Title: Conducting Experiments
1Conducting Experiments
2Announcements
- Midterm grades will be posted by Tuesday (this is
my goal) - Mook reading can be found on online reserve (use
same password)
3Outline
- Brief Review of Experimental Research
- Causation
- Manipulation Measurement
- Control Extraneous Variables
- Teasing out Effects Basic Experiments
- Before-after (prettest-posttest) design
- Posttest only design
- The Solomon design
- Artifacts, Experimenter Expectancy, and their
Control - Experiments Realism
4Brief Review of Experimental Research
- Elements of a true experiment
- Establish the existence of a cause-effect
relationship between two variables - Manipulation
- Must rule out alternative explanations
- Control
5Causation
- 3 Criteria of Causation
- Temporal precedence
- Cause before effect
- Covariation
- Cause present, effect present
- Cause absent, effect absent
- Alternate explanations
- Only cause gt effect
6Causation and the 3rd Variable Problem
- Its a Wonderful Experiment
- Experiments show us exactly how the
- world changes when you change only
- one thing
- If you want to know how whether one
- thing alone affects something else, you
- must hold everything else but that one
- thing constant
-
- http//www.angryalien.com/1204/wonderful_lifebuns.
asp
7Causation and the 3rd Variable Problem, cont.
- Freedman et al. (1974)
- Examined trends in family planning and birth
control - Results showed a strong positive relationship
between television watching and use of birth
control - Do you accept this as a cause effect
relationship?
8Causation and the Directionality Problem
- Researchers may establish a relationship between
2 variables - Which is the cause and which is the effect?
- Example (Eagley Steffen, 1984)
- Relation between gender roles and gender
stereotypes
9Manipulation and Measurement
10Manipulation and Measurement, cont.
- Example
- Relation between insomnia and depression
- How can we determine the causal relationship?
11Manipulation and Measurement, cont.
- Manipulation typically involves creating 2
treatment conditions - Examines the relation between the IV and the DV
- Any other variables that exist in an experiment
are called
12Control Extraneous Variables
- Must eliminate all confounding variables
- Characteristics of Confounding Variables
13Dealing with Extraneous Variables
- Control by Holding Constant
- e.g., test Ps in same room, at same time of day,
but same researcher, etc. - Control by Matching
- (e.g., same mean age or IQ per cond.)
- Control by Randomization
14Comparing Methods of Control
15Advantages and Disadvantages of Control Methods
- Holding Constant and Matching
- Useful when one or two variables are identified
as potential confounds - Randomization
16Basic Experiments
17Basic Experiments, cont.
- Prettest-Posttest Design
- Participants are given a pretest before the
experimental manipulation is introduced - Pretest sensitization is a potential problem
18The Solomon Design
- Uses a combination of the posttest-only and
pretest-posttest design
Pretested?
19The Solomon Design, cont.
20Participant-Related Artifacts
- Artifact
- Good subject
- Evaluation apprehension
- How to Overcome?
21Experimenter Expectancy Effects
- Systematic error/bias resulting from the
uncontrolled intentions or actions of the
experimenter - Rosenthal Fodes (1963) study of maze-blind vs.
maze bright rats
22How to Control Experimenter Expectancy Effects
- Blind vs. Double-Blind Procedures
- Blind Experimenters
- Double-Blind Procedure
23Experiments Realism
- Experiments are often criticized because of their
artificial nature - Can we really generalize the results from our
experiments to the real world? - Mook article
- Consider an example
24Aschs (1955) Study of Conformity
- Ps told they were taking part in a study of
visual perception - Asked to make judgments about the lengths of lines
Standard
A B C
25Experiments Realism, cont.
- Two Forms of Realism
- Mundane Realism
- Experimental Realism
26Tuesday
- Between-Subjects Designs
- Cozby, Chapter 8