe how to use a t-table to find a range of p-values for a one-sample t-test.
1. A p-value range can be found by examining which alpha values the null hypothesis is rejected or retained for. If the null hypothesis has been rejected, then p is lower than alpha. If the null hypothesis has been retained, then p is higher than alpha.
What does the term “statistical significance” indicate?
1. “Statistical significance” indicates that the null hypothesis has been rejected.
Give an example of how the concepts of practical significance or clinical significance can sometimes be more useful than statistical significance.
1. Responses will vary, but student examples should be situations where the context of importance or the clinical usefulness (e.g., whether people actually are well enough to not need treatment) is important.
Explain why a point estimate is always inside the confidence interval.
1. Because the confidence interval is the range of plausible values for a population parameter, while the point estimate is the best estimate of the parameter. The best estimate will always be plausible—and therefore inside the confidence interval.
Explain why a researcher might use a one-sample t-test to compare a sample mean to a number that is not the population mean.
1. Sometimes a researcher might want to compare a sample mean to another number because that other number might have an important interpretation (e.g., for diagnosis) and it would be important to determine whether the sample mean is equal to—or not equal to—that number.

Chapter 9

Which of the following is an example of paired data?
1. Scores from child participants and scores from adult participants
2. Two scores collected from the same individual
3. Scores from native English speakers and scores from native Spanish speakers
4. All of the above
Which of the following is an example of paired data?
1. A set of scores from older adults and scores from younger adults
2. Scores from a population of religious people and scores from a population of non-religious people
3. Scores from a sample of parents and scores from each’s parent’s child
4. All of the above
What is the example of an artificial pairing of scores?
1. Scores connected to people’s therapeutic treatments
2. A researcher pairing up two similar people in different groups
3. Two scores collected by a computer from the same person
4. All of the above
What is the null hypothesis for a paired-samples t-test?
1. = 0
2. > 0
3. < 0
4. ≠ 0
What is the alternative hypothesis for a two-tailed paired-samples t-test?
1. = 0
2. > 0
3. < 0
4. ≠ 0
Which alternative hypothesis would be appropriate for a one-tailed paired-samples t-test when all of the rejection region is on the right side of the distribution?
1. = 0
2. > 0
3. < 0
4. ≠ 0
Which alternative hypothesis would be appropriate for a one-tailed paired-samples t-test when all of the rejection region is on the left side of the distribution?
1. = 0
2. > 0
3. < 0
4. ≠ 0
In a paired-samples t-test the sampling distribution is made up of
1. Differences between sample means
2. Sample means
3. Population means
4. Sample standard deviations
As the sample size increases in a paired-samples t-test increases, the sampling distribution more closely resembles a
1. Uniform distribution
2. Normal distribution
3. Leptokurtic distribution
4. Skewed distribution
Glass’s Δ is appropriate when
1. A treatment could change the standard deviation of scores
2. The null hypothesis in a paired-samples t-test has been rejected
3. The null hypothesis in a paired-samples t-test has been retained
4. The effect size is greater than 1.0
The pooled standard deviation is
1. The standard deviation of the effect size
2. Another name for the variance of a sample
3. The overall standard deviation of multiple groups combined
4. The population standard deviation estimate
Because there are three effect size options for a paired-samples t-test, a researcher must
1. Justify the choice of effect size
2. Clearly explain which effect size option you choose
3. Understand the differences between the effect size options
4. All of the above
Which is more serious, a Type I or a Type II error?
1. Type I error
2. Type II error
3. Both errors are equally serious
4. It depends on the situation
Why do the one-sample t-test and the paired samples t-test share so many similarities?
1. Both are members of the General Linear Model
2. The sampling distributions for both procedures are the same shape
3. A paired samples t-test uses differences scores to conduct a one-sample t-test
4. All of the above
Which two characteristics make it easier to reject the null hypothesis?
1. A higher α and larger sample size
2. A higher α and smaller sample size
3. A lower α and larger sample size
4. A lower α and a smaller sample size
The n used to calculate the t-observed value in a paired-samples t-test refers to the
1. Total number of scores
2. Number of pairs of scores
3. Number of degrees of freedom
4. Total number of scores minus 1
If a paired-samples t-test result is statistically insignificant, this means that
1. The p-value is greater than the α value
2. The null hypothesis was retained
3. The t-observed value is closer to zero than the t-critical value is
4. All of the above
Why is it is never possible to know with complete certainty whether a Type I or Type II error has occurred?
1. The sample sizes are usually too small to conclusively produce a result about the null hypothesis
2. Because we never fully know whether the null hypothesis is true or not
3. The paired-samples t-test is not a suitable method for investigating Type I or Type II error
4. None of the above
The formula for Glass’s Δ is most like the formula for
1. Cohen’s d
2. The sample mean
3. The sample standard deviation
4. The t-observed value
In a paired-samples t-test, the number of scores in each group is
1. Larger in the first group
2. 25 or more in both groups
3. Equal in the two groups
4. Larger in the second group
How are the two sets of scores in a paired samples t-test prepared for analysis?
1. By subtracting one score in a pair from the other to produce a difference score
Why does Glass’s Δ use the control group’s standard deviation to compare the means for the two groups in a t-test?
1. Because often interventions can change the standard deviation of the data and make individuals more variable. This would reduce the effect size, so it may make more sense to compare groups using a standard deviation of the control group because they will usually better resemble the population. This is because the population will usually not have been exposed to the treatment.
What is the impact of decreasing alpha from .05 to .01 on the probability of committing a Type II error?
1. Decreasing alpha will always increase the probability of a Type II error because it makes it harder to reject a null hypothesis.
Provide an example of a situation where a researcher can tolerate high Type II error but wants to minimize a Type I error as much as possible.
1. Responses will vary, but student examples should be situations where a Type I error/false positive is a very costly/unpleasant/undesirable outcome, and a Type II error/false negative is comparatively less serious or damaging.
In a z-test and a one-sample t-test, the sampling distribution consisted of sample means. However, this is not true for a paired samples t-test. What does the sampling distribution of a paired samples t-test consist of?
1. The sampling distribution of a paired samples t-test consists of differences between means.

