Deck 13: Nominal Categorical Data and the Chi Square

Question 1

With chi square, the more categories there are, then

A) the fe's must each equal the fo's
B) the fe's must all equal each other
C) no single fe may ever equal a given fo
D) none of these

Answer 1

none of these

Question 2

If a certain nominal scale, applied to a single group of subjects, showed observed frequencies of 30, 25 and 20, the frequencies expected on the basis of chance would be

A) 30, 25 and 20
B) 30, 30 and 25
C) 25, 30 and 25
D) 25, 25 and 25

Answer 2

D

Question 3

If during a presidential campaign, a certain random sample of voters indicated 30 persons for Clinton, 25 for Bush, and 10 for Perot, the resulting scale would be

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

Answer 3

A

Question 4

The degrees of freedom on a 3 by 3 chi square are equal to

A) 9
B) 4
C) 3
D) cannot tell from the above information

Question 5

With chi square, the more categories there are

A) the fewer the degrees of freedom
B) the greater the degrees of freedom
C) the higher the fo values
D) the higher the fe values

Question 6

When using the nominal scale to assess political-party affiliation,

A) the strength of each individual's party loyalty must be established
B) the subjects within each category must be rank-ordered with respect to affiliative strength
C) each and every subject must receive a different scale number
D) the individuals must be independently grouped as to party preference

Question 7

When the chi square fo's deviate significantly from the fe's, then

A) the independent rule has been violated
B) the data are not nominal
C) the chi square cannot be calculated
D) null is rejected

Question 8

When a single group of subjects is allocated among any number of categories, the chance fe's for the chi square are

A) equal to the size of the sample
B) equal to the fo's, divided by the sample size
C) equal to the sample size, divided by the fo's
D) equal to the fo's

Question 9

On a 1 by k chi square with a sample size of 100, the degrees of freedom must always equal

A) 100
B) 99
C) k-1
D) none of these

Question 10

When the chi square value is negative, it means

A) the relationship is inverse
B) the second group is higher than the first group
C) the fo's are greater than the fe's
D) none of these, since obtaining a negative chi square is impossible

Question 11

All nominal categories

A) must share the same trait
B) must overlap
C) must be totally independent of each other
D) must be rank-ordered

Question 12

The degrees of freedom for a 2 by 2 chi square must always equal

A) 4
B) 2
C) 1
D) none of these

Question 13

For any chi square, the degrees of freedom are equal to

A) the size of the sample
B) sample size, minus 1
C) sample size, minus 2
D) none of these

Question 14

When the fe's are determined, not on the basis of strict chance, but on the basis of some previous theory, then the tested hypothesis is

A) accepted
B) rejected
C) no longer appropriate to a chi square analysis
D) a priori

Question 15

When measuring individuals with the nominal scale, the key question is,

A) how much of the given trait does the individual possess?
B) does individual A possess more of the given trait than individual B?
C) does the individual possess or not possess the trait?
D) all of these are key questions when using the nominal scale

Question 16

The major mathematical rule underlying nominal scaling is

A) the greater-than, or less-than rule
B) the equality versus non-equality rule
C) the equal-appearing-intervals rule
D) none of these, since nominal scaling is not mathematical

Question 17

When a single group of subjects is allocated among any number of categories, the chi square analysis is

A) r by k
B) r by 1
C) 2 by 2
D) 1 by k

Question 18

On a 1 by k chi square, the tabled value needed for significance, at a given alpha level, cannot be found unless

A) the number of subjects is known
B) the number of categories is known
C) the number of rows is known
D) all of these

Question 19

The null hypothesis for chi square states that the

A) means for each group are equal
B) means for each group are equal to the fo's
C) fo's are equal to the fe's
D) population mean is equal to the sample mean

Question 20

When the obtained value of chi square is equal to the critical, table value for a given number of degrees of freedom, then

A) Ho is rejected
B) There is a failure to reject Ho
C) the alternate hypothesis is accepted
D) none of these, since no conclusions are possible when the obtained value

Question 21

This is an example of

A) post-facto research
B) experimental research
C) correlational research
D) none of these

