Deck 4: Causation, Classical Experiments and Quasi Experiments

Question 1

What is a rival causal factor?

Answer 1

Anything other than X that explains changes in Y.

Question 2

What is a spurious relationship?

Answer 2

A relationship that is not real (example, Margarine consumption and divorce rates).

Question 3

Name two types of threats to validity.

Answer 3

History, maturation, mortality, selection-maturation interaction, testing, selection bias, statistical regression, instrumentation, diffusion of treatment, compensatory equalization of treatment, demoralization, compensatory rivalry, uncertainty about time-order, statistical conclusion validity)

Question 4

Why is it not advisable to do many criminal justice studies in the laboratory setting?

Question 5

How can we reduce our chances of making a Type II error?

Question 6

In social science statistics, how confident do we like to be before we say we are comfortable rejecting the null hypothesis and declaring support for our research hypothesis?

Question 7

Why do we avoid stating that we "proved" anything in social science research?

Question 8

Name one technique that we can use with a quasi-experiment to get the experimental and comparison groups to be as comparable as possible.

Question 9

If we are studying juvenile delinquency by surveying the same group of students on the first Monday of March in school each year, explain how mortality will likely lead to a biased sample.

Question 10

Provide an example of why diffusion of treatment might occur in a criminal justice program.

Question 11

How does a factorial design differ from a classical experiment?

Question 12

How does reactivity generate problems for generalizability?

Question 13

Define correlation or association.

Question 14

If we find that variable A is correlated with variable B, what can we conclude?

A) A causes B
B) A might cause B, but we really don't know yet
C) B causes A
D) None of the above

Question 15

The _________ represents the researcher's prediction of what kind relationship there is between X and Y.

A) Null hypothesis
B) Research hypothesis
C) Theory
D) Inductive analysis

Question 16

The ________ predicts that there will not be a relationship between X and Y.

A) Null hypothesis
B) Research hypothesis
C) Theory
D) Inductive analysis

Question 17

If we find through statistical analysis that there does appear to be a relationship between X and Y, we__________.

A) Reject the null hypothesis
B) Accept the null hypothesis
C) Accept the research hypothesis
D) Reject the research hypothesis

Question 18

If it appears that ice cream and crime rates are related, but what is actually happening here is that both are related to weather patterns, what kind relationship do we have between ice cream and crime?

A) Causal
B) Incidental
C) Spurious
D) False

Question 19

For the social sciences, how confident do we like to be before we declare that we are ready to reject our null hypothesis?

A) 50%
B) 68%
C) 95%
D) 99%

Question 20

All of the following are necessary components of classical experiments EXCEPT:

A) Randomization
B) Experimental and control groups
C) Instrumentation
D) Pretests and posttests

Question 21

Which of the following is the strongest type of study?

A) Time series
B) One-group before-after design
C) Classical experiment
D) Two group posttest only design

Question 22

Factors other than the independent variable that might impact the dependent variable are known as:

A) Control groups
B) Threats to internal validity
C) Threats to external validity
D) Noise

Question 23

When we select people with extreme opinions or neighborhoods with very high crime rates, we are likely to see change regardless of any type of intervention we might introduce because of ________.

A) Statistical regression
B) History
C) Testing effects
D) Mortality

Question 24

_____________ involves a comparison or control group that tries to outperform the experimental group.

A) Demoralization
B) Compensatory equalization of treatment
C) Compensatory rivalry
D) Diffusion of treatment

Question 25

In the classic Kansas City Preventive Patrol Experiment, police efforts to make their presence known in the reactive beats locations that were supposed to receive the least amount of police attention was known as _____________.

A) Demoralization
B) Compensatory equalization of treatment
C) Compensatory rivalry
D) Diffusion of treatment

Question 26

What is the best way to combat the problem of statistical conclusion validity?

A) Run only multiple regression
B) Increase the sample size
C) Only work with random samples
D) Get a control group

Question 27

All of the following are problematic because they might hide the existence of an actual relationship between X and Y EXCEPT:

A) Compensatory equalization of treatment
B) Demoralization
C) Compensatory rivalry
D) Diffusion of treatment

Question 28

The Solomon four-group design is best for avoiding which threat to internal validity?

A) Testing
B) History
C) Compensatory rivalry
D) Statistical conclusion validity

Question 29

The ___________ is similar to a classical experiment but allows us to test more than two groups, so we can compare different types of interventions.

A) Quasi-experiment
B) Time series
C) Factorial design
D) Meta analysis

Question 30

People attempting to be more productive once they realize they are being supervised is called _______

A) Compensatory equalization
B) Compensatory rivalry
C) Generalizability
D) Hawthorne effect

Question 31

Which if the following is not a threat to internal validity?

A) Statistical regression
B) Noise
C) Mortality
D) Diffusion of Treatment

Question 32

It is easier to find causation between two variables than it is to find a correlation.

Question 33

An apparent association between X and Y that appears because both are actually correlated with Z is called a spurious relationship.

Question 34

Quantitative research is largely based on collection of anecdotal information.

Question 35

Classical experiments are considered to be better designs than quasi-experiments.

Question 36

Quasi-experiments involve random allocation of individuals to experimental and control groups.

Question 37

A time-series is a type of quasi-experiment.

Question 38

Classical experiments tend to be better at controlling threats to internal validity than addressing external validity.

Question 39

Multiple treatment interference is an external validity issue.

Question 40

Causation between two variables is easier to predict in the social sciences than in the natural sciences.

Question 41

What are the three requirements of causation?

Question 42

Explain how classical experiments are designed to address some threats to internal validity. What threats to internal validity remain a problem?

Question 43

Why is it a bad idea to draw conclusions about a subject based on individual cases or anecdotal information? What should we do instead? Provide a real-word example.

Question 44

The _________ hypothesis is what we actually test during statistical analysis.

Question 45

Large research studies tend to focus on ___________, while individuals who witness one event are likely to rely on _________ information.

Question 46

A ____________ is characterized by random assignment to experimental and control groups and pretests and posttests.

Question 47

A ___________ is similar to a classical experiment, except it is missing random assignment to treatment and control groups.

Question 48

________ is the change of data collection tools being used over time.

Question 49

The __________ allows us to test for not only a relationship between X and Y but also to see if the pretest generated any biases.

Question 50

_______ involves pairing people in the experimental group with those in the comparison group to get the two groups to be as similar as possible.

Question 51

A __________ is a type of quasi-experiment with several data collection points.

Question 52

Classical experiments tend to be stronger at minimizing threats to __________ validity than threats to __________ validity.