Chapter 10

What is the null hypothesis in an unpaired two-sample t-test?
What level of data is the independent variable in an unpaired two-sample t-test?
1. Interval
2. Ordinal
3. Nominal
4. Ratio
Which of the following is the alternative hypothesis for a two-tailed test in an unpaired two-sample t-test?
Which of the following is the alternative hypothesis for a one-tailed test in an unpaired two-sample t-test where the rejection region is on the left?
Which of the following is the alternative hypothesis for a one-tailed test in an unpaired two-sample t-test where the rejection region is on the right?
A _____________________ unpaired two-samples t-test is appropriate for situations where there is little prior research or theory guiding expectations of results.
1. Unpaired two-sample t-test
2. One-tailed test
3. Two-tailed test
4. None of the above
What is the pooled variance?
1. The square of the pooled standard deviation
2. The variance of scores from two samples that are combined together
3. Part of the t-observed formula
4. All of the above
What is statistical power?
1. The ability for a statistic to produce a large effect size
2. The power of a dataset to produce generalizable results
3. The probability that a false null hypothesis will be rejected in a study
4. The frequency that the critical value is on the right side of the sampling distribution
High statistical power reduces the likelihood of a
1. Type I error
2. Type II error
3. Rejecting the null hypothesis
4. Retaining the null hypothesis
Which of the following will raise statistical power?
1. Increase the sample size
2. Decrease the α value
3. Using an unpaired two-sample t-test instead of a paired samples t-test size
4. None of the above
Studies with larger effect sizes tend to have higher statistical power. But why doesn’t this impact study design?
1. Designs with larger effect sizes are harder to carry out than designs with smaller effect sizes
2. Larger effect sizes usually require smaller α values
3. The effect size is usually not under the researcher’s control
4. All of the above
When a study has low statistical power and the null hypothesis is rejected
1. A Type II error has occurred
2. The likelihood that the null hypothesis really is false is high
3. The results are untrustworthy and seem unstable
4. All of the above
What is the relationship between group size and statistical power?
1. Studies with equally sized groups usually have more power
2. Studies with smaller groups tend to have more power
3. A combined group size of 100 or more is necessary for a study to have high statistical power
4. All of the above
The correct interpretation of a p-value is the
1. Probability that the researcher has committed a Type I error
2. Likelihood that the results of a study would replicate in a future sample drawn from the population
3. Likelihood that the results of a study are important and relevant to the researcher’s hypothesis
4. Probability that a random sample would produce similar results, assuming the null hypothesis were perfectly true
A p-value indicates the probability that
1. The null hypothesis is false, given the results of the statistical analyses
2. The strength of the effect size in the study
3. The results of the statistical test are due to chance alone
4. None of the above
The homogeneity of variance assumption means that
1. The sample’s standard deviation is the square root of the sample’s variance
2. The two groups have similar levels of variability
3. The dependent variable scores are interval- or ratio-level data
4. All of the above
The homogeneity of variance assumption is met when
1. The sample’s standard deviation is the square root of the sample’s variance
2. The dependent variable’s scores meet the requirements for interval- or ratio-level data
3. The ratio of the two groups’ variance does not exceed 10:1
4. All of the above
One assumption of unpaired two-sample t-tests is that the sample sizes for the two groups are approximately equal. To meet this assumption, the larger group must not have a sample size more than _________ times larger than the smaller group
1. 1
2. 2
3. 4
4. 10
If the relationship between the two variables in an unpaired two-sample t-test is very strong
1. The two groups’ distributions will have little overlap
2. The two groups’ distributions will overlap almost completely
3. The α value will be high
4. The t-critical value will be outside the tail of the distribution
As the relationship between variables in an unpaired two-sample t-test decreases,
1. The two groups’ distributions overlap more
2. The overlap between the two groups will decrease
3. The α value decreases
4. The t-critical value moves towards the tail of the distribution
Explain why the null hypothesis of in a paired-sample t-test is not appropriate for an unpaired samples t-test.
1. Because calculating D values requires pairing scores across groups, and there is no logical way to pair data from independent groups.
If a researcher has paired data, it is mathematically possible to still conduct an unpaired samples t-test. But why is this inadvisable?
1. Because an unpaired samples t-test has lower statistical power than a paired samples t-test.
Why is it that an extreme t-observed value and a very small p-value do not prove a null hypothesis true?
1. Because it is never fully known whether the null hypothesis is true or not. Even if the null hypothesis is very easily rejected, it is still possible that the results are a Type I error. Rejecting the null hypothesis once in a single study does not conclusively disprove the null hypothesis forever.
A student said, “When I rejected the null hypothesis, I obtained a p-value of .035. This means that there is only a 3.5% probability that the null hypothesis is true.” What is the problem with the student’s logic?
1. A p-value is the probability that a random sample would produce the observed results, given the assumption that the null hypothesis was perfectly true. The student has the logic reversed: p does not state the probability that the null hypothesis is true, given the observed data.
Why is the homogeneity of variance assumption important for an unpaired samples t-test?
1. Because an unpaired samples t-test only tests whether two sample means are equal. But if two distributions have very different variances, the two samples could be very different while still having the same mean.

