Chapter 13- Quasi experiments and small N-designs

Question 1

Quasi experiments

Answer 1

Different from true experiments- the researchers do not have full experimental control. They select an independent and dependent variable, then study participants who are exposed to each level of the independent variable. Participants are not assigned to each level by researchers- assignment occurs by choice, acts of nature, teachers, political regulations, etc.

Question 2

Quasi-independent variable

Answer 2

The independent variable in a quasi-experiment, researchers do not have full experimental control over this variable.

Question 3

4 types of quasi-experiments

Answer 3

1. Nonequivalent control group posttest-only design
2. Nonequivalent control group pretest/posttest design
3. Interrupted time-series design
4. Nonequivalent control group interrupted time-series design

Question 4

Nonequivalent control group posttest-only design

Answer 4

Organ donation study- the quasi-independent variable is the two default options for consent to organ donation, and the dependent variable is the rate of organ donation. Researchers can’t control which countries had which defaults, and people weren’t randomly assigned to live in certain countries. This design is called nonequivalent control group posttest only design because the participants were not randomly assigned to groups and were tested only once, after exposure to one level of the independent variable.

Question 5

Nonequivalent control group pretest/posttest design

Answer 5

People can’t be randomly assigned to undergo plastic surgery, but a research team recruited a group of participants already getting plastic surgery and measured certain variables (self esteem, etc.) before surgery and at intervals afterward. The control group was a group of people who had registered at a plastic surgery clinic but decided not to get any procedures. Independent variable is having cosmetic surgery or not, dependent variable is the measures of self esteem. This is a nonequivalent control group pretest/posttest design because the participants were not randomly assigned to groups and were tested both before and after some intervention.

Question 6

Interrupted time-series design

Answer 6

When 13 reasons why was released, researchers used social media data to determine when public attention was highest (April 2017). They then checked to see if suicide rates increased during that month (there was, compared to the suicide rates in April of previous years). This was an interrupted time-series design because it measures a variable repeatedly (suicide rates) before, during, and the after the “interruption” caused by an event (the show’s release).

Question 7

Nonequivalent control group interrupted time-series design

Answer 7

Did Florida’s 2010 policies to control opioid drug use work? Researchers tracked opioid overdose deaths from 2003-2013 in Florida and North Carolina. Rates were increasing in both places until 2010, but started to decline in Florida only afterward. This design combines the nonequivalent control group design (states weren’t randomly assigned to have the laws) and interrupted time-series design (no experimental control over the year that the laws were passed). Quasi independent variables- whether the state passed the laws and time period (before and after the laws).

Question 8

What is the biggest validity concern for quasi-experiments?

Answer 8

Internal validity due to lack of full experimental control

Question 9

Selection effects

Answer 9

Only relevant for independent groups designs, not repeated measures. This threat applies when the kinds of participants at one level of the independent variable are systematically different from those at the other level.

Question 10

Wait list design

Answer 10

All participants plan to receive treatment but are randomly assigned to do so at different times. Controls for selection effects.

Question 11

Design confound

Answer 11

Some outside variable accidentally and systematically varies with the levels of the targeted independent variable. Ex- if another government policy or knowledge of organ donation influenced the results in ALL countries with an opt in/opt out organ donation policy.

Question 12

Maturation threat

Answer 12

Occurs with a pretest/posttest design, when an observed change could have emerged spontaneously over time. Did plastic surgery really cause an improvement in self esteem, or do those things normally improve over time? A comparison group can resolve this threat

Question 13

History threat

Answer 13

When an external, historical event happens for everyone in a study at the same time as the treatment. Maybe suicide rates increased in 2017 not from 13 reasons why, but from another event like a celebrity suicide.

Question 14

Regression to the mean

Answer 14

Occurs when an extreme outcome is caused by a combination of random factors that aren’t likely to happen in the same combination again, so the extreme outcome gets less extreme over time. Maybe people who got plastic surgery were feeling really bad about themselves at the time, which then improved. However, this is usually only a problem when a group is selected because of their unusually high or low scores. A comparison group also helps to eliminate this.

Question 15

Attrition threat

Answer 15

Occurs when systematic kinds of people drop out of a study over time. Maybe self image only improved in the plastic surgery study because people who were disappointed with their outcome stopped responding to the survey over time.

