CHAPTER 11
CORRELATIONAL RESEARCH
TEACHING ACTIVITIES FOR CHAPTER 11
Activity 1. Demonstrating a scattergram.
We have found that students can develop a good basic understanding of correlational statistics and correlational
research by having them generate data for two variables and then plot the data on a scattergram. This is more
helpful for many students than to explain correlation in terms of the mathematical logic and procedures for
calculating a correlation coefficient.
You can ask students to write down on a slip of paper (1) the height and weight of an approximately 20-year-old
individual they know. All the students should select an individual who is the gender you specify. Students can use
themselves as the data source, if they wish. Reassure students that the data will be anonymous.
Have someone collect the slips of paper. Construct a scattergram on an overhead or the blackboard, with, let’s
say, 10-pound increments on the weight axis and one-inch increments on the height axis. Ask the person to read
each pair of scores (height and weight) and plot it as a data point on the scattergram.
Activity 2. Group comparisons versus correlation.
To help your students see the similarity and differences between correlational research and group-comparison
research, which is the subject of Chapter 11, you can use the results of the scattergram activity described in Activity
1 above.
Ask a student to read the height and weight on each slip of paper. Draw two columns on an overhead or
blackboard, labeling one column “Height” and the other “Weight.” As the students reads the data on each slip of
paper, write them in the appropriate columns. After this is done, ask students to think about how they could
determine whether the height and weight scores are related to each other without the use of a scattergram.
Activity 3. Analysis and evaluation of the article reprint.
You can ask students to read the reprinted article about teachers’ metacognition at the end of Chapter 11 and
evaluate it using the criteria in Appendix 3 and those specified on pages 280-281. In class, you can call on students
to state what they consider strengths and weaknesses of the study. Also, you can ask them to form small groups in
which they discuss the practical significance of the study’s findings.
MULTIPLE-CHOICE TEST ITEMS FOR CHAPTER 11
1. In correlational research, the presumed cause in a cause-and-effect relationship is the
a. dependent variable.
b. independent variable.
c. y variable.
d. fixed variable.
2. The primary difference between correlational and group-comparison research designs is that correlational
research designs
a. cannot be used to predict matters of interest to educators
b. cannot provide descriptions of samples’ score descriptions.
c. are more similar to experiments in their sampling procedures.
d. typically involve the measurement of more values of variables.
3.correlation coefficients can provide information about
a. the direction, but not the magnitude, of the relationship between two variables.
b. the magnitude, but not the direction, of the relationship between two variables.
c. the direction and magnitude of the relationship between two variables.
d. the extent to which each of two variables are continuous.
4. Scattergrams are used primarily to display
a. a sample’s score distribution on one variable.
b. the effect size for two samples’ score distributions on one variable.
c. the relationship between a sample’s score distributions on two variables.
d. All of the above.
5. Each data point in a scattergram represents
a. the score of one individual on one variable.
b. the scores of one individual on two variables.
c. the reliability of an individual’s scores for two variables.
d. the mean scores of a sample on two variables.
6. A positive correlation between two variables indicates that
a. individuals with higher scores on one variable tend to have higher scores on the other variable.
b. individuals with higher scores on one variable tend to have lower scores on the other variable.
c. the sample tends to have high scores on both variables.
d. one variable has a beneficial effect on the other variable.
7. A line of best fit provides the best prediction of scores on one variable from scores on another variable
if the
a. relationship between the two variables is curvilinear.
b. relationship between the two variables is linear.
c. correlation coefficient has a positive sign.
d. correlation coefficient has a negative sign.
8. High values of a correlation coefficient indicate that
a. the two variables are positively related to each other.
b. the two variables are negatively related to each other.
c. one of the variables has a nonlinear relationship with the other variable.
d. one of the variables is a good predictor of scores on the other variable.
9. The symbol r refers to
a. a correlation coefficient for two variables that are true dichotomies.
b. a correlation coefficient for two variables that are artificial dichotomies.
c. a Pearson product-moment correlation coefficient.
d. a tetrachoric correlation coefficient.