Chapter 11

Instead of conducting an ANOVA, a researcher chooses to conduct a series of unpaired two-sample t-test. This increases which kind of Type I error?
1. Testwise Type I error
2. Experimentwise Type I error
3. Aggregate Type I error
4. Individual Type I error
As the number of groups in an independent variable increases, the number of unpaired two-sample t-tests
1. Reaches a plateau
2. Increases faster
3. Increases at the same rate
4. Changes in an unpredictable fashion
How does one perform a Bonferroni correction?
1. Use a post hoc test that combines the data from some groups
2. Expand the null hypothesis to accommodate multiple group means
3. Divide the α by the number of tests
4. None of the above
Exploratory research
1. Does not have an effect size
2. Is not driven by pre-existing hypotheses
3. Is not appropriate for the first research study on a topic
4. Requires conducting an ANOVA
The sampling distribution in an ANOVA is called a(n)
1. ANOVA distribution
2. Normal distribution
3. t-distribution
4. F-distribution
The alternative hypothesis in an ANOVA is always
1. That any two group means are not equal
2. Appropriate for a one-tailed test
3. The same for any ANOVA
4. All of the above
____________________ created the ANOVA
1. Karl Pearson
2. Sir Ronald Fisher
3. George W. Snedecor
4. William Gosset
In the ANOVA information table, the values are
1. Predicted dependent variable values
2. Calculated for each group and for the total sample
3. Dependent variables averages
4. All of the above
The for all groups combined is the
1. Predicted dependent variable when sample members’ group membership is unknown
2. Grand mean
3. Helpful for interpreting an ANOVA’s effect size
4. All of the above
In an ANOVA, the observed value is obtained from the
1. Table of critical values
2. Information table
3. ANOVA table
4. Post hoc test table
Every number in the ANOVA table will always be
1. Close to the sample average
2. Whole numbers
3. Zero or positive
4. All of the above
When the null hypothesis of an ANOVA is rejected, the observed value is
1. Less than the critical value
2. Greater than 1.96
3. Greater than the critical value
4. Greater than the α value
The effect size for ANOVA is
1. η²
2. Glass’s Δ
3. Cohen’s d
4. Standardized mean difference
The range of possible values for the effect size for an ANOVA is
1. -∞ to +∞
2. -3 to +3
3. 0 to 1
4. 0 to 10
The interpretation of the effect size for an ANOVA is
1. The improvement of predictions when compared to the grand mean
2. The proportion of dependent variable variance shared with the independent variable
3. The degree to which the squared residuals are reduced when using group membership to predict dependent variable values
4. All of the above
Some people like to interpret effect sizes by labeling them as “small,” “medium,” or “large.” What is the problem with this tendency?
1. The same standards for “small,” “medium,” or “large” effect sizes may not be appropriate for all topics of study
2. These labels do not correspond to the importance of an effect size
3. The creator of these labels said that they were not appropriate for all research situations
4. All of the above
Which of the following is not an important component of effect size interpretation?
1. Prior research
2. Group standard deviations
3. Real-life applicability
4. Context
In an ANOVA, the independent variable is _________________ data, and the dependent variable is _________________ data.
1. Nominal-level, nominal-level
2. Interval- or ratio level, interval- or ratio-level
3. Nominal-level, interval- or ratio-level
4. Ratio-level, interval- or ratio-level
It is unnecessary to conduct a post hoc test when the
1. Null hypothesis is rejected
2. Null hypothesis is retained
3. Researcher conducts a one-tailed test
4. Effect size is large
The purpose of a post hoc test is to determine why the ANOVA
1. Sampling distribution is not normal
2. Information table cannot be completed
3. Null hypothesis was rejected
4. Alpha value needs to be adjusted
The harmonic mean is a measure of central tendency that
1. Is equal to the 5% trimmed mean
2. Compensates for the shortcomings of an ANOVA
3. Can be calculated with ordinal-level data
4. Gives disproportionate weight to smaller sample sizes
Instead of examining every possible group difference (as in a Tukey’s post hoc test), some post hoc tests examine only selected group differences based on theory. These post hoc tests are called
1. Cohen’s contrasts
2. Planned contrasts
3. A priori contrasts
4. Explanatory contrasts
Which post hoc test is most appropriate when there is a clear baseline group for other groups to be compared to?
1. Tukey’s post hoc test
2. Bonferroni-Dunn post hoc test
3. Scheffé’s post hoc test
4. Dunnett’s test
Which is the most common post hoc test in the social sciences?
1. Tukey’s post hoc test
2. Bonferroni-Dunn post hoc test
3. Scheffé’s post hoc test
4. Dunnett’s test
_______________ measures accuracy of predictions at the individual level.
1. η²
2. A Bonferroni correction
3. The residual
4. Tukey’s post hoc test
_______________ measures accuracy of predictions at the overall level.
1. η²
2. A Bonferroni correction
3. The residual
4. Tukey’s post hoc test

Which value will always be largest?
1. SOS_Total
2. SOS_Within
3. SOS_Between
4. None of these is consistently larger than the others
Which value will always be largest?
1. df_Total
2. df_Within
3. df_Between
4. None of these is consistently larger than the others
What is another term for the residual?
1. Statistical power
2. Shared dependent variable variance
3. Independent variable condition
4. Error
What are the two disadvantages to conducting a Bonferroni correction?
1. (1) The correction is too severe and can lower alpha values too much, and (2) it does not address the problem of conducting a large number of unpaired samples t-tests.
In ANOVA groups, what is the best value for predicting a sample member’s dependent variable score?
1. The mean for the group that the individual belongs to.
Explain the relationship between an unpaired samples t-test and an ANOVA in the context of the General Linear Model.
1. The unpaired two-sample t-test is a special case of an ANOVA (i.e., the ANOVA is a generalization of the unpaired two-sample t-test). This means that if a person conducts an ANOVA with two groups, the results will be the exact same as the results of an unpaired two-sample t-test.
What is the null hypothesis of an ANOVA, and what does it mean when that hypothesis is rejected?
1. The null hypothesis for an ANOVA is that all group means are equal to one another (i.e., ). When this null hypothesis is rejected, it means that at least one group mean differs from at least one other group mean.
How is the sampling distribution in an ANOVA different from the sampling distribution from a z-test and every type of t-test?
1. The sampling distribution in an ANOVA (i.e., the F-distribution) is always positively skewed, whereas the z-distribution and t-distribution are symmetrical/unskewed.