Question 16

Testing threat

Answer 16

A kind of order effect in which participants tend to change as a result of having been tested before. Repeated testing might cause people to improve regardless of the treatment, or cause performance to decline due to fatigue or boredom.

Question 17

Instrumentation threat

Answer 17

A measuring instrument could change over repeated uses and threaten internal validity. This would occur if one version of a test was more difficult or if coders changed their standards over time.

Question 18

Observer bias

Answer 18

Threatens internal validity when the experimenters’ expectations influence their interpretation of the results. To interrogate- who measured the behaviors? Was the study double blind or blind?

Question 19

Demand characteristics

Answer 19

When participants guess what the study is about and change their behavior in the expected direction. To interrogate- think about whether participants were able to detect the study’s goals and respond accordingly.

Question 20

Placebo effects

Answer 20

When participants improve, but only because they believed they were receiving an effective treatment. You should ask whether the design of the study included a comparison group that received a placebo treatment.

Question 21

Why are quasi experiments important for real life events?

Answer 21

These designs provide real world opportunities for studying important events. Research teams take advantage of events that happen in real world settings when they can’t manipulate variables.

Question 22

How do quasi-experiments enhance external validity?

Answer 22

Because these experiments represent real world settings, they can likely be generalized to other circumstances and individuals. We already know they can apply to real world settings because that’s where they took place.

Question 23

How is the construct validity of a quasi-experiment assessed?

Answer 23

You would interrogate how successfully the study manipulated or measured its variables. Usually, there’s excellent construct validity for the quasi-independent variable. You should also ask how successfully the dependent variable was measured.

Question 24

How is the statistical validity of a quasi-experiment assessed?

Answer 24

You could ask how large the group differences are estimated to be (effect size), and you could evaluate the confidence interval (precision) of that estimate.

Question 25

Are quasi experiments the same as correlational studies?

Answer 25

Both may use independent groups designs. Neither use random assignment, and neither use manipulated variables. Researchers for quasi-experiments also tend to select their samples more intentionally.

Question 26

Henry Molaison (HM)

Answer 26

Suffered from repeated, debilitating epileptic seizures- a surgeon excised the front half of his hippocampus on both the right and left sides. It made sense to do a small case study due to HM’s unique condition. Neuroscientists were able to learn that the hippocampus was responsible for memory. HM’s declarative memory was impaired, but nondeclarative and working memory were intact.

Question 27

Experimental control in case studies

Answer 27

We can draw conclusions from case studies when they are carefully designed. HM’s performance on tasks was compared to controls, and he was given many tasks to assess different types of memory.

Question 28

Disadvantages of small N studies

Answer 28

In the case of HM, brain damage usually involves more than one region, and can therefore influence the outcome. There is also an external validity issue, because participants in small studies might not represent the general population very well. People in brain damage studies often have other health issues, and it would unethical to give others the same brain damage to compare.

Question 29

3 small N designs

Answer 29

1. Stable baseline designs
2. Multiple baseline design
3. Reversal design

Question 30

Stable baseline design

Answer 30

A study in which a practitioner or researcher observes behavior for an extended baseline period before beginning a treatment or other intervention. If behavior is stable at baseline, you can be more sure that the treatment was effective.

Question 31

Multiple-baseline design

Answer 31

Researchers stagger their introduction of an intervention across a variety of individuals, times, or situations to rule out alternative explanations

Question 32

Reversal design

Answer 32

Researchers observe a problem behavior both with and without treatment, but take the treatment away for a while (reversal period) to see whether the problem behavior returns. They then reintroduce the treatment to see if the behavior improves again. This establishes internal validity and can allow for causal claims.

Question 33

Piaget

Answer 33

studied cognitive development in children by observing his own children

Question 34

Ebbinghaus

Answer 34

studied himself to develop the forgetting curve- measuring how fast information is lost

Question 35

Evaluating the 4 validities in small N designs

Answer 35

1. Internal validity- can be very high if the study is carefully designed
2. External validity- can be problematic depending on the goals of the study
3. Construct validity- can also be very high if definitions and observations are precise
4. Statistical validity- not always relevant to small N studies