Experimentation

EXPERIMENTATION

1

Variables in Experiments

2

•

Independent variables

•

Dependent variables

Advantages of an Experiment?  Researcher’s

ability to manipulate the independent variable  Contamination from extraneous variables can be controlled more efficiently  Convenience  Cost  Replication 3

Disadvantages of Experiments  Artificiality

of the laboratory  Generalization from nonprobability samples  Larger budgets needed  Restricted to problems of the present or immediate future  Ethical limits to manipulation of people 4

Experimentation Process  Select

relevant variables  Specify the treatment levels  Control the experimental environment  Choose the experimental design  Select and assign the participants  Pilot-test, revise, and test  Analyze the data 5

Ways to Assign Subjects  Random

Assignment  Matching Assignment

• Quota matrix

6

Does a Measure Accomplish What it Claims?  Internal

validity

 External

7

validity

Threats to Internal Validity  History  Maturation  Testing  Instrumentation

8

 Selection  Statistical

Regression  Experimental Mortality

Threats to External Validity  The

Reactivity of Testing on X  Interaction of Selection and X  Other Biasing Effects on X

• Artificial setting of testing • Respondents knowledge of testing

9

Experimental Designs  Preexperimental

10

designs

 True

experimental designs

 Field

experiments

Design Symbols

X

the introduction of an experimental stimulus to the participant

0

a measure or observation activity

R

an indication that sample units have been randomly assigned

11

Preexperimental Designs  One-shot

case study

 One-group

 Static

12

pretest-posttest design

group comparison

True Experimental Designs  Pretest-posttest

 Posttest-only

13

control group design

control group design

Operational Extensions of True Designs  Completely

randomized designs  Randomized block design  Latin square  Factorial design  Covariance analysis

14

Field Experiments: Quasi- or Semi-Experiments  Non

Equivalent Control Group Design

 Separate

Sample Pretest-Posttest

Design  Group

15

Time Series Design