Deck 15: Repeated-Measures and Matched-Subjects Designs Interval Data

Question 1

With any t test, paired or independent, the final comparison is always drawn between measures of the

A) independent variable
B) dependent variable
C) control variable
D) none of these

Answer 1

dependent variable

Question 2

When subjects are measured in the "pre" condition, then the independent variable is inserted, and the subjects are measured again in the "post" condition, the experimental design is

A) test-retest
B) repeated-measures
C) post-facto
D) none of these

Answer 2

B

Question 3

The between-subjects design, with random assignment of subjects, should yield between the scores in the two groups.

A) a significant positive correlation
B) a significant negative correlation
C) no correlation
D) all of these, depending on the degrees of freedom

Answer 3

no correlation

Question 4

Whenever the repeated-measures design is used

A) the same subjects are used as their own controls
B) there should be a positive correlation between the pre and post scores
C) the researcher is using the post-facto method
D) both a and b, but not c

Question 5

A t ratio which is calculated on data from a repeated-measures design, using 20 subjects, should yield

A) 19 degrees of freedom
B) 38 degrees of freedom
C) 18 degrees of freedom
D) cannot tell from the above information

Question 6

The paired t is to the independent t, as

A) chi square is to ANOVA
B) the Pearson r is to the Spearman rs
C) the Pearson r is to the chi square
D) the repeated-measures is to the between-subjects design

Question 7

Whenever Ho is accepted, when, in fact, it should have been rejected, the error has been committed.

A) Type 1
B) Type 2
C) standard
D) sampling

Question 8

Whenever the value of the estimated standard error of difference is increased, then

A) the positive correlation was significant
B) the zero correlation was not significant
C) the repeated-measures design was used
D) none of these

Question 9

When a paired t is applied to between subjects data with independently assigned subjects then

A) the positive correlation should be significant
B) the correlation should approximate zero
C) the negative correlation should be significant
D) the multiple R should be used

Question 10

The final goal of the independent t is to

A) test the hypothesis of association
B) test the hypothesis of difference
C) match the independent and dependent variables
D) match the independent and control variables

Question 11

Of the following research designs, the one(s) requiring active manipulation of the independent variable is (are) the

A) between-subjects
B) repeated-measures
C) matched-subjects
D) all of these

Question 12

The final goal of the paired t, is to

A) test the hypothesis of association
B) test the hypothesis of difference
C) match the independent and dependent variables
D) match the independent and control variables

Question 13

For a given number of scores, the independent t has the degrees of freedom as the paired t.

A) one-half
B) twice
C) the same
D) cannot tell from the above information

Question 14

In comparing the between-subjects with the matched-subjects designs, the matched-subjects design should yield

A) a significant positive correlation between the paired scores
B) no correlation between paired scores
C) a significant negative correlation between the paired scores
D) a partial correlation

Question 15

In a paired t, when the correlation is significant,

A) the degrees of freedom become less
B) the alpha error becomes smaller
C) the estimated standard error of difference is increased
D) the estimated standard error of difference is reduced

Question 16

With any t test, independent or paired, the more subjects there are being measured

A) the greater the number of degrees of freedom
B) the fewer the degrees of freedom
C) the greater the alpha level
D) none of these

Question 17

With interval data and random assignment of subjects, the appropriate analysis of data from an between subjects (two group) design would be

A) the paired t
B) the Pearson r
C) the independent t
D) the Spearman correlation

Question 18

Of the following research designs, the one(s) in which the sample scores must be independent of each other, is (are) the

A) between-subjects
B) repeated-measures
C) matched-subjects
D) all of these

Question 19

When experimental research yields a significant correlation between pairs of scores, the experimental design was

A) either between-subjects, or repeated-measures
B) either between-subjects, or matched-subjects
C) either repeated-measures, or matched subjects
D) all of these

Question 20

Of the following experimental designs, the one(s) in which sample scores are correlated with each other is (are) the

A) between-subjects
B) repeated-measures
C) matched-subjects
D) both b and c, but not a

Question 21

Accepting the null hypothesis, when null should have been rejected causes the

A) Type 1 error
B) Type 2 error
C) loss of degrees of freedom
D) all of these, depending on the sample size

Question 22

The relationship between the size of the values for the estimated standard error of difference and the t ratio, is

