Deck 13: Interpreting the Strength and Importance of Relationships

Question 1

We would increase statistical power if we:

A)Increase the sample size.
B)Assume a smaller effect size in the population.
C)Decrease the significance level from .05 to .01.
D)Decrease the sample size

Answer 1

Increase the sample size.

Question 2

A statistically significant finding:

A)Is substantively significant.
B)Would certainly be substantively significant if the effect size is strong.
C)Could not be substantively significant if the effect size is weak.
D)None of these

Answer 2

None of these

Question 3

Which of the following is true about correlation statistics?

A)For two variables to be correlated, there can be no exceptions to the pattern in which increases in one variable is accompanied by increases in the other variable.
B)Two variables can be correlated if increases in one variable are accompanied by decreases in the other variable with most cases, but not necessarily with all cases.
C)Two variables cannot be correlated if one decreases while the other increases.
D)When both variables are at the ratio level of measurement, a nonparametric correlation coefficient must be used.

Answer 3

Two variables can be correlated if increases in one variable are accompanied by decreases in the other variable with most cases, but not necessarily with all cases.

Question 4

A Cohen's d of .30 and a correlation coefficient of .30 both signify the same degree of relationship strength.

Question 5

Which of the following statements is/are true about meta-analyses?

A)They are unable to merge disparate findings from disparate studies.
B)Their statistical procedures ensure the avoidance of the problem of author biases.
C)Their statistical procedures ensure the avoidance of the problem of methodological flaws in the reviewed studies.
D)Their findings could be skewed by the problem of publication bias (or the file drawer effect).

Question 6

Effect-size statistics enable us to compare the effects of different interventions across studies using different types of outcome measures.

Question 7

Which of the following is true about correlation statistics?

A)A negative correlation of -.50 is weaker than a positive correlation +.30.
B)A positive correlation of +.50 is stronger than a negative correlation -.50.
C)A negative correlation of -.30 is stronger than a positive correlation of +.25.
D)A negative correlation of -.30 is stronger than a positive correlation +.30.

Question 8

Which of the following statements is true about correlation statistics?

A)They cannot be calculated with nominal level variables.
B)They can only be calculated with ratio-level variables.
C)They can only be calculated with ordinal level variables.
D)They can be calculated with variables at any level of measurement.

Question 9

All strong and statistically significant relationships are considered to be substantively significant.

Question 10

Interpreting the strength and practical implications of effect sizes based on correlation coefficients is:

A)The same as interpreting percentage scores of exams.
B)The same as interpreting correlation coefficients in studies of instrument reliability.
C)Less meaningful than just asking whether the correlation is statistically significant.
D)None of these.

Question 11

Which of the following is true about effect-size statistics?

A)A coefficient of determination that equals .09 signifies a trivial relationship.
B)The odds ratio is an effect-size statistic used in logistic regression with dichotomous data.
C)Eta is an effect-size statistic for two ratio-level variables.
D)An eta of .50 is weaker than a Cohen's d of .60

Question 12

If statistical power is .80, then the probability of a Type II error is:

A)0.80
B)1.80
C)0.20
D)0.60

Question 13

According to approximate guidelines for interpreting the strength of effect sizes, a Cohen's d of .55 is considered to be:

A)Very strong
B)Strong
C)Moderate
D)Weak

Question 14

In an evaluation of an intervention that seeks to increase self-esteem scores, the experimental group mean is 40, and the control group mean is 22.The pooled standard deviation is 6.Therefore, Cohen's d equals:

A)0.60
B)1.0
C)3.0
D)18.0

Question 15

If a correlation coefficient equals .30, then:

A)The relationship is weak.
B)9 percent of the dependent variable variation is explained.
C)30 percent of the dependent variable variation is explained.
D)The relationship is strong

Question 16

According to approximate guidelines for interpreting the strength of effect sizes, a Pearson's r of .30 is considered to be:

A)Strong
B)Moderate
C)Weak
D)Very weak

Question 17

If Study A reports a Cohen's d of .60, and Study B reports a correlation coefficient of .60, then:

A)Both studies have reported equal effect sizes.
B)Study A's effect size is stronger than Study B's effect size.
C)Study B's effect size is stronger than Study A's effect size.
D)Both studies have reported substantively significant findings.

Question 18

Trivial relationships:

A)Can be statistically significant with very large sample sizes.
B)Are more likely to be statistically significant with smaller sample sizes.
C)Are never statistically significant.
D)None of these

Question 19

Which of the following is true about correlation statistics?

A)A negative correlation is always undesirable.
B)A strong positive correlation is always desirable.
C)A weak correlation in one study might be more desirable than a strong correlation in another study if the two studies differ with regard to what variables are being correlated.
D)A strong correlation is always more desirable than a weak correlation,

Question 20

A correlation of -.40 indicates that 40 percent of the variation in the dependent variable is explained.

Question 21

Some studies can have sample sizes that are so small that researchers have little likelihood of being able to refute sampling error even with effect sizes that most would agree are substantively or clinically significant.

Question 22

Clinical significance can refer to the meaningfulness and practical value of the benefits of an intervention for each individual recipient of the evaluated intervention.

Question 23

The probability of committing a Type II error is one minus statistical power.

Question 24

Each of the three statistical approaches that have been proposed for measuring clinical significance is considered by some to be controversial.

Question 25

Differences in effect sizes automatically indicate which intervention is more effective or valuable.

Question 26

The statistical power of a study is its probability of correctly rejecting a null hypothesis that is false.

Question 27

Substantive significance refers to how meaningful a finding is to clients, significant others, society or practitioners concerned about a problem.