Chapter 7: Experimental Control 1: Obscuring Factors

1
Chapter 7 Experimental Control 1 Obscuring
Factors

• What is experimental control?
• Two kinds of control problems
• Influences that can obscure an effect--Ineffecti
  ve manipulation--Measurement error--Ceiling/Floo
  r Effects--Variability in data
• Reducing these Problems
• Laboratory vs. Field Experiments
• Control Procedures and Significance Tests

2
The Ideal
3
The Reality Many things can affect the DV
4
To see the effect of the IV on the DV we can
5
Two Kinds of Control Problems

• Obscuring Factors
• Confounding Factors or Confounded Variables
• No effect -gt Worry about obscuring factors
• Manipulation did have an effect -gt Worry about
  confounding factors

6
Obscuring Factors

• Ineffective Manipulations
• Measurement Error
• Ceiling/Floor Effects
• Variability

7
Ineffective Manipulations

• Adolph, 1947
• Do rats eat for taste or calories?
• Importance of a manipulation check
• Mood and Memory

8
Measurement Error

• Administer a drug to reduce hunger
• Later ask each subject how hungry he/she is
• Suppose the observed data (subjects ratings of
  hunger do not show expected effect) -- B of
  Figure 7.2
• Possibility of measurement error

9
Measurement Errors
10
Ceiling Effects
11
Reducing Variability Finding an Appropriate
Design
12
Variability The Situation

• Variability in external situation can affect data
• Noise in external environment
• Internal Factors (e.g. tiredness, hunger, etc.)
• Meticulous control?

13
Pavlovs Lab Exterior
14
Pavlovs Lab Floor Plan
15
Pavlovs Lab Floor Plan
16
How Much Control?

• Varies
• Cant control everything
• Control those variables that have the greatest
  effect on the DV
• Randomize what you cant control
• Randomization is the single best method of control

17
Sacrificing Control The Field Experiment

• Lab Simplified Situation
• Field Experiment Experiment in a natural setting
• The Liquor Store Robbery Example
• An Injury-faking Bird

18
A plover faking an injury
19
Ristaus Experiment
20
Field or Laboratory?

• Field Study increases confidence in IVs effect
• A continuum
• Controlling for observer effects bias
• What is a natural setting

21
Degree of Control
22
How Do We Choose?

• Field Experiments Pick up Powerful Effects
• Field Data May be More Generalizable
• The Question of Real Life
• How much to control?

23
Field Experiments

• Large Effects
• Subtle Influences
• If we get an effect in a Field Experiment
• Some effects cant be obtained in the laboratory
  OR vice versa

24
Field Data May be More Generalizable

• External Validity Issue
• Greater variability in conditions under which an
  effect may be obtained

25
Question of Real Life

• Laboratory is an unnatural environment
• Separation of confounded variables
• Does this apply in the natural situation?
• Field Experiment--automatically know whether it
  applies to a natural situation
• Another view Predicting a specific laboratory
  result from a theory

26
How Much to Control?

• Select IV and DV
• Decide what to control and what not
• Literature and other researchers as guide
• Importance of pilot studies

27
Field or Laboratory Experiment?

• Importance of the Power of an Effect
• Power of the Manipulation versus the Power (or
  sensitivity) of the Design
• Field Experiments only Pick up Powerful Effects
• Generalizability
• The Real Life Question

28
Significance Tests Experimental Control

• The importance of proper design
• Significance depends on--The effect of the
  manipulation (treatment)--Our control over
  non-systematic variability (error)--Sample Size
• Trade-offs

29
?
30
Summary

• An experiment looks at effect of one or more
  variables on a Dependent Variable
• Control other (extraneous variables)
• Obscuring Factors--Weak manipulations--Measurem
  ent Error--Ceiling/Floor Effects--Variability
  in data