10. Correlational statistics can be used to analyze data
a. from experiments, correlational studies, and group-comparison studies.
b. only from experiments and correlational studies.
c. only fromcorrelational and group–comparison studies.
d. only from studies that used a correlational research design.
11. The usual null hypothesis that is subjected to a test of statistical significance is that the correlation
coefficient for the population represented by the sample is
a. +.05 or –.05 or greater.
b. .00.
c. nonlinear.
d. positive rather than negative in direction.
12. The use of multiple independent variables in a multiple regression equation
a. makes it easier to reject the null hypothesis.
b. simplifies the statistical analysis.
c. can make it more difficult to identify which of the variables predict scores on the dependent
variable.
d. can make it easier to predict scores on the dependent variable.
13. Discriminant analysis and logistic regression are used in situations where
a. the dependent variables are categorical.
b. the independent variables are categorical.
c. at least one of the dependent variables is an interval scale.
d. at least one of the dependent variables is a ratio scale.
14. Canonical correlation is appropriate in situations where there are
a. only two dependent variables.
b. two or more dependent variables that are unrelated to each other.
c. two or more dependent variables that are related to each other.
d. two or more independent variables that are related to each other.
15. Hierarchical linear modeling is used in situations where
a. the variables to be correlated are nested within different organizational levels.
b. independent variables can be ordered in a logical relation to each other.
c. two or more dependent variables are related to each other.
d. All of the above.
16. Path analysis is most similar in purpose to
a. canonical correlation.
b. structural equation modeling.
c. hierarchical linear modeling.
d. product-moment correlation coefficients.
17. In contrast to path analysis, structural equation modeling
a. yields more valid and reliable measures of the variables being measured.
b. yields less valid and reliable measures of the variables being measured.
c. has the goal of helping explain causal relationships between the variables being measured.
d. has the goal of maximizing prediction of a criterion variable from the variables being
measured.
18. A moderator variable is
a. an independent variable that has a nonlinear relationship with another variable.
b. an independent variable whose correlation with a dependent variable is between .30 and .50.
c. a variable that strengthens the correlation between two or more dependent variables.
d. a variable that affects the strength, direction, or both, of the correlation between two other variables.
19. Researchers use factor analysis to
a. determine whether the variables measured in the study reflect a smaller number of underlying
variables.
b. test the validity of the tests used to measure variables.
c. test the statistical significance of a multiple regression.
d. identify which of the independent variables is most effective in predicting the outcome variables.
20. Correlational research has the limitation that
a. only paper-and-pencil instruments can be used to measure the variables of interest.
b. it is not possible to form samples by randomly selecting individuals from a defined population.
c. correlational statistics can be used for prediction but not for explanation.
d. correlational results cannot demonstrate definitively whether two variables are causally related to
each other.
SHORT-ANSWER TEST ITEMS FOR CHAPTER 11
1. A researcher correlates a number of measures that he has collected on a sample of college seniors. He obtains a
correlation coefficient of .65 between the amount of time a student has spent working at a paid job during his
college years and a paper-and-pencil measure of the student’s personal maturity. Determine whether the
researcher would be justified in drawing each of the following conclusions on the basis of this finding, and
explain your answers.
a. The college should institute a work-study program to increase the maturity level of its students.
b. It is probable that the more a student has worked during college, the higher will be his or her score
on the personal maturity measure.
c. More mature students are better able to obtain jobs while in college than their less mature peers.
2. State two types of information that can be provided by a scattergram of the scores on two measured variables
for a research sample.
3. Researchers have measured three predictor variables and want to know how well they predict an
outcome variable. To determine predictability, they correlate each predictor variable with the outcome
variable. Is this the appropriate procedure?
CHAPTER 11
CORRELATIONAL RESEARCH
Multiple-Choice Test Items
Short-Answer Test Items