A) inverse
B) positive
C) curvilinear
D) impossible to determine

Question 23

Whenever the value of the estimated standard error of difference is reduced,

A) the value of alpha is increased
B) the value of the sampling error is increased
C) the value of the t ratio is increased
D) the value of the t ratio is reduced

Question 24

The more powerful a given statistical test is, the greater its likelihood of

A) rejecting Ho
B) failing to reject Ho
C) detecting the standard error
D) reducing the degrees of freedom

Question 25

When equating subjects on a matched-subjects design, the researcher should choose some measure which is correlated with the

A) independent variable
B) dependent variable
C) measurement scale
D) confounding variable

Question 26

When analyzing interval data from a repeated-measures design, with a separate control group, the t ratio must be established between

A) the "pre" and "post" scores
B) the "post" scores, without reference to the "pre" scores
C) the "pre" scores, without reference to the "post" scores
D) the "change" scores of the experimental and control groups

Question 27

Compared to the independent t, the paired t, on data which are really correlated, is

A) a less powerful test
B) a more powerful test
C) equal in power with the independent t
D) none of these, since power is never a function of which statistical test is

Question 28

In a repeated-measures design, with a separate control group that is not matched to the experimental group, the "change" scores should be

A) correlated with each other
B) independent of each other
C) positive in sign
D) none of these

Question 29

When the independent t is used on data which are really correlated, there is an increased likelihood of

A) rejecting Ho
B) failing to reject Ho
C) committing the Type 1 error
D) none of these

Question 30

When attempting to establish differences among three sets of matched interval scores, the appropriate test is the

A) paired t
B) factorial ANOVA
C) within-subjects ANOVA
D) multiple R

Question 31

Compared to the independent t, the paired t for a given sample size always

A) has fewer degrees of freedom
B) has more degrees of freedom
C) uses degrees of freedom which are independent of the sample size
D) none of these, since when sample sizes are fixed, so too are the degrees of

Question 32

Subtracting beta from the constant 1.00, yields

A) the sampling error
B) the deviation error
C) bias
D) power

Question 33

The more powerful a given statistical test is, the more apt it is to

A) reject Ho when Ho should be rejected
B) failing to reject Ho when Ho should be rejected
C) inflate beta
D) trigger a statistical decision

Question 34

In order to calculate a within-subjects ANOVA, one must first

A) establish the value of the multiple R
B) establish the values of the separate t tests
C) establish the value of the Pearson r
D) none of these

Question 35

Rejecting the null hypothesis when it should have been accepted, causes the

A) Type 1 error
B) Type 2 error
C) loss of degrees of freedom
D) all of these, depending on the sample size

Question 36

The within-subjects ANOVA should never be used

A) on interval data
B) on a matched-group design
C) when more than two groups have been equated
D) none of these

Question 37

When the paired t is used on data which are really independent, there is an increased likelihood of

A) Failing to reject Ho
B) rejecting Ho
C) committing the Type 1 error
D) none of these

Question 38

Using a repeated-measures design with a separate control group, helps to prevent

A) sampling error
B) score dispersion
C) individual differences
D) confounding variables

Question 39

The probability of rejecting null always increases when the

A) Type 1 error is decreased
B) sample size is decreased
C) sample size is increased
D) power of the test is reduced

Question 40

When the correlation between two sets of interval measures is significant and positive, the value of the paired t will always be

A) higher than the independent t
B) lower than the independent t
C) equal to the independent t
D) none of these

Question 41

For the t ratio, the only time a negative t value may occur is when

A) sample mean 1 is smaller than sample mean 2
B) sample mean 2 is smaller than sample mean 1
C) the sample means are equal
D) none of these, since t may never be negative

Question 42

With the within-subjects ANOVA, when null is rejected at an alpha of .05, then

A) the alpha becomes .10
B) the alpha becomes .01
C) the alpha becomes .95
D) none of these

Question 43

Regardless of the size of the sample, or the scale of measurement, when correlated samples are involved, the researcher may never use

A) the paired t
B) the within-subjects ANOVA
C) the appropriate degrees of freedom
D) none of these

Question 44

When the probability of beta increases, then

A) so, too, does alpha
B) so, too, does sampling error
C) so, too, does the power of the test
D) power decreases

Question 45

With an equal mean difference, when the correlation is significant, the paired t results in a higher t ratio than does the independent t.