Chapter 12

When two variables are correlated, it indicates that the two variables are
1. Interval-level data
2. Related
3. Continuous
4. Predictable
The two components of Pearson’s r to interpret are the
1. Level of data and number
2. Sign and level of data
3. Sign and number
4. Direct and inverse correlations
A positive correlation is also called a(n)
1. Direct correlation
2. Nonlinear correlation
3. Linear correlation
4. Inverse correlation
A negative correlation is also called a(n)
1. Direct correlation
2. Nonlinear correlation
3. Linear correlation
4. Inverse correlation
What are the range of possible values for Pearson’s r?
1. -∞ to +∞
2. -3 to +3
3. -1 to +1
4. 0 to +1
Which Pearson’s r value indicates the strongest correlation?
1. +.42
2. -.28
3. -.70
4. +.55
Which Pearson’s r value indicates the weakest correlation?
1. +.42
2. -.28
3. -.70
4. +.55
Which Pearson’s r value would indicate a dataset that has very few exceptions in its relationship between variables?
1. -.52
2. -.22
3. +.05
4. +.83
Which Pearson’s r value would indicate a subtle and extremely weak relationship between its variables?
1. -.52
2. -.22
3. +.05
4. +.83
The Pearson’s r equation automatically converts all the data into
1. z-scores
2. Standard deviations
3. Covariances
4. None of the above
Which correlation will produce a scatterplot that consists of a series of dots on a straight line that extends from the upper left quadrant to the bottom right quadrant of the scatterplot?
1. -1.0
2. 0.0
3. +0.5
4. +1.0
Which correlation will produce a scatterplot that consists of a series of dots that are scattered more-or-less randomly throughout the scatterplot?
1. -1.0
2. 0.0
3. +0.5
4. +1.0
When the correlation is negative, sample members with _____________ scores on the independent variable tend to have _____________ scores on the dependent variable.
1. High, high
2. Low, high
3. Average, average
4. Average, low
When the correlation is positive, sample members with _____________ scores on the independent variable tend to have _____________ scores on the dependent variable.
1. High, high
2. Low, high
3. Average, average
4. Average, low
When a correlation is _________________ the dots in a scatterplot tend to be spread apart, while in a ____________________ correlation the dots in a scatterplot tend to be close together.
1. Strong, weak
2. Positive, negative
3. Weak, strong
4. Negative, positive
The null hypothesis for a correlation coefficient is
r = 0
None of the above
Which of the following is the alternative hypothesis for a two-tailed null hypothesis test of Pearson’s r?
1. r = 0
2. r ≠ 0
3. r < 0
4. r > 0
Which of the following is the alternative hypothesis for a one-tailed null hypothesis test of Pearson’s r where the rejection region is completely on the left side of the sampling distribution?
1. r = 0
2. r ≠ 0
3. r < 0
4. r > 0
Which of the following is the alternative hypothesis for a one-tailed null hypothesis test of Pearson’s r where the rejection region is completely on the right side of the sampling distribution?
1. r = 0
2. r ≠ 0
3. r < 0
4. r > 0
To determine that the null hypothesis for a test of Pearson’s r should be rejected, a researcher should
1. Expand the rejection region until it includes the critical value
2. Follow the same steps as for any other null hypothesis test
3. Ensure that the sampling distribution is non-normal and centered on zero
4. Calculate the Pearson’s r first and then use that information to determine whether to conduct a one- or two-tailed test
The effect size r² is equal to
1. Pearson’s r
2. η²
3. Glass’s Δ
4. Cohen’s d
One interpretation of r² is to compare the improvement of predictions when compared to the
1. Dataset’s z-scores
2. Residuals calculated from the correlation coefficient
3. Sample mean
4. Baseline model without predictors
According to Schönbrodt and Peruigini (2013), a sample size of approximately __________ is needed to produce stable correlations.
1. 25
2. 50
3. 100
4. 250
When two variables are correlated, it is because
1. Changes in the independent variable cause changes in the dependent variable
2. Changes in the dependent variable cause changes in the independent variable
3. Changes in a third variable cause changes in both the independent and dependent variables
4. All of the above are possible reasons why variables are correlated
If two variables are correlated, why is it impossible to state that changing one variable score will result in a change in the other variable’s score?
1. Interventions or treatments may be very complex to design and implement
2. The two variables may be correlated simply because they both relate to a third variable.
3. Dependent variables are often not sensitive to changes from other variables
4. All of the above
Which of the following is an example of a positive correlation?
1. Adolescents who earn higher grades use fewer drugs
2. People who use more social media are better informed of news events
3. Higher income countries have lower illiteracy rates
4. All of the above
Which of the following is an example of a negative correlation?
1. As individuals watch more television, they read less
2. Children with higher reading skills often have higher reading skills
3. Larger cities usually have higher levels of pollution
4. Prisoners who break fewer rules in prison usually commit fewer crimes after being released
Which of the following is an example of a negative correlation?
1. Younger adults tend to have fewer health problems
2. Families with higher incomes generally eat at restaurants more often
3. There is no relationship between people’s height and their political beliefs
4. Countries with higher immunization rates tend to have lower disease rates
Which of the following is an example of a negative correlation?
1. Children who are more involved with sports are more physically fit
2. Adults who vote more often are more involved with community affairs
3. Individuals who are low in neuroticism tend to stay in their jobs longer
4. People with more friends tend to have better coping skills
Which of the following is an example of a positive correlation?
1. Children who play more video games have shorter reaction times
2. Schools where teachers are better trained have fewer disciplinary problems
3. As the temperature increases, more people purchase ice cream
4. Doctors who prescribe more antibiotics have fewer patients who need additional treatment
Which of the following is an example of a positive correlation?
1. Individuals with lower self-control often have fewer friends
2. Smaller universities generally have students who are more involved with activities
3. People who travel less usually feel more comfortable at home than people who travel more
4. Students who send text messages frequently in class generally earn lower grades
Give an example of two uncorrelated variables.
1. Responses will vary, but the student’s response should be too variables that are unrelated (e.g., eye color and height).
Explain the meaning of the phrase “correlation is not causation.”
1. The phrase indicates that the existence of a correlation does not—by itself—prove that changes in one variable cause changes in the other.
Under the baseline model for a correlation, what is every sample member’s predicted score on the dependent variable?
1. The grand mean for the dependent variable (i.e., the mean of all dependent variable scores in the data).
What happens to the pattern of dots on a scatterplot as the correlation between two variables strengthens?
1. The dots group closer together.

