Deck 13: Testing for Significance in Multiple Groups: the Analysis of Variance Statistic

Question 1

An independent variable with 6 values (groups)can be used with a the t-test,but you would have to make several ?t-tests.Of the various problems below,which is the one reflected by making multiple t-tests?

A)you would have an experiment-wise error rate problem
B)the groups would present a normal-distribution problem
C)there is no way to choose correctly which t-test you should use
D)there is no problem with making multiple t-tests as long as you meet other assumptions

Answer 1

A

Question 2

The analysis of variance is part of the nonlinear specific model,a family of statistical techniques.

Answer 2

False

Question 3

The analysis of variance is designed for which level of measurement?

A)nominal
B)ordinal
C)interval
D)ratio

Answer 3

C

Question 4

A reasonable alternative to the ANOVA statistic is the

A)Chi-square test
B)Kruskal-Wallis H test
C)Mann-Whitney U test
D)Fisher's Exact test

Question 5

Multiple comparison tests compare only the highest and lowest group values.All others have to be estimated using.

Question 6

If data are not normally distributed,which of the two inferential tests below should not be used?

A)Chi-square and t-test
B)Chi-square and analysis of variance
C)T-test and analysis of variance
D)All of the tests can be used,a normal distribution is not required for any of them.

Question 7

The concept of "error" is expressed in the ANOVA by

A)the total mean sums
B)the between sums of squares
C)the within sums of squares
D)the between mean squares

Question 8

The ANOVA cannot be used for a 2-group comparison.

Question 9

If you are doing 20 two-at-a-time t-tests,you would expect 1 test result to be the result of error.

Question 10

If an analysis of variance is significant,you should use a multiple comparison test to locate groups with significantly different comparisons.

Question 11

A determination of which groups,if any,are significantly different from each other cannot be done with the analysis of variance test.

Question 12

The best approach,of those listed below,to determining differences between individual groups is to use

A)the analysis of variance test
B)the Bonferroni test
C)the t-test
D)the Mann-Whitney U test

Question 13

One of the assumptions of the Analysis of Variance test is homogeneity of variances.If the data violate this assumption they are referred to as:

A)nonhomogenic
B)heteroscedastic
C)homoscedastic
D)lacking variation

Question 14

ANOVA can accommodate multiple dependent variables.

Question 15

How many groups can be used in the analysis of variance?

A)4
B)as many as you want
C)only 2
D)between 3 and 15

Question 16

If there are multiple violations of the assumptions of ANOVA,it doesn't really make a difference because ANOVA is very robust to these violations.

Question 17

If the group variances are unequal,you should use the unequal variance ANOVA test.

Question 18

Which reason below best answers why a two-tailed hypothesis should be used with an analysis of variance.

A)the probability levels for ANOVA do not easily divide by 2 as in other tests
B)the dependent variable in an ANOVA is nominal level
C)the normal distribution assumption is not used,so a one-tailed curve does not exist
D)it is too difficult to make predictions for multiple group comparisons in a single statement

Question 19

Only a one-tailed hypothesis should be used with the analysis of variance statistic.

Question 20

The effect of an independent variable in the ANOVA is found

A)in the degrees of freedom used to create the F-ratio value
B)in the product of the between mean squares divided by the within mean squares
C)in the product of the between sums of squares divided by the within sums of squares
D)you can't find this affect in ANOVA,you have to use a multiple comparison test