Deck 11: Testing for Relationships

Question 1

If the data for the variables being tested are not normally distributed, the statistical test:

A) cannot be used.
B) is not affected.
C) provides only an approximation of the relationship.
D) must be done on other variables.
E) can control for this by adding additional degrees of freedom.

Answer 1

provides only an approximation of the relationship.

Question 2

A correlation is the statistical test for:

A) testing the linear relationship between two continuous level variables.
B) testing the linear relationships among three continuous level variables.
C) determining degrees of freedom.
D) determining the significance level of other statistical tests.
E) examining two variables of nominal level data.

Answer 2

testing the linear relationship between two continuous level variables.

Question 3

A spurious correlation is:

A) the relationship between two unrelated variables.
B) a nonsignificant relationship.
C) one in which a third variable influences the nature of the relationship between the two variables tested.
D) a positive linear relationship.
E) a negative linear relationship.

Answer 3

one in which a third variable influences the nature of the relationship between the two variables tested.

Question 4

Researchers compute the coefficient of determination to:

A) assess the linear characteristic of the correlation.
B) determine the amount of shared variance for a significant r value.
C) determine the amount of shared variance for an non-significant r value.
D) assess the amount of error in a statistical test.
E) determine if the correlation is positive or negative.

Question 5

A correlation matrix is used to:

A) display how every variable is correlated with every other variable.
B) display the shared variance between pairs of variables.
C) identify all of the possible hypotheses in a research study.
D) determine which relationships are significant and worthy of further study.
E) none of the above.

Question 6

Correlations can also be computed for variables based on ordinal or nominal data. These tests include:

A) point biserial correlation
B) spearman correlation coefficient
C) coefficient of determination
D) a and b
E) b and c

Question 7

Regression is superior to correlation because it:

A) is more flexible.
B) can predict some variables by knowing others.
C) can be used with more than two variables.
D) can test for the influence of more than one independent variable.
E) all of the above.

Question 8

Exogenous variables are:

A) Used in structural equation modeling (SEM)
B) Also known as endogenous variables
C) Variables that are not caused by another variable
D) Variables that are caused by another variable
E) Both a and c

Question 9

Tests for relationship are most commonly computed for variables of continuous level data.

Question 10

Statistical tests for relationship are limited to two variables.

Question 11

Researchers use inferential statistics to determine if the relationship observed in the data is stronger than the relationship that might occur due to chance.

Question 12

Statistical tests were designed to test for relationships and differences on normally distributed variables.

Question 13

The significance level is usually set at .05 as the criterion for making the decision to accept the research hypothesis.

Question 14

Regardless of how the research hypothesis is stated, there is a complementary null hypothesis.

Question 15

To use a correlation as a statistical test, each participant must have provided measurements on two separate variables.

Question 16

Correlation tests for curvilinear relationships.

Question 17

A significant r value would cause the researcher to retain the null hypothesis.

Question 18

Regression is of particular advantage to communication researchers because it allows researchers to study variables that cannot be experimentally manipulated.

Question 19

Because there is no theoretical limit to the number of predictor variables tested in multiple regression, it is common for researchers to include 10 or more.

Question 20

Multiple regression allows the researcher to determine the relative importance of each variable to the regression relationship.

Question 21

Statistically significant results ensure practical application of the results.

Question 22

Explain why statistical tests are designed to test the scores of a sample, rather than the scores of one individual participant.

Question 23

Describe how the researcher interprets a correlation if the .05 significance level is not achieved.

Question 24

Describe the steps the researcher takes to interpret the regression statistic once it can be determined that the test is significant.

Question 25

Describes the steps in interpreting a correlation coefficient.

Question 26

Describe the limitations of correlation as a statistical test.

Question 27

Explain the phrase "variance accounted for"as it would be used with a test of regression.