Chapter 13

The standardized regression equation is appropriate for situations where
1. The independent and dependent variables are both interval- or ratio-level data
2. The independent and dependent variables are both z-scores
3. The independent variable is nominal-level data and the dependent variable is interval-level data
4. An ANOVA requires a post hoc test
The standardized and unstandardized regression equations are both used to
1. Convert data into z-scores
2. Link regression to the other procedures in the General Linear Model
3. Make predictions about the dependent variable
4. Enlarge the residuals of each sample member
The regression line
1. Is created via the regression equations
2. Is also called the line of best fit
3. Minimizes the vertical distance between each point and the regression line
4. All of the above
All regression predictions of dependent variable scores are
1. On the regression line
2. Larger than the actual dependent variable scores
3. Larger than the independent variable scores
4. All of the above
It takes a minimum of ______________ points to draw a regression line
1. One
2. Two
3. Three
4. Four
Converting all the data to the z-scores
1. Strengthens the correlation
2. Weakens the correlation
3. Does nothing to the pattern of data in the scatterplot
4. Converts the data to ratio-level scores
The regression line will always pass through
1. (10, 10)
2. (0, 1)
3. (, )
4. (-1, +1)
In the standardized regression equation, the y-intercept is always
1. 0
2. 1
3. 5
4. 10
A positive residual indicates
1. The person scored higher than predicted on the dependent variable
2. The correlation is positive
3. The person scored lower than predicted on the dependent variable
4. The correlation is negative
A negative residual indicates
1. The person scored higher than predicted on the dependent variable
2. The correlation is positive
3. The person scored lower than predicted on the dependent variable
4. The correlation is negative
The vertical distance between a person’s point on the scatterplot and the regression line is the
1. Predicted dependent variable value
2. Residual
3. z-score
4. Standard deviation
In the baseline model, every sample member is predicted to have a score equal to
1. Their actual dependent variable score
2. Their z-score
3. The dependent variable standard deviation
4. The dependent variable mean
The regression line and the line for the baseline model intersect at
1. (-1, +1)
2. (, )
3. (-3, +3)
4. (-2, +2)
The r² is
1. The degree of improvement in the predictions when compared to the baseline model
2. The percentage decrease in the sum of the squared residuals when calculated the regression line compared to the sum of the squared residuals calculated from the baseline model
3. The square of the correlation coefficient
4. All of the above
An r² value of 0 indicates that
1. Models based on the regression line are as good as predictions from the baseline model
2. The residuals for the regression line are all zero
3. The baseline model is a better tool for making predictions than the regression line
4. The sum of the residuals using the regression line is zero
An r² value of 1 indicates that
1. The residuals for the data add to 1
2. The baseline model is a line that has a slope of +1
3. All the predictions are perfectly accurate
4. All of the above
Because of regression towards the mean
1. Over time the average score in a population will decline
2. Outliers are difficult to predict
3. The mean of a variable will increase when using a regression equation
4. All of the above
Regression towards the mean will be greatest for
1. Outliers
2. Scores close to the average
3. Correlations of r > .5
4. z-scores
Regression towards the mean will be zero for
1. Outliers
2. Correlations of r > .5
3. Subjects with independent variable scores that are equal to the mean
4. z-scores
Regression towards the mean was discovered by
1. Karl Pearson
2. Francis Galton
3. William Gosset
4. John Tukey
Which would be a real-life situation where regression towards the mean would apply?
1. A sports team that wins a championship is not as good of a team the following year
2. After writing her best novel yet, an author’s following novel is not as good
3. The worst salesperson in a company for a month improves in their performance the following month
4. All of the above
Graphing the regression line requires knowing at least ________ points that fall on the line.
1. 2
2. 3
3. 5
4. 10
Which of the following is not an assumption of regression?
1. A linear relationship between variables
2. Interval- or ratio-level data for both variables
3. Homogeneity of residuals
4. All of the above
What is the usual impact of restriction of range on a correlation coefficient?
1. Its strength is reduced and the value is closer to zero
2. It increase in strength and is closer to -1.0
3. It increases in strength and is closer to +1.0
4. It switches from positive to negative (or from negative to positive)
Which of the following is an example that could cause restriction of range?
1. A sample of English speakers is not representative of the population of people worldwide
2. Requiring a minimum test scores for college admission and then finding the correlation between test scores and college GPA
3. The outliers in a sample have the largest influence over the correlation value
4. All of the above
To determine whether outliers are distorting the relationship, the researcher should ______________________ and determine whether it changes the correlation.
1. Re-calculate the correlation without the outliers
2. Convert the data to z-scores
3. Find the regression line
4. Divide the sample in half
If the data in a correlation coefficient consists of two or more groups,
1. The data should be combined into one group before calculating a correlation coefficient or performing regression
2. The groups should each occupy a different portion of the scatterplot
3. The correlation for each subgroup should be similar to the correlation for the entire dataset
4. All of the above
As a member of the General Linear Model, the point-biserial correlation is a connection between Pearson’s r and the
1. z-test
2. one-sample t-test
3. unpaired samples t-test
4. ANOVA
The slope of the standardized regression line is equal to what value?
1. -1
2. 0
3. 1
4. r
Describe the appearance on the baseline model on a scatterplot.
1. On a scatterplot, the baseline model is a horizontal line that at the mean for the dependent variable.
Explain what homogeneity of residuals means and why it is important.
1. Homogeneity of residuals is an assumption of correlation and regression which states that the dots in a scatterplot form an oval or circle (instead of a trumpet or a funnel). It is important because if the assumption is violated, it indicates that the regression line is a model that fits some of the data better than other parts of the data.
Draw an example of a nonlinear relationship.
1. Student responses will vary, but a correct response will be a scatterplot where the dots cannot be summarized as a straight line. They may instead for a parabola, or possibly will be a pattern of dots that change direction (e.g., by leveling off after an increase or after a decrease).
Explain regression towards the mean.
1. Regression towards the mean is a phenomenon where—on average—predicted dependent variable z-scores will be closer to the mean than the same individuals’ independent variable z-scores were.
2. Also acceptable: Regression towards the mean occurs when researchers predict that individuals will be more average on the dependent variable than they were on the independent variable.
What is the homogeneity of subgroups assumption, and what happens when it is violated?
1. The homogeneity of subgroups assumption is that any subgroups within the sample have similar correlations between the independent and dependent variables. When this assumption is violated, then the correlation for the entire dataset will not be similar to the correlations calculated using each subgroup’s data

