Internal Validity and Alternative Research Designs

Question 1

What is internal validity?

Answer 1

Deals with experimental control. Extent to which we can be sure the IV is the cause of the Dv (elimination of confounds, nuisance, and bias)

Question 2

What is external validity?

Answer 2

Extent to which we can be sure we can generalize our results to different populations

Question 3

What happens to external validity when internal validity goes up?

Answer 3

External validity goes down.

Question 4

What are the 8 threats to internal validity?

Answer 4

Selection, History, Maturation, Testing, Mortality, Instrumentation, Regression to the Mean, Interactions with Selection (MMM… SHIIT)

Question 5

What is selection (threat)

Answer 5

If groups are not equal prior to the experiment, you don’t know whether the IV has the effect OR if it’s the preexisting condition.

Question 6

What is history (threat)

Answer 6

Something happens between the IV and the measurement of the DV that could effect the response.

Question 7

What is maturation (threat)

Answer 7

Time elapsed between trials or length of experiment can effect response.

Question 8

What is testing (threat)

Answer 8

2 groups are part of the same condition-measuring DV the first time causes a change in effects of the 2nd time, because it teaches participants what the experiment is about.

Question 9

What kind of test gets the biggest effect out of the testing threat?

Answer 9

Microgenetic as they have lots of practice with all conditions in a short amount of time.

Question 10

What is instrumentation?

Answer 10

Measurement criterion changes during testing-can apply to objects and people. Usually happens at the beginning of the experiment, as experimenter is still working out how to word questions and interact with participants.

Question 11

What is regression to the mean?

Answer 11

When we’re at extreme ends, the only place left to go is back towards the middle (issue with testing learning disabilities)

Question 12

What is experimental mortality?

Answer 12

Dropout in studies (longitudinal)- systematic dropout is worse than random dropout as there is a bigger effect on results. Big problem in aging studies.

Question 13

What is interactions with selection

Answer 13

When systematic differences exist between or among groups based on maturation, history, or instrumentation.

Question 14

What is the most obvious fix to internal validity?

Answer 14

Random assignment! Idea is you can get rid of any systematic, unaccounted for differences between groups. Also can equally distribute potential issues

Question 15

What kinds of validity do random selection and random assignment help with respectively?

Answer 15

Random selection- External

Random assignment- internal.

Question 16

What is a pre-test post-test design?

Answer 16

A test where you give 2 groups a pretest to see where they’re at, assign them to a control or IV, then test the DV again.

Question 17

What types of things does the pre-test post test design control for?

Answer 17

Selection (pretest assures groups are equal), regression to mean, maturation, history, testing, interactions with selection

Question 18

What are some issues with the pre-test post-test design?

Answer 18

Expensive, time consuming, not applicable IRL because people don’t do a pretest IRL. Can taking a pretest change how you respond? (testing effects).

Question 19

What is a post-test only design?

Answer 19

Assumes groups are equal from the start so no pretest is done.

Question 20

What are the benefits of post-test only?

Answer 20

Protection from testing effects if you are worried about how the pretest could effect the post test

Question 21

What is the downside of the post-test only?

Answer 21

Protection from selection goes down, as you haven’t done a pretest to assure that the groups are equal.

Question 22

What is the Solomon 4 Group Design?

Answer 22

Combines pre-test and post-test together. Random assignment to 4 groups, 1/2 of sample gets pretest, all get control and IV.

Question 23

What is the main issue with the Solomon 4 Group Design?

Answer 23

Need a MASSIVE sample.

Question 24

What is the main advantage of the Solomon 4 Group design?

Answer 24

There is a gain in external validity

Question 25

What are the group sizes needed for random assignment to work?

Answer 25

n= 20-30

Question 26

What kind of test works best for participatory action research?

Answer 26

Pre-test post-test

Question 27

What kind of test would you use if you had a low n?

Answer 27

PRetest posttest because we aren’t confident that groups will be equal.

Question 28

What are single-case experimental designs?

Answer 28

Experiments with one participant (n=1). You take one person and you study them LOTS. Attempts to determine cause and effect with one person. Normally used in the past due to stats not being invented yet

Question 29

Why would you use a single-case experimental design?

Answer 29

If there are limitations on the opportunity to observe a behaviour, if there is no variability between people, when a negative instance would refute a theor.y

Question 30

What are the 3 types of procedures that are important in a single-case design?

Answer 30

1) Baseline measurement-where is the participant at the start?
2) Change 1 variable at a time-avoid confounding explanations
3) Test again
4) Baseline again
ABAB design.

Question 31

What is a quasi-experimental design? (non-equivalent group design)

Answer 31

When we can’t always randomly assign the IV-instead of assignment, we attempt to match. Compare people in the “IV” to people who are a “close enough match”.

Question 32

What is an interrupted time series design (quasi-experimental)

Answer 32

Test groups… then wait for something to happen. Measure again. Used when you can’t manipulate the IV, also used when a law is passed.