Deck 7: T Tests and Analysis of Variance

Question 1

Two sample t tests are designed to learn if there is support for a relationship between variables when:

A) both variables are interval or ratio level.
B) the dependent variable is interval or ratio level and the independent variable is nominal level and dichotomous.
C) both variables are nominal level.
D) all of the above

Answer 1

the dependent variable is interval or ratio level and the independent variable is nominal level and dichotomous.

Question 2

The degrees of freedom for the one sample t test are:

A) sample size plus one.
B) sample size plus two.
C) sample size minus one.
D) sample size minus two.

Answer 2

sample size minus one.

Question 3

Simple ANOVA is described as "an omnibus test." What does this mean?

A) It can be used with any level of data.
B) It does not produce a p value.
C) It can substitute for any of the t tests.
D) It is designed to see if there is a relationship between variables, but does not reveal which measurement of the independent variable is most closely related to the dependent variable.

Answer 3

It is designed to see if there is a relationship between variables, but does not reveal which measurement of the independent variable is most closely related to the dependent variable.

Question 4

The two sample dependent t test should be used when two samples consist of the same people or objects measured "before" and "after" some event occurs, or when two samples consist of

A) randomly selected experimental and control groups.
B) two naturally different groups such as males and females.
C) individuals or objects that have been matched to make them comparable in relation to certain variables.
D) all of the above

Question 5

Which statistical test is sometimes used to determine if a research sample is sufficiently representative in relation to some potentially confounding variable?

A) simple ANOVA.
B) two sample independent t
C) one sample t
D) all of the above

Question 6

Using a measurement scale believed to produce interval level measurement, an instructor wishes to learn if the variable "attitude toward statistics" will show improvement over the course of a semester in her class. She administers the scale the first and last days of the class and compares the pairs of scores for every student. What test should she use to learn if there is statistical support for her hypothesis that students will show improved attitudes following her course?

A) one sample t
B) two sample independent t
C) two sample dependent t
D) simple ANOVA

Question 7

When do researchers use t distributions and the family of t tests to seek support for a relationship between variables rather than simply computing confidence intervals based on the normal distribution (as described in Chapter 5)?

A) When the mean of the dependent variable within the research sample is not known.
B) When the mean of the dependent variable within the population is badly skewed.
C) When the standard deviation of the dependent variable within the population is not known.
D) When the dependent variable is less than interval level.

Question 8

Multifactor ANOVA is appropriate for use when there are:

A) one nominal level independent variable and one interval or ratio level dependent variable.
B) two or more nominal level independent variables and one interval or ratio level dependent variable.
C) one nominal level independent variable and two or more interval or ratio level dependent variables.
D) two or more nominal level independent variables and two or more interval or ratio level dependent variables.

Question 9

The non-parametric test that requires at least ordinal level data and that examines the entire distribution of a variable within two samples (and identifies the point at which the two samples differ the most) is:

A) Mann-Whitney U
B) Kolmogorov-Smirnov
C) Kruskal-Wallis
D) Wilcoxon sign

Question 10

A social worker wishes to examine the effectiveness of an innovative method for the treatment of people with a diagnosis of pedophilia. She suspects that it may improve symptoms among some clients, but may make them worse among others. However, she would like to learn if the intervention makes a statistically significant difference in client behaviors. What statistical test should she use?

A) simple ANOVA
B) Mann-Whitney U
C) Kolmogorov-Smirnov
D) Moses test of extreme reactions

Question 11

A non-parametric alternative to the one sample t test is:

A) Mann-Whitney U
B) Wilcoxon Sign
C) Chi-square goodness of fit
D) simple ANOVA

Question 12

The independent t test is often well-suited for comparing the effectiveness of two methods of treatment intervention for addressing some problem such as substance abuse because:

A) the independent variable is nominal level and dichotomous.
B) the dependent variable often can be measured at the interval or ratio level.
C) the two subsamples of clients can be of different sizes.
D) all of the above

Question 13

The parametric test designed to examine the relationship among one dichotomous independent variable and a set of interval or ratio level dependent variables is:

A) Hottellings T².
B) One factor ANOVA.
C) Moses test of extreme reactions.
D) Multi-factor MANOVA

Question 14

If the instructor in the previous question had wished to learn instead whether there was a statistically significant difference in attitudes toward statisics among the three sections of the course that she teaches following completion of it, what statistical test should she use?

A) one sample t
B) two sample independent t
C) two sample dependent t
D) simple ANOVA

Question 15

If the problem being addressed (the dependent variable) in the previous question can be measured at only the ordinal level, what would be an appropriate statistical test to use?

A) Kruskal-Wallis
B) Mann-Whitney U
C) Chi-square goodness of fit
D) Wilcoxon sign