Chapter 14

A one-variable chi-squared test most closely resembles which other member of the General Linear Model?
1. z-test
2. one-sample t-test
3. Pearson’s r
4. ANOVA
The one-variable chi-square test has the null hypothesis that the independent variable groups’ data are equal to
1. Theorized proportions of the dependent variable categories
2. Mean scores on the dependent variable
3. Standard deviations on the dependent variable
4. z-scores of the dependent variable
A chi-squared test is always a
1. Test of residual variance
2. Test that requires an α value of .05
3. One-tailed test
4. Two-tailed test
The chi-squared sampling distribution is
1. Normally distributed
2. Negatively skewed
3. Similar in shape to a t-distribution
4. Positively skewed
For a one-variable chi-squared test, the chi-squared observed equation produces a series of fractions that are added together. The number of fractions will always be equal to
1. The number of groups in the dependent variable
2. The number of groups in the independent variable
3. n - 1
4. n / 2
The possible range of chi-squared observed values is
1. -∞ to +∞
2. -3 to +3
3. 0 to 10
4. 0 to +∞
One criticism of null hypothesis statistical significance (NHST) testing procedures is that
1. The p-value is continuous, yet NHST produces a dichotomous decision
2. It can encourage researchers to tweak their study to make it artificially easier to reject the null hypothesis
3. The difference between “significant” and “non-significant” results may be very subtle
4. All of the above
Publication bias
1. Occurs when studies that reject the null hypothesis are more likely to get published
2. Fixes the file drawer problem that plagues the social sciences
3. Is the result of logical decisions about the priorities of scientists
4. All of the above
Publication bias is a result of problematic decisions from
1. Journal editors
2. Researches
3. Peer reviewers
4. All of the above
When an odds ratio is equal to 1.0
1. The two groups’ means are one standard deviation apart
2. The effect size is statistically significant
3. The observed and expected values are equal
4. The groups are all the same size
In a one-variable chi-squared test, if the odds ratio is 2.5, it indicates that
1. There are 2.5 times more independent variable groups than were theorized
2. There are 2.5 times more sample members in a group than were theorized
3. There are 2.5 times fewer independent variable groups than were theorized
4. There are 2.5 times fewer sample members in a group than were theorized
In a one-variable chi-squared test, if the odds ratio is 4.1, it indicates that
1. There are 4.1 times more independent variable groups than were theorized
2. There are 4.1 times more sample members in a group than were theorized
3. There are 4.1 times fewer independent variable groups than were theorized
4. There are 4.1 times fewer sample members in a group than were theorized
In a one-variable chi-squared test, if the odds ratio is 0.5, it indicates that
1. There are 0.5 times as many independent variable groups than were theorized
2. There are 0.5 times as many sample members in a group than were theorized
3. There are 0.5 times as many dependent variable groups than were theorized
4. The overall sample size is 0.5 times as large as was theorized
When an odds ratio is precisely equal to 1.0
1. The null hypothesis is a perfect model for the data
2. The null hypothesis has been rejected
3. The smallest group in the data has the highest mean
4. All of the above
If a one-variable chi-squared test retains the null hypothesis, it indicates that
1. The results are statistically significant
2. The proportion of the sample in each group is similar to the theorized group proportions
3. The p-value is less than the α value
4. The observed value is inside the rejection region of the sampling distribution
The null hypothesis in a two-variable chi-squared test is that
1. Group sizes are all equal to one another
2. The proportions of each group are equal to their theorized proportions
3. The independent and dependent variables are correlated
4. The independent and dependent variable groups are unrelated
For a two-variable chi-squared test, the chi-squared observed equation produces a series of fractions that are added together. The number of fractions will always be equal to
1. The number of groups in the dependent variable
2. The number of groups in the independent variable
3. The number of cells in the table
4. The sum of the number of independent and dependent variable groups added together
The expected values in the rows of a two-variable chi-squared table will always sum up to
1. The sum of the observed values in the same row
2. The sum of the observed values in the columns
3. The number of sample members divided by the number of cells in the row
4. The sum of the number of independent and dependent variable groups added together
The chi-squared sampling distribution is a theoretical distribution of
1. Sample means
2. Sample variances
3. Sample standard deviations
4. Differences between means
Calculating an odds ratio for a two-variable chi-squared test requires choosing a
1. Baseline group
2. Outcome of interest
3. Non-baseline group
4. All of the above
If the baseline and non-baseline groups are reversed, the odds ratio changes to the
1. Original odds ratio multiplied by negative 1
2. Original odds ratio multiplied by the sample size
3. Reciprocal of the original odds ratio
4. Original odds ratio divided by the number of independent variable groups
