Stats

Question 1

Selection bias

Answer 1

When people who enter a study earlier may be different from those who enter it later. Best control for this is random assignment

Question 2

Solomon 4 group design

Answer 2

Used to control for testing practice threat to internal validity

• Pre and post w/ intervention
• Pre and post w/o intervention
• Post w/ intervention
• Post wi/o intervention

Question 3

Attrition or experimental mortality

Answer 3

Occurs when subjects drop out of a study.

To assess for attrition is to run t-tests to compare drop outs to those who stayed in the study.

Question 4

Internal validity

Answer 4

How well a study is done in terms of avoiding methodological errors.

When internal validity is strong there is good confidence that any changes in the DV are due to the IV.

Question 5

Construct validity

Answer 5

Determined by if the intended factor of the intervention caused change or if it was some other aspect of the intervention.

Attention and contact with clients
Experimenter expectancies (Rosenthal effect)
Demand characteristics
John Henry effect (compensatory rivalry)

Question 6

Rosenthal effect

Answer 6

Threat to construct validity

Involves wxperimenter expectancies impacting the results.

Experimenter blindness controls this

Question 7

Demand characteristics

Answer 7

Threat to construct validity

When subjects have too much knowledge about their assignment based on experimental procedure

This can be controlled with subject blindness

Question 8

John Henry effect (compensatory rivalry)

Answer 8

Threat to construct validity

When control group tries to compete with the experimental group

To control for this the groups should not know about each other.

Question 9

External validity

Answer 9

Factors that limit generalizability of findings

Sample characteristics (ie volunteers)
Stimulus characteristics ( lab study versus real life)
Context characteristics (reactivity, Hawthorne effect)

Question 10

Hawthorne effect (reactivity)

Answer 10

When subjects in a study behave differently than normal bc they are being observed.

Question 11

Statistical conclusion validity

Answer 11

Statistical problems that result in lack of statistical significance

Low power
Unreliable measures
Variability only procedures
Subject heterogeneity