Deck 23: Inferential Data Analysis

Question 1

If your probability of a Type Il error is .30,then your statistical power is:

A).30
B).05
C).70
D).80

Answer 1

C

Question 2

An a priori statistical power analysis can be used to determine how large a sample to select.

Answer 2

True

Question 3

Ignoring statistical power can result in

A) conducting a study with a sample size that is too small.
B) wasting resources by selecting a sample that is much smaller than needed.
C) overlooking the probability of committing a Type I error.
D) nonparametric outcomes.

Answer 3

A

Question 4

The probability of a Type II error increases when

A) lower significance levels are used.
B) larger sample sizes are selected.
C) a stronger relationship in the population is assumed.
D) statistical power is calculated on a post hoc bsis.

Question 5

The t-test is appropriate to use when testing the significance of the relationship between two nominal level variables.

Question 6

Which statistical significance test should be used in a study of whether the mean number of days worked by case-managed clients is greater than the mean worked by non-case-managed clients?

A) t-test.
B) chi-square.
C) Pearson's product-moment r.
D) Fisher's exact test.

Question 7

Parametric tests assume

A) interval- or ratio-level measurement of all variables.
B) a normal sampling distribution of variables with interval- or ratio-level data.
C) nonprobability sampling.
D) nominal level measurement of all variables.

Question 8

A statistical power analysis tells us:

A)How large our effect size is.
B) The probability of avoiding a Type II error assuming our hypothesis is true.
C) The probability of committing a Type I error.
D) How much substantive significance our findings have.

Question 9

Which of the following statistics can be used to determine whether or not there is a statistically significant relationship between two nominal-level variables?

A) Gamma
B) Chi-square
C) Analysis of variance
D) Pearson's product-moment correlation coefficient

Question 10

Which of the following statistics can be used to determine whether there is a statistically significant difference between three treatment groups in number of days hospitalized?

A) t-test
B) chi-square
C) analysis of variance
D) Pearson's product-moment r

Question 11

Which of the following statements is/are true about a multivariate statistical analyses in which variable A has a .20 correlation with variable C,and variable B has a .20 correlation with variable C?

A)Variables A and B together must accounor 40 percent of the variation in variable C.
B) Variables A and B together must accounor 4 percent of the variation in variable C.
C) If variables A and B are correlated with each other, then variables A and B together will accounor less than 40 percent of the variation in variable C.
D) If variables A and B are correlated with each other, then variables A and B together will accounor more than 40 percent of the variation in variable C.

Question 12

Nonparametric tests are intended for use when some assumptions of parametric tests cannot be met.

Question 13

A rigorous and trustworthy meta-analysis worthy of guiding evidence-based practice decisions should:

A) include published studies, only.
B) be transparent about potential conflicts of interest.
C) calculate one mean effect size based on all included studies, regardless of their methodological strengths or weaknesses.
D) avoid calculating the overall level of statistical significance.

Question 14

If two correlated variables each have a bivariate correlation of .25 with a third variable,then the multiple correlation of the three variables combined will be .50.

Question 15

A researcher conducts 20 tests of statistical significance,one for each independent variable,wita.05 significance level.Only one test yields a probability value that is not greater than .05.What can we infer about the latter test?

A) The probability of committing a Type I error in rejecting the null hypothesis is .05.
B) It is a statistically significant predictor of the dependent variable.
C) Because so many tests were conducted, the probability of committing a Type I error is greater than .05.
D) Because so many tests were conducted, the probability of committing a Type I error is much smaller than .001.

Question 16

Multiple regression analysis shows the overall correlation between each of a set of independent variables and an interval- or ratio-level dependent variable.

Question 17

Chi-square should be used to test the significance of relationships between the ratio level variables.

Question 18

Statistical power increases when

A) larger sample sizes are selected.
B) lower significance levels are used.
C) a weaker relationship in the population is assumed.
D) it is done on a post hoc basis.

Question 19

An experiment is conducted to test the effectiveness of a social work intervention.The experimental group improves more than does the control group,but the results are not statistically significant.This means that

A) the intervention was not effective.
B) the intervention was effective, but not at a significant level.
C) we cannot rule out sampling error in inferring about the effectiveness of the intervention.
D) the sample size was too large.

Question 20

If your statistical power is .60,then your probability of a Type II error is .40.

Question 21

Identify the key things to look for when critically appraising the trustworthiness of meta-analysis.

Question 22

Meta-analysis procedures can be used to calculate the statistical significance of results aggregated across various studies that by themselves did not have statistically significant results.

Question 23

Identify four different tests of statistical significance and the conditions under which each would be used.

Question 24

Describe the controversy regarding statistical significance testing with population data.

Question 25

Define statistical power and explain its utility in planning research as well as in interpreting the implications of findings that are not statistically significant.

Question 26

Two different studies assessing the effectiveness of two different types of social work intervention in preventing psychiatric hospitalization are conducted in two different regions by two different teams of researchers.One study,conducted in a rural area,finds that the mean number of days hospitalized for 10 treated clients is 60,as compared to a mean of 80 for 10 untreated clients.But this difference is not statistically significant.The other study,conducted in an urban area,finds a statistically significant p<05 difference in which 48 percent of 1,000 treated clients received some hospitalization,as compared to 52 percent of 1,000 untreated clients.What can and cannot be inferred about the comparative effectiveness of the two interventions? Explain.Can we infer that one intervention appears to be more effective than the other? Why or why not?

Question 27

One of the best ways to determine which hypotheses have been tested is to firsind which variables are significantly related.