Deck 12: Testing for Significance in Two Groups: The T-Test

Question 1

The t-test has some stringent assumptions, if the data reasonably depart from those assumptions

A) the t-test results can probably still be used
B) the t-test can no longer be used
C) it is best to use the Levene's Test instead of the t-test
D) you should use both the t-test and the Levene's Test to determine how much assumptions have been violated

Answer 1

A

Question 2

The Levene's test is used for which purpose below?

A) to determine whether a t-test result is significant
B) to test the F-ratio hypothesis that two means are equal
C) to determine whether the data are related or independent
D) to determine which t-test to use

Answer 2

D

Question 3

Assuming all assumptions are met, which version below is the most powerful (accurate) t-test?

A) k-samples t-test
B) unequal variance t-test
C) equal variance t-test
D) independent samples t-test

Answer 3

C

Question 4

The t-distribution for smaller sample sizes is slightly different from a normal distribution because:

A) it has more data in the middle of the curve
B) it is shaped more like a rectangle
C) it is a positively skewed distribution
D) it has more data in the tails of the curve

Question 5

A t-test can compare as many as four groups at a time without losing power.

Question 6

If a variable is measured in an experiment at time 1 (pretest) and again at time 2 (posttest), the most appropriate t-test would be

A) independent samples t-test
B) equal variances t-test
C) related samples t-test
D) none of the above, you cannot use a t-test with this data

Question 7

If you propose a one-tailed hypothesis, you must examine the group means before proceeding.

Question 8

If the F-ratio test (Levene's test) produces a significant result, which of the choices below would be correct?

A) you use the unequal variance t-test
B) you use the equal variance t-test
C) you skip the t-test because the direction is not correct
D) you must complete the t-test because the direction is correct

Question 9

If you have two groups with equal variances, which of the t-tests below would you use?

A) unequal variance t-test
B) equal variance t-test
C) equal means t-test
D) unequal means t-test

Question 10

An equal-variance t-test is best used when the groups have essentially the same amount of dispersion.

Question 11

There are two forms of variation that a t-test can measure. The type of variation used to test your hypothesis is called

A) unequal variance
B) equal variance
C) within-group variance
D) none of the above

Question 12

The t-test is designed to examine the number of cases in two groups to determine which one is the group with the most cases.

Question 13

If the probability associated with a t-test is larger than your alpha level, what do you do?

A) accept your hypothesis and reject the null hypothesis
B) reject your hypothesis and accept the null hypothesis
C) reject both your hypothesis and the null hypothesis
D) use the version of the F-test associated with larger probabilities.

Question 14

The t-test is properly used with which level of measurement below?

A) nominal level
B) ordinal level
C) interval level
D) it can be used with all levels

Question 15

When used with small samples, the t-test produces a conservative estimate of differences.

Question 16

The best alternative to a independent-samples t-test if assumptions are violated is the Mann-Whitney U-test.

Question 17

Which of these assumptions below apply to the t-test?

A) the marginals are equal and normally distributed
B) the dependent variable is normally distributed and there are two groups
C) large sample sizes and equal marginals
D) large sample sizes and normally-distributed independent variables

Question 18

Typical data used with the t-test are those from survey and secondary datasets.

Question 19

The t-test is a

A) nonparametric statistic
B) parameteric statistic
C) predictive statistic
D) descriptive statistic

Question 20

A two-tailed hypothesis is the most appropriate for a t-test because of the difficulty of determining direction in the variances.

Question 21

Because of its stringent assumptions, the t-test is not very commonly used in criminal justice or criminological research.

Question 22

The problem with violating assumptions of the t-test is that probability estimates would have error.