Question 46

Dividing the mean difference by the estimated standard error of difference always yields the

A) standard error of the mean
B) standard error of estimate
C) standard error of proportion
D) t ratio

Question 47

When a researcher manipulates two independent variables simultaneously, the appropriate analysis must always utilize the

A) independent t
B) paired t
C) the one-way ANOVA
D) none of these

Question 48

With a paired t, a significant correlation has the effect of reducing the value of the estimated standard error of difference.

Question 49

When the correlation is not significant, the major difference between the paired and independent t ratios, is in the

A) degrees of freedom
B) the value of the estimated standard error of difference
C) the value of the estimated standard error of the mean
D) the value of the mean difference

Question 50

The paired t should not be used for the analysis of data from independent sample groups.

Question 51

Sampling error is assumed

A) only when the F ratio is calculated
B) only when the t ratio is calculated
C) only when the Pearson r is significant
D) whenever samples are measured

Question 52

When using the independent t on interval data from a one-group, before- after design, there is an increased likelihood of a

A) sampling error
B) Type 2 error
C) Type 1 error
D) rejection of Ho

Question 53

When the calculated value of a paired t, for a given number of degrees of freedom, equals the critical, table value of t, then

A) there is a failure to reject Ho
B) Ho is rejected
C) no statistical decision can be made
D) the alpha error is reduced

Question 54

For a given number of scores, the paired t has one-half the degrees of freedom of an independent t.

Question 55

For a given number of scores, the within-subjects ANOVA

A) always has more degrees of freedom than does the independent F
B) always has the same number of degrees of freedom as the independent F
C) always has fewer degrees of freedom than does the independent F
D) has more degrees of freedom than the independent F when Ho is rejected

Question 56

For the t test, other things being equal, the smaller the estimated standard error of difference, the greater the likelihood of

A) committing the beta error
B) failing to reject Ho
C) rejecting Ho
D) eliminating the test's power

Question 57

Accepting Ho when it should have been accepted causes an increase in the

A) Type 1 error
B) Type 2 error
C) sampling error
D) none of these

Question 58

With ratio data, calculating the t ratio is identical with the procedures used on

A) nominal data
B) ordinal data
C) interval data
D) none of these, since with ratio data the t test cannot be used

Question 59

With any t test, independent or paired, the loss of degrees of freedom causes

A) an increased likelihood of rejecting Ho
B) a decreased likelihood of rejecting Ho
C) an increase in the standard error of difference
D) the mean difference to assume a negative value

Question 60

With a within-subjects ANOVA, a significant difference implies that

A) the difference occurs only in the sample
B) the difference occurs in the population, but not in the sample
C) the difference occurs both in the sample and in the population
D) there is no difference

Question 61

With interval data and a two-group, matched-subjects design, the appropriate analysis is via the independent t.

Question 62

When there is a correlation across the rows, the within-subjects ANOVA results in a higher value than when there is no correlation across the rows.

Question 63

With 20 pairs of interval scores, the paired t has 38 degrees of freedom.

Question 64

When analyzing data from a repeated-measures design, if the correlation is significant, the paired t increases the test's power.

Question 65

The lower the value of beta, the higher the power of any statistical test.

Question 66

An F ratio may never be used on data from a matched-subjects design.

Question 67

The paired t may only be applied to experimental data when there are at least two independent variables.

Question 68

The power of a test is equal to the summation of both the alpha and beta probabilities.

Question 69

For a given mean difference, a significant correlation increases the value of a paired t ratio.

Question 70

With a paired t, the greater the number of paired scores, the greater the number of degrees of freedom.

Question 71

Other things being equal, the smaller the value of the estimated standard error of difference, the less likely the t ratio will be significant.

Question 72

With interval data from a three-group, matched design, the appropriate test would be the within-subjects ANOVA.

Question 73

The use of the repeated-measures design with a separate control group, prevents the possibility of manipulating the independent variable.

Question 74

The beta level defines the probability of accepting null when, in fact, null should be rejected.

Question 75

The within-subjects ANOVA may never result in a negative value.

Question 76

With an extremely powerful test, the probability of a Type 1 error is eliminated.

Question 77

Failing to reject Ho can never occur when alpha has been set at .01

Question 78

When analyzing interval data from a repeated-measures design with a separate control group, an independent t may be applied to the two sets of "change" scores.

Question 79

Degrees of freedom are a direct function of the sample size for all statistical tests except the paired t.

Question 80

The paired t may never result in a negative value.