Question 22

When several groups of subjects are being compared with regard to some observed frequencies, the appropriate chi square analysis would be

A) 1 by k
B) r by k
C) N by df
D) none of these

Question 23

The independent variable in this study was

A) assigned
B) not allowed to vary
C) manipulated
D) significant

Question 24

The Yates correction may only be applied

A) when any fo is less than 5
B) when any fo is less than fe
C) when chi square equals 0
D) none of these

Question 25

The dependent variable in the study was

A) whether the subjects caught the flu
B) whether the subjects received the vitamin C
C) the control group
D) the placebo

Question 26

A researcher obtains a 2 by 2 chi square value of 3.83. The appropriate degrees of freedom indicate required table values of 3.84 at .05, and 6.64 at .01. The researcher should

A) Fail to reject Ho
B) reject Ho at .05
C) reject Ho at .01
D) only test the a priori hypothesis

Question 27

In setting up the contingency table, whether the groups received the vitamin C should be placed in the

A) columns
B) rows
C) marginal totals
D) lower, right-hand corner

Question 28

When comparing two or more groups with regard to some observed frequencies, the dependent variable should be set in the

A) rows
B) columns
C) marginal totals
D) lower, right-hand corner

Question 29

When Yates is applied, the likelihood of obtaining a significant chi square is

A) slightly decreased
B) slightly increased
C) greatly increased
D) eliminated

Question 30

When doing a 2 by 2 chi square, when any expected frequency is less than 10, the Yates correction suggests that the researcher should

A) reduce the fo-fe value by .5
B) increase the fe by .5
C) reduce the fe by .5
D) not calculate a chi square because of the small fe

Question 31

Regardless of the fo or fe values, the Yates correction need only be applied when

A) the chi square is significant
B) the difference between a given fo and fe is less than 10
C) the df equals 1
D) none of these

Question 32

The independent variable in the study was

A) whether the subjects caught the flu
B) whether the subjects received the vitamin C
C) the control group
D) the placebo

Question 33

The effect of the Yates correction on the chi square value is that chi square is

A) slightly decreased
B) slightly increased
C) greatly increased
D) none of these, since Yates does not influence the chi square value

Question 34

In setting up the contingency table, whether the groups caught the flu should be placed in the

A) columns
B) rows
C) marginal totals
D) lower, right-hand corner

Question 35

When fe's are obtained by multiplying row totals by column totals, and then dividing by N, the chi square is a (n)

A) 1 by k
B) r by k
C) 1 by r
D) all of these, depending on the degrees of freedom

Question 36

Whenever a single group of subjects is being allocated among four independent categories, the chi square analysis would be

A) 1 by 3
B) 1 by 4
C) 2 by 2
D) cannot tell from the above information

Question 37

The measurement scale employed in the study was

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

Question 38

The appropriate analysis of these data would involve

A) a 1 by k chi square
B) a 1 by r chi square
C) a 2 by 2 chi square
D) an F ratio, followed by a chi square

Question 39

In the chi square table of critical values, as degrees of freedom increase

A) so too do the critical values
B) the critical values decrease
C) the critical values remain constant
D) none of these, since for chi square the critical table values are

Question 40

When comparing two or more groups with regard to some observed frequencies, the independent variable should be set in the

A) rows
B) columns
C) marginal totals
D) lower, right-hand corner

Question 41

When using chi square, the larger the alpha level

A) the more difficult it becomes to reject Ho
B) the easier it becomes to reject Ho
C) the greater the degrees of freedom
D) none of these

Question 42

Chi square, like t and F, always tests the hypothesis of

A) association
B) correlation
C) difference
D) a priori events

Question 43

Chi square may never be used with nominal data.

Question 44

The use of (C), the correlational technique cited in the text for nominal data, first requires

A) a manipulated independent variable
B) an a priori hypothesis
C) that the constant, .50, be subtracted from the chi square
D) that the chi square be significant

Question 45

With chi square, the larger the sample size

A) the smaller the alpha error
B) the larger the alpha error
C) the larger the critical, tabled value needed for significance
D) none of these

Question 46

Chi square is a non-parametric statistical test.

