Types of Bias in Experiments

Name seven ways to minimize experimenter's expectancy

1. Make experimenter unaware of condition
2. Make experimenter unaware of hypothesis
3. Have multiple experimenters within sessions
4. Have multiple experimenters across sessions
5. Make experimenter unaware of condition as long as possible
6. Randomly assign experimenters to conditions
7. Make sure experimenters are trained

Types of Bias in Experiments

Name two ways to minimize experimenter's role in experiment

1. Use other media to communicate instructions

2. Minimize experimenter/subject interactions

Questionable Practices

Limitations of Free Choice

1. Invovling people without their consent
2. Participation through coercion
3. Invasion of privacy

Questionable Practices

Deception

1. Withholding information

2. Providing incorrect information

Importance of debriefing

1. Make sure subjects understand what happened
2. Rectify negative feelings
3. Remove harmful after effects
4. Return subjects same psychological state

The Big 10 Design Errors

1. No control group

2. Using the wrong statistical test

3. Misuse of Correlation (no causality, assuming no correlation)

4. Misuse of ANCOVA (no equivalency b/w groups if no ran. assn)

5. Misue of Standard Error of Measurement (measure of precision not to know which test is better or valid measure)

6. Regression toward the mean (scores becoming less extreme naturally)

7. Sampling problems

8. Using non-equivalent groups

9. Confusing discriminant & convergent validity (want high convergent w/ similar measures)

10. Poor reliability of measure/ Confusing reliability w/ validity

t-test for independent means

Use this test when you have two-groups and a single measurement of the DV

t-test for dependent means

one group and you measure the DV twice. It tells you if the mean of the first measurement is different from the mean of the second.