Deck 10: Regression Analyses

Question 1

If a researcher wants to predict the value category of the variable "high school graduation" (yes or no) based upon several predictor variables of different levels of measurement (nominal, ordinal, and interval), she should use a form of:

A) logistical regression
B) discriminate analysis
C) multiple linear regression
D) simple linear regression

Answer 1

discriminate analysis

Question 2

A regression equation cannot be used to predict the value of an outcome variable if:

A) the correlation between variables is not a perfect one.
B) the value of the predictor varible is larger or smaller than any value that was used to produce the regression equation.
C) the value of the predictor value falls between two values that were used to produce the regression equation.
D) the value of the predictor variable is not a whole number.

Answer 2

the value of the predictor varible is larger or smaller than any value that was used to produce the regression equation.

Question 3

When a potential predictor variable is only nominal level (e.g., marital status), it can still be included in multiple linear regression analysis by the use of:

A) simultaneous data entry
B) incremental data entry
C) stepwise data entry
D) creation of dummy variables

Answer 3

creation of dummy variables

Question 4

If a correlation between variables is quite weak why would a researcher not conduct a regression analysis?

A) It would be impossible to construct a regression equation.
B) The correlation between variables could not possibly be statistically significant.
C) Predicted values would be so inaccurate as to be virtually useless.
D) all of the above

Question 5

If another student taking the make-up test did not study at all for the test (just like three other students who took the original test), what grade can we predict that she will receive?

A) 85
B) 55
C) 25
D) 65

Question 6

If all of the predictor variables in the previous question had been interval or ratio level and normally distributed, the statistical analysis of choice would have been:

A) logistic regression
B) discriminate analysis
C) multiple linear regression
D) simple linear regression

Question 7

In a regression equation, what symbol is used to denote the point where the regression line (if continued) would intercept the y axis?

A) Y'
B) b
C) a
D) X

Question 8

Marcia also was sick the day of the test. How many hours would we estimate that she would need to study to get a grade of 91 (an A) on the make-up test?

A) 10
B) 6
C) 9
D) 4

Question 9

In a social work research class, the instructor asked each student how many hours he or she had studied for a test and then compared this (predictor) variable with his or her test grade (the outcome variable). Data analysis produced a high positive correlation and the regression equation Y' = 55 + 6(X). Mark was sick when the exam was given. He studied 5 hours prior to taking the make-up test. What is Mark's predicted grade on the test?

A) 61
B) 30
C) 85
D) 94

Question 10

What do statisticians call the estimate of how well a regression line will predict measurements of the outcome variable?

A) the confusion matrix
B) the standard error
C) the slope
D) the discriminant analysis

Question 11

The terms "simultaneous," "heirarchal," and "stepwise" describe different ways that when conducting multiple linear regression analysis.

A) variables are categorized
B) regression equations are created
C) data are entered
D) findings are interpreted

Question 12

When conducting regression analysis, can the predictor variable and the outcome variable be reversed?

A) No, when two variables are correlated one always produces the values of the other and not vice versa.
B) No, it would be impossible to construct a regression equation.
C) Yes, the regression equation would be unchanged. However, it would not produce accurate predictions.
D) Yes, but a different regression equation would be produced.

Question 13

What statistical analysis allows a researcher to predict the value of an interval or ratio level outcome variable from the values of two or more interval or ratio level predictor variables?

A) multiple linear regression
B) logistical regression
C) simple linear regression
D) discriminant analysis

Question 14

What can regression analysis tell the researcher that correlation analysis alone cannot?

A) the strength of a relationship between variables
B) the direction of a relationship between variables
C) the probability that the relationship between variables within a research sample could have been produced by sampling error
D) the predicted value of an outcome variable if we know the value of a case's predictor variable.

Question 15

In a regression equation, what symbol is used to denote the predicted value of the outcome variable?

A) Y'
B) b
C) a
D) X