In a two-variable chi-squared test, if the odds ratio is 1.0, it indicates that
1. The two nominal variables are unrelated
2. The null hypothesis needs to be retained
3. The outcomes are equally likely for the baseline and the non-baseline groups
4. All of the above
In a two-variable chi-squared test, if the odds ratio is less than 1, it indicates that
1. The outcome of interest is more likely for the baseline group
2. The outcome of interest is more likely for the non-baseline group
3. The baseline outcome is less likely for the non-baseline group
4. The outcome of interest is equally likely for both groups
In a two-variable chi-squared test, if the odds ratio is greater than 1, it indicates that
1. The outcome of interest is more likely for the baseline group
2. The outcome of interest is more likely for the non-baseline group
3. The baseline outcome is more likely for the non-baseline group
4. The outcome of interest is equally likely for both groups
An odds ratio can be converted into ____________________ for every combination of groups and outcomes.
1. The null hypothesis
2. A z-score
3. A predicted probability
4. A proportion
Meta-analysis
1. Is possible because effect sizes can be averaged together
2. Overcomes this subjectivity inherent in literature reviews
3. Can resolve contradictions in the scientific literature
4. All of the above
When there are more than two groups in the independent or dependent variable in a two-variable chi-squared test,
1. The researcher may need to calculate more than one odds ratio
2. Multiple groups must serve as the baseline group
3. The equation for an odds ratio becomes much more complex
4. None of the above
The relative risk describes the
1. Odds ratio for z-scores
2. Probability of the outcome of interest occurring for the non-baseline group compared to the baseline group
3. Probability that the null hypothesis will be rejected in a two-variable chi-squared test
4. Results of a meta-analysis for the dependent variable
What does a relative risk of 1.0 indicate?
1. The outcome of interest is more likely for the baseline group
2. Equal proportions of the two groups experience the outcome of interest
3. The outcome of interest is more likely for the non-baseline group
4. The odds ratio would be a more appropriate effect size for the data
What does a relative risk of value that was less than 1.0 indicate?
1. The outcome of interest is more likely for the baseline group
2. Equal proportions of the two groups experience the outcome of interest
3. The outcome of interest is more likely for the non-baseline group
4. The odds ratio would be a more appropriate effect size for the data
What does a relative risk of value that was more than 1.0 indicate?
1. The outcome of interest is more likely for the baseline group
2. Equal proportions of the two groups experience the outcome of interest
3. The outcome of interest is more likely for the non-baseline group
4. The odds ratio would be a more appropriate effect size for the data
A correlation coefficient for nominal data is called
1. Pearson’s r
2. The relative risk
3. The odds ratio
4. Phi (ϕ)
Given an odds ratio of 4.0, what would the odds ratio be if the baseline and non-baseline groups were switched?
1. 0.10
2. 0.25
3. 0.40
4. 1.40
Given an odds ratio of 0.33, what would the odds ratio be if the baseline outcome and the outcome of interest were switched?
1. 1.00
2. 1.33
3. 3.00
4. 33.00
What is a problem with the effect size Cramér’s V?
1. It can only have a negative value
2. It requires interval- or ratio-level data to calculate
3. When the null hypothesis is rejected, the value of V will always be zero
4. Different numbers of categories in the variables can inflate or depress the V value
State a consequence of publication bias.
1. Publication bias distorts the literature and (usually) inflates effect sizes.
2. Also acceptable: it encourages researchers to engage in p-hacking to tweak their studies to produce a p-value below .05.
3. Also acceptable: it prevents readers from having access to the full body of research on a topic.
Explain the similarities between the two-variable chi-squared test and Pearson’s r in the context of the General Linear Model.
1. A two-variable chi-squared test is just a Pearson’s r correlation for nominal data. Also acceptable: the null hypothesis for both procedures is that there is no relationship between independent and dependent variables.
What is the purpose of performing a meta-analysis?
1. A meta-analysis combines results from several studies into one large study. This increases the sample size of the analysis beyond the sample size of any one study.
Give an example of the type of data that would be appropriate for a two-variable chi-squared test?
1. Responses will vary, but correct student responses will consist of two categorical (i.e., nominal or ordinal) variables.
To calculate a chi-squared observed value, it is first necessary to find a series of expected counts. What is the meaning of these expected counts?
1. Expected counts are the number of individuals that would be expected in the cell in a contingency table if the data perfectly matched the null hypothesis.
What is the purpose of creating a contingency table in a chi-squared test?
1. The contingency table displays all possible combinations of the independent and dependent variable values and the number of sample members in each group. This information can then be used to calculate expected counts, which are necessary to calculate the chi-squared observed value.

