Deck 5: Hypothesis Testing: Basic Principles

Question 1

The term "Type II error" refers to

A) mistakenly concluding that variables are unrelated when a true relationship exists.
B) mistakenly concluding that there is a true relationship when variables are really unrelated.
C) failing two recognize how sampling bias may have affected the research findings.
D) using a non-directional research hypothesis when a directional one would have been justified.

Answer 1

mistakenly concluding that variables are unrelated when a true relationship exists.

Question 2

The "null hypothesis" states that:

A) a relationship between variables within a research sample is a true relationship.
B) a relationship between variables will not exist within a research sample.
C) a relationship between variables within a research sample was caused by sampling bias.
D) a relationship between variables within a research sample was caused by sampling error.

Answer 2

a relationship between variables within a research sample was caused by sampling error.

Question 3

A null research hypothesis predicts that:

A) variables will be found to be related, but in an unknown direction.
B) variables will found to be unrelated.
C) variables will be found to be related and in a certain direction.
D) variables will not be found to be related in a research sample, but they are related in the population from which it was drawn.

Answer 3

variables will found to be unrelated.

Question 4

What factor that influences the statistical power of an analysis of the relationship between variables can be most easily controlled?

A) sample size
B) the actual strength of the relationship between variables
C) the rejection level that should be used
D) the amount of variability in the values of the variables

Question 5

A p value is the mathematical probability that a relationship between variables within a research sample could have been produced by:

A) unethical research methods.
B) sampling error.
C) sampling bias.
D) some other variable or variables.

Question 6

If a researcher reduces the likelihood of making a Type I error by using a .01 rejection level rather than the traditional .05 level, she:

A) also reduces the likelihood of commiting a Type II error.
B) increases the statistical power of the data analysis.
C) increases the likelihood of commiting a Type II error.
D) eliminates the possibility that sampling error might produce a relationship between variables within a research sample.

Question 7

The natural tendency of research samples to differ from the population from which they were drawn and from each other is known as:

A) sampling bias
B) random error
C) sampling error
D) design flaw

Question 8

A statistically significant relationship between variables within a research sample may not be:

A) valuable to the social work practitioner.
B) surprising or previously unknown.
C) very strong.
D) all of the above

Question 9

When variables are related within data drawn from a research sample, the relationship may have been caused by:

A) design flaws
B) rival hypotheses
C) sampling bias
D) all of the above

Question 10

A "null finding" that two variables believed to be related are really unrelated:

A) is of no value.
B) is less valuable than if the variables had been found to be related.
C) can be just as valuable as the finding of a relationship between the variables.
D) suggests that the research was methodologically flawed in some way.

Question 11

When social workers tend to answer questions about social issues based on what they think they should believe about them (rather than what they really believe) this is an example of:

A) systematic error
B) random error
C) social desirability bias
D) cultural bias

Question 12

Effect size refers to:

A) how strongly the values of one variable affect the values of another variable.
B) the mathematical probability that a relationship between variables within a sample may have been caused by sampling error.
C) the likelihood that design flaws may have produced a relationship between variables within a research sample.
D) the ability of a statistical analysis to detect a true relationship between variables.

Question 13

If you were to have simply guessed the answer for each of the two previous questions, what is the mathematical probability that you would have gotten them both right?

A) one in twelve
B) one in sixty-four
C) one in sixteen
D) one in four

Question 14

A statistically significant relationship between variables within a research sample is one in which:

A) The relationship must be a true one.
B) The relationship cannot possibly have been caused by sampling error.
C) The relationship cannot have been caused by research design flaws, sampling bias, or other variables.
D) The mathematical probability that sampling error could have caused the relationship is very small.

Question 15

A "true relationship" between variables is one that exists:

A) within a research sample.
B) because of rival hypotheses.
C) within the population from which a research sample is drawn.
D) because of sampling bias