Deck 11: Other Ways That Statistical Analyses Contribute to Evidence-Based Practice

Question 1

Parametric statistical tests generally are considered inappropriate for analysis of single system research designs because:

A) sample sizes are usually too large.
B) the research samples probably should not be considered random.
C) the dependent variable is never at least interval level.
D) the independent variable is not nominal level and dichotomous.

Answer 1

the research samples probably should not be considered random.

Question 2

In social work meta-analysis, the independent variable is always:

A) the method or type of intervention.
B) the problem that the intervention seeks to address.
C) the cost of the intervention.
D) the research design used to evaluate the intervention.

Answer 2

the method or type of intervention.

Question 3

In a program evaluation, qualitative data such as audiotapes of personal interviews must be analyzed. A first step is often to eliminate "noise" prior to transcribing them. To what does the term "noise" generally refer?

A) static created by the recording device
B) background conversations that may have been accidentally picked up
C) the body language of people who provide research data
D) interchanges not relevant to the data being sought, such as exchange of greetings or other social amenities

Answer 3

interchanges not relevant to the data being sought, such as exchange of greetings or other social amenities

Question 4

In what type of program evaluation would a researcher be most likely to use inferential statistical analyses?

A) single system research
B) a needs assessment
C) a formative evaluation
D) an outcome evaluation

Question 5

Why are parametric tests generally considered inappropriate for statistical analysis of meta- analysis data?

A) The dependent variable is never interval or ratio level.
B) The dependent variable is never normally distributed.
C) Research samples cannot be considered random.
D) all of the above

Question 6

In an outcome evaluation, a researcher decides to use two sample independent t to learn whether clients in a program showed more reduction in a problem than those in a randomly selected control group. What is her dichotomous independent variable?

A) gender
B) the amount of reduction in the problem
C) which group clients were in
D) length of the program

Question 7

What types of statistical analysis are most likely to occur in a formative evaluation?

A) frequency distributions and graphs
B) t tests and ANOVA
C) multiple regression
D) logistic regression

Question 8

In analyzing single system data, which statistical test would be appropriate to use if the dependent variable is dichotomous and nominal level (e.g., medication compliance/non- compliance) and the client is in compliance only 30 percent of days during a four week baseline (A) phase?

A) the mean effect size
B) the median test
C) the two-standard deviation method
D) the proportion/frequency method

Question 9

In social work meta-analysis, statistical analysis is used to learn if a method or type of intervention can be considered effective in addressing some problem. In meta-analysis, a research sample is comprised of a group of:

A) clients.
B) social agencies.
C) social programs.
D) reports of research that evaluated the intervention.

Question 10

In an outcome evaluation, Hottelling's T² might be an appropriate statistical test to use if:

A) there are more than two control groups.
B) there are multiple program objectives whose achievement can be measured at the interval or ratio level.
C) there are qualitaive indicators of program success such as direct evaluator observations or notes from focus groups.
D) there are two or more independent variables.

Question 11

The theory underlying the two standard deviation methods for analysis of single system data relies on:

A) the laws of probability.
B) the normal distribution.
C) the construct of the sampling distribution.
D) all of the above

Question 12

What generally has to be done before data collected using qualitative methods can be statistically analyzed?

A) Graphs and frequency distributions must be created.
B) Words must be transformed into numbers.
C) Numbers must be transformed into words.
D) Nothing. Qualitative data cannot be analyzed using statistics.

Question 13

In another outcome evaluation, a researcher wishes to use statistical analysis to learn if a new program produced more reduction in a problem than the program used in the past. What is his dependent variable?

A) the amount of reduction in the problem
B) which program clients were in
C) the length of the two programs
D) the amount that the programs differ from each other

Question 14

A commonly used form of statistical analysis in meta-analysis is:

A) the two standard deviation method
B) the proportion/frequency method
C) the mean effect size
D) all of the above

Question 15

In single system research, research samples consist of:

A) a group of clients who all possess the same problem.
B) the number of A phases and B phases in the research design.
C) the number of clients in a social program.
D) groups of measurements of the target problem in the A phases and B phases of the research design.