Chapter 15

Which procedure(s) was/were designed to analyze multiple dependent variables simultaneously?
1. Nonparametric statistics
2. Multiple regression
3. Logistic regression
4. Multivariate statistics
Which is the proper sequence for designing a research study?
1. Formulate research questions 🡪 Choose methodology 🡪 Plan statistical analysis
2. Create research survey 🡪 Formulate research questions 🡪 Plan statistical analysis
3. Plan statistical analysis 🡪 Choose methodology 🡪 Formulate research questions
4. Select sample 🡪 Plan statistical analysis 🡪 Formulate research questions
Which procedure is appropriate when a dataset has two or more independent variables and one interval-level dependent variable?
1. Multiple regression
2. Logistic regression
3. Kruskal-Wallis test
4. Wilcoxon rank-sum test
Comparing ______________________ can help determine which independent variable has the strongest relationship with the dependent variable.
1. Logistic regression results
2. Sample sizes
3. β weights
4. p-values
The null hypothesis in multiple regression is
1. r = 0
2. b = 0
3. α = 0
The effect size for multiple regression is
1. r²
2. Cohen’s d
3. The main effect
4. R²
An interaction occurs when
1. The effect size in a multiple regression increases when independent variables are added to the data
2. The impact of an independent variable varies according to the value of another independent variable
3. The sample size increases in a study due to the recruitment of more subjects
4. There is a relationship between the independent and dependent variables in a one-way ANOVA
Multicollinearity occurs when
1. Multiple regression fails to reject the null hypothesis
2. The addition of an independent variable lowers a statistical model’s effect size
3. Independent variables are highly correlated with one another
4. The effect size is lowered because of the presence of an interaction
A two-way ANOVA has
1. Two independent variables
2. An independent and dependent variable
3. No ability to test for an interaction
4. Cannot produce just one effect size value
On a line graph, the presence of an interaction is indicated by
1. Parallel lines
2. Nonparallel lines
3. Vertical lines
4. None of the above
When an interaction is present
1. The relationship between the independent variables and the dependent variable is nuanced
2. Results will not replicate in a future study
3. The assumptions of multiple regression have been violated
4. The effect size is not trustworthy
Dummy coding is a way of preparing _____________ data for analysis in multiple regression
1. Dependent variable
2. Ordinal
3. z-score
4. Nominal
Logistic regression requires
1. z-score data
2. Multiple dependent variables
3. A nominal-level dependent variable
4. All of the above
The effect size for logistic regression and multiple logistic regression is
1. η²
2. The odds ratio
3. Cohen’s d
4. R²
Simpson’s paradox occurs when
1. Independent variables in a multiple regression are statistically significant, but the effect size is close to zero and indicates that the overall model is not statistically significant
2. An independent variable’s β value changes its sign when another independent variable is added to the statistical model
3. The results in a study contradict previous literature, yet are still correct
4. A sample produces strong results for independent variables that are weakly correlated with the dependent variable
A 3 x 4 ANOVA has ___________ independent variable(s) and ___________ dependent variable(s).
1. 2, 1
2. 3, 4
3. 2, 12
4. 12, 1
Although nonparametric statistics have the advantage that they can be used for nominal and/or ordinal data, what is their disadvantage?
1. It is impossible to convert interval-level data down to the level needed to conduct nonparametric statistics
2. Nonparametric statistics cannot detect relationships between independent and dependent variables
3. Nonparametric statistics have lower statistical power than parametric statistics
4. All of the above
Spearman’s ρ is a nonparametric statistic that most closely resembles
1. ANOVA
2. Tukey’s post hoc test
3. The unpaired-samples t-test
4. Pearson’s r
What is the minimum level of data that Spearman’s ρ is appropriate for?
1. Nominal
2. Ordinal
3. Interval
4. Ratio
The null hypothesis for the Wilcoxon rank-sum test is that
1. The correlation between two nominal variables is zero
2. The standard deviations of two groups are equal to one another
3. Two groups’ medians are equal
4. The distributions of ranks for two groups’ ordinal data are equal
The AUC is the effect size for a Wilcoxon rank-sum test. The range of AUC values is
1. -∞ to +∞
2. -3 to +3
3. -1 to +1
4. 0 to +1
What is the interpretation of the AUC effect size?
1. The probability of randomly selecting a person in the first group who outscores a randomly selected person from the second group
2. The probability that a person in the group of interest will experience the outcomes of interest
3. The number of standard deviations between medians in the two groups
4. None of the above
The Kruskal-Wallis test most closely resembles
1. ANOVA
2. Tukey’s post hoc test
3. The unpaired-samples t-test
4. Pearson’s r
Meta-analytic thinking can involve
1. Reporting as much detailed statistical information as possible
2. Provide enough information for a reader to replicate the study
3. Considering the results of a study in the wider research context
4. All of the above
The alternative of conducting multivariate statistics is to
1. Convert dependent variables into independent variables
2. Use parametric statistics in the analysis
3. Analyze each dependent variable separately
4. All of the above
A latent variable is a variable that is
1. Created by summing together the scores on a series of dependent variables
2. A theorized underlying variable that is the common source of the scores on multiple dependent variables
3. Cannot be observed without distorting the data in the dependent variables
4. All of the above

Factor analysis, structural equation modeling, and MANOVA are all types of
1. Multiple regression
2. Nonparametric statistics
3. Descriptive statistics
4. Multivariate statistics
What are the two data characteristics that determine the appropriate statistical analysis procedure?
1. The number of independent and dependent variables and the level of data for each variable
What is the interpretation of the R² effect size?
1. The R² effect size is interpreted the same way as the r² effect size. That is, it is a measure of the improvement of predictions when using the statistical model compared to the baseline model (or the percentage of dependent variable variance shared with the independent variable’s variance).
Why is it important to investigate the presence of an interaction?
1. Because sometimes the overall average impact of a treatment may not apply to a particular sample member—or any sample member at all. Interactions tell researchers whether an intervention will be more effective for some sample members than for others.
2. Also acceptable: if a student makes the same point but in terms of the strength of a relationship between an independent variable and a dependent variable varying according to individuals’ scores on a second independent variable.
What is the purpose of the Variance Inflation Factor (VIF) or Tolerance statistics?
1. Both of these statistics test for the presence of multicollinearity among independent variables.
Explain why Simpson’s paradox can be difficult to detect.
1. Detecting Simpson’s paradox can be difficult because it requires including all relevant independent variables that contribute to the paradox in the statistical model. However, there is no way to know beforehand which variables may be relevant.

Document Information

Document Type:

DOCX

Chapter Number:

All in one

Created Date:

Aug 21, 2025

Chapter Name:

Statistics for the Social Sciences 1st Edition Test Bank

Author:

Russell T. Warne

Connected Book

Statistics for the Social Sciences 1st Edition Test Bank

By Russell T. Warne

Test Bank General

View Product →

Explore recommendations drawn directly from what you're reading

Statistics for the Social Sciences 1st Edition Test Bank

DOCX Ch. All in one Current

All chapters in this product are shown above

$24.99

100% satisfaction guarantee

Buy Full Test Bank

Quick Navigation

Preview Document Info Connected Book Recommendations

Benefits

Immediately available after payment

Answers are available after payment

ZIP file includes all related files

Files are in Word format (DOCX)

Check the description to see the contents of each ZIP file

We do not share your information with any third party

Warne 1st edition Full Test Bank

Document Information

Connected Book

Statistics for the Social Sciences 1st Edition Test Bank

Explore recommendations drawn directly from what you're reading

$24.99

Quick Navigation

Benefits

Report Unauthorized Use

Added to Cart