Question 47

The McNemar test may be used on nominal data taken from

A) a repeated-measures design
B) a matched-subjects design
C) both of these
D) neither of these

Question 48

In order to use the McNemar test, the samples should be

A) independent of each other
B) dependent on each other
C) free of sampling error
D) normally distributed

Question 49

The null hypothesis for chi square states that the observed frequencies are equal to the expected frequencies.

Question 50

In analyzing the data from this research, the Yates correction should only be applied if

A) one or more of the fo's was less than 10
B) one or more of the fo's was greater than 10
C) one or more of the fe's was less than 10
D) none of these

Question 51

The McNemar test may only be used when the contingency table is

A) 1 by k
B) 2 by 2
C) 2 by 3
D) none of these, since the McNemar test may be applied to contingency tables

Question 52

Whenever the chi square is applied to interval data, the resulting value will be

A) inflated
B) lowered
C) the same as when applied to nominal data
D) none of these, since chi square may not be applied to interval data

Question 53

The chi square is a

A) parametric test of mean differences
B) a non-parametric test of frequency differences
C) a parametric test of frequency differences
D) an interval test of rank differences

Question 54

In evaluating these data, the chi square's degrees of freedom would equal

A) 1
B) 4
C) 2
D) cannot tell from the information provided

Question 55

The chi square statistical test may handle data from

A) experimental research
B) post-facto research
C) both of these
D) neither of these

Question 56

In order to calculate a chi square, it is assumed that

A) the underlying distributions are normal
B) the population mean is a known value
C) the mean of the sample does not deviate significantly from the population mean
D) none of these

Question 57

The chi square test may not be used when

A) the underlying distributions are skewed
B) the variability within the sample differs
C) the intervals between successive scale points are not equal
D) none of these

Question 58

When the 2 by 2 chi square contingency table contains expected frequencies of less than 5, most statisticians agree that

A) the Yates correction must be applied
B) coefficient of contingency must be applied
C) the degrees of freedom then must increase by 5
D) the chi square should not be calculated

Question 59

In order to use the chi square

A) all data must be in nominal form
B) the categories must be independent of each other
C) the sum total of all fo's must equal N
D) all of these

Question 60

The coefficient of contingency (C) may be applied when

A) using a 2 by 2 contingency table
B) using a 3 by 4 contingency table
C) both of these
D) neither of these

Question 61

When any expected frequency is less than 10, some statisticians argue that Yates should be applied when df=1).

Question 62

Every 2 by 2 chi square, regardless of sample size, contains only 1 degree of freedom.

Question 63

The degrees of freedom for chi square are based on the sample size.

Question 64

Correlations may never be established when the data form is nominal.

Question 65

The Yates correction is designed to reduce the value of chi square.

Question 66

The chi square test demands that cell entries be independent of each other.

Question 67

Chi square may only be used to compare frequencies, never sample means.

Question 68

Chi square may not be used when comparing two or more samples.

Question 69

Because of the independence restriction, the McNemar test cannot be used on dependent samples.

Question 70

When applying chi square to a 3 by 5 contingency table, Yates must be used whenever any fe is less than 10.

Question 71

When chi square equals zero, there is a failure to reject Ho

Question 72

When applying chi square to a 2 by 2 contingency table, each fo must be at least equal to 10.

Question 73

Other things being equal, the larger the chi square value, the greater the likelihood of rejecting Ho.

Question 74

Unlike t of F, as the degrees of freedom increase for chi square, the critical, table values also increase.

Question 75

Chi square may only be used for the analysis of post-facto research data.

Question 76

Applying the Yates correction, makes it somewhat easier to accept Ho.

Question 77

In order to use chi square, the underlying distribution must be normal.

Question 78

In a 1 by k chi square, the expected frequency is always equal to k-1.

Question 79

As long as the data are in nominal form, the McNemar test may be applied to results from matched-subjects' designs.

Question 80

When analyzing research data via the chi square, the independent variable should be placed in the rows.