Deck 9: A Synopsis of Selected Statistical Tests for Examining Nominal Data

Question 1

The chi square statistic and the binomial test are tools for determining statistical significance when the data are measured at the:

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

Answer 1

C

Question 2

The two tests which measure the magnitude of the relationship between two variables measured at the nominal level and are used in examining practical significance discussed in this text are:

A) the phi coefficient and the Binomial test
B) the contingency coefficient and the Fisher Exact probability test
C) the phi coefficient and the p value
D) the phi coefficient and the contingency coefficient

Answer 2

D

Question 3

The test that is useful, when you want to know whether the frequencies in the categories of a nominal variable are significantly different from one another, is:

A) the one-sample t Test
B) the binomial test
C) the contingency coefficient
D) the one-sample chi square

Answer 3

D

Question 4

The most frequently used chi square statistic, in examining questions such as the relationship between ethnicity and employment when measured as nominal variable is:

A) the binomial test
B) the chi square test for independent samples
C) the contingency coefficient
D) the one-sample chi square

Question 5

The McNemar test for the significance of changes examines:

A) the relationship between two nominal variables when the data are drawn from independent samples
B) the relationship between two nominal variables when the data are drawn from related samples
C) the relationship between two nominal variables, measured dichotomously, when the data are drawn from related samples
D) the relationship between two interval variables when the data are drawn from related samples

Question 6

In order to determine whether the frequencies in a dichotomous variable are significantly different from a 0-50 split, with 50% of cases in each of the two categories of the variable, you would engage in:

A) the McNemar's test
B) the binomial test
C) the one-sample chi square
D) the independent samples chi square test

Question 7

For dichotomous variables, the strength of the relationship between the variables and practical significance can be established through:

A) either chi square or the contingency coefficient
B) either chi square or the binomial test
C) either effect size or the p value
D) either the p value or correlation

Question 8

The chi square statistic examines:

A) only the frequency that would occur by chance
B) the expected frequencies when there is a relationship between the nominal variables
C) whether the frequencies in the various categories of a nominal variable are different from expected frequencies
D) the likelihood of the relationship between two ordinal or interval level variables

Question 9

When you compare two independent groups such as males and females on the basis of a dichotomous dependent variable such as "improvement on the social skills" or "doesn't show improvement on social skills" measured at the nominal level, you can use:

A) the McNemar's test
B) the binomial test
C) the one-sample chi square
D) the independent samples chi square test

Question 10

When analyzing data from small samples, Chi square is not an appropriate test:

A) if 80% or more of the cells have expected frequencies less than 5
B) if 70% or more of the cells have expected frequencies less than 5
C) if 20% or more of the cells have expected frequencies less than 5
D) if 66% or more of the cells have expected frequencies less than 5

Question 11

If you were comparing three different treatments on the basis of each person categorized either as recovered or not recovered for each treatment, you would have three groups to compare and therefore you would use:

A) contingency coefficient
B) phi coefficient
C) chi square
D) binomial test

Question 12

An evaluative researcher measured a group of teenage parents on sleep deprivation before and after the birth of their children and categorized each teenage parent as high or low on sleep deprivation according to some threshold of sleep deprivation and decides to use the McNemar test because:

A) this test can be useful in comparing the group who experienced high sleep deprivation before the birth and low sleep deprivation after the birth with the group who experienced low sleep deprivation before and high sleep deprivation after the birth of their children to see if the former is significantly greater than the latter
B) it compares the number of people who experienced a gain in sleep deprivation
C) this test compares the number of people who experienced an improvement in sleep with the number of people who experienced a decline in sleep
D) this test measures the decline in sleep patterns of the teenage parents

Question 13

The McNemar test for the significance of changes is a variation of the:

A) binomial test
B) phi Coefficient measure
C) chi square test
D) t Test

Question 14

When you have the pretest-posttest one group design with the dependent variable that is dichotomous, you will use:

A) McNemar test
B) phi coefficient
C) chi square
D) binomial test

Question 15

Although you can make up your own mind about how noteworthy a given coefficient is by just examining the relationship as displayed in the output table in SPSS, it is a common practice to assume that a phi coefficient of .70 reveals:

A) a weak relationship between the two nominal variables
B) a strong relationship between the two nominal variables
C) a moderate relationship between the two nominal variables
D) absence of a relationship between the two nominal variables

Question 16

The binomial test is an appropriate statistical measure for determining:

A) whether the frequencies between two categories of a nominal variable are significantly different
B) whether the two values of an interval variable are significantly different
C) whether the frequencies between two ratio level measurements are significantly different
D) whether the observations between two categories of a variable are significantly different

Question 17

When you want to know whether a given set of proportions of a dichotomous variable are different from a 50-50 split, you will use:

A) McNemar's test
B) phi coefficient
C) Pearson's coefficient
D) binomial test

Question 18

When a researcher wants to compare the proportion of perpetrators of domestic violence who completed their anger management exercises during the pretest period with the number of perpetrators who completed the anger management exercises during the posttest period, you would use:

A) McNemar's test
B) binomial test
C) one-sample t Test
D) t Test for paired data

Question 19

Phi coefficient is for examining the relationship between:

A) two nominal variables
B) two nominal variables drawn from dependent data
C) for examining the relationship between two dichotomous variables when the data are drawn from independent samples
D) for examining the relationship between two nominal variables with many categories

Question 20

The McNemar's test is used with:

A) data from independent samples
B) comparison group design
C) paired data
D) ordinal data from different groups

Question 21

Sample size is not one of the likely determinants of statistical significance.

Question 22

Phi coefficient is for examining the relationship between two dichotomous variables when the data are drawn from independent data and just like the Spearman rank correlation coefficient, the higher the score the stronger is the relationship between the two variables.

Question 23

The McNemar test is useful for the examination of the relationship between two nominal variables when the data are related.

Question 24

A negative sign in front of the phi coefficient value in the output generated from SPSS is meaningless because you cannot have a positive or negative relationship between two nominal variables as this type of variable puts people into categories that have no order.

Question 25

You cannot report the difference in proportions as your phi coefficient because it is only a quick-and-dirty estimate to be used before you actually calculate the real value.

Question 26

The binomial test could be used with the posttest only data when you have a threshold proportion for comparison.

Question 27

Chi square is an appropriate test when analyzing data from small samples.

Question 28

The contingency coefficient is used when one or more of your two nominal variables have more than two categories.

Question 29

Practical significance is determined by the application of statistical tests.

Question 30

The expected frequency is the frequency for a given cell that would represent a failure to find statistical significance.

Question 31

Explain the importance of the need to observe the expected frequencies in the cells when you compute a chi square test.

Question 32

When you compute chi square, the importance of expected frequencies and the logic behind it is evident in the method of calculating the expected frequencies for the cells. Please explain this process.

Question 33

The phi coefficient does not calculate numerical order of the variables. Please explain.

Question 34

Explain the process of using the Binomial test to compare the categories of a dichotomous variable.