Experimental Research Strategy

1
Chapter 8

• Experimental Research Strategy

2
The purpose of an experiment is to establish the
existence of a cause and effect relationship.
3
To accomplish this goal

• The researcher must demonstrate that changes in
  one variable cause changes in another variable
• The researcher must rule out the possibility that
  the changes are caused by some other variable.

4
Demonstrating cause
To demonstrate that changes in one variable
causes changes in the other, the researcher must
manipulate the independent variable and observe
the consequence on the dependent variable.
5
Manipulation Checks
Sometimes it is essential that the researcher
demonstrate that the independent variable has
actually been manipulated.
6
Ruling out other variables
To rule out other variables, the researcher must
take steps to prevent confoundings
7
Confounding
A confounding occurs when an extraneous variable
varies systematically with the independent
variable.
8
Types of extraneous variables

• Participant variables Participants differ with
  regard to many characteristics.
• Environmental variables The environmental
  conditions under which a participant is tested
  may vary.
• Measurement variables The process of taking
  repeated measurements from a participant can
  introduce change.

9
Means of controlling extraneous variables

• Control by randomization
• Random assignment of participants to conditions
• Random order of conditions
• Control by holding the variable constant
• Control by matching across treatment conditions
• Control through manipulation

10
Experimental and control groups

• Experimental group a group of individuals who
  receive a particular treatment
• Control group a group of individuals who are
  like the experimental group in all respects
  except that they dont receive a particular
  treatment

11
Placebo Controls
Sometimes just the expectation of a treatment
effect is sufficient to produce one. To control
for these expectancy effects, placebo controls
are used.
12
Placebo Controls
Sometimes just the expectation of a treatment
effect is sufficient to produce one. To control
for these expectancy effects, placebo controls
are used. Placebo controls establish the same
expectancy as the experimental group, but with
out the essential treatment.
13
The problem of external validity
Control of extraneous variables can reduce
external validity.
14
Experimental research designed to improve
external validity

• Simulation experiments
• Field experiments