Producing Data: Experiments

1
Chapter 8

• Producing Data Experiments

2
Explanatory and Response Variables

• Response variable measures what happens to the
  individuals in the study
• In an experiment, the investigator controls the
  values of the explanatory variable in individuals
  to see if he or she can influences the response
  variable

3
Experiments Vocabulary

• Subjects people studied in an experiment
• Factors explanatory variables (specific
  experimental condition applied to subjects)
• Treatment a combination of a specific set of
  factors

4
Case Study
Effects ofTV Advertising
p. 200 in text
5
Case Study
Effects ofTV Advertising
Objective To determine the effects of length of
message and how often message is repeated
6
Case Study

• Subjects a sample of undergraduate students
• Subjects viewed a 40-minute television program
  that included ads for a digital camera

7
Case Study

• Factors
• Some subjects saw a 30-second commercial others
  saw a 90-second version
• Commercial were shown either 1, 3, or 5 times
  during the program
• Thus two factors
• length of the commercial (2 levels)
• number of repetitions (3 levels)

8
Case Study

• The 6 combinations of factors, i.e., 6 treatments

Factor B Repetitions Factor B Repetitions Factor B Repetitions
1 time 3 times 5 times
Factor A Length 30 seconds 1 2 3
Factor A Length 90 seconds 4 5 6
9
Case Study

• After viewing, subjects answered questions about
• recall of the ad
• their attitude toward the camera
• their intention to purchase
• Three response variables

10
Comparative Experiments

• Comparison is the leading principle
• Effects of treatment can be judged only in
  relation to what happens in a similar group
• This sorts out changes that are unrelated to
  treatment
• You cannot assess the effect of a treatment in
  isolation because
• Many factors contribute to a response
• Conditions change on their own
• People are open to suggestion (Placebo effect)
• Observations changes things (Hawthorne effect)

11
Randomized Experiments

• Randomization is the second leading principle
• Randomization use impersonal chance mechanisms
  to assign treatments
• Randomization balances lurking variables across
  treatments groups

12
Example Quitting Smoking with Nicotine Patches
(JAMA, Feb. 23, 1994, pp. 595-600)

• Subjects 60 people
• Explanatory variable Treatment assignment
• Nicotine patch
• Control (placebo) patch
• Random assignment of treatment!
• Response variable Cessation of smoking (yes/no)

13
Outline of Experiment
Group 130 smokers
Treatment 1 Nicotine Patch
Random Assignment
CompareCessation rates
Treatment 2 Placebo Patch
Group 230 smokers
14
Randomizing the Treatment

• Assign labels 01,,60 to subjects
• Use random digits (TABLE B)
• Select line at random (say 102)73676 47150
  99400 01927
• First four subjects are 50, 40, 19, and 27
• Keep using table until you get 30 subjects in
  Group 1
• Remaining subjects are assigned to Group 2

15
Mozart, Relaxation and Performance on Spatial
Tasks (Nature, Oct. 14, 1993, p. 611)

• Subjects (30 undergraduate students)
• Variables
• Explanatory 3 treatments (see below)
• Response Change in Stanford-Binet IQ score
• Treatment
• Group 1 Listen to Mozart
• Group 2 Listen to relaxation tapes
• Group 3 Silence
• Random assignment of treatments

16
Outline of Experiment
Group 110 students
Treatment 1 Mozart
Random Assignment
CompareChange in IQ score
Group 210 students
Treatment 2Relaxation
Group 310 students
Treatment 3Silence
17
Logic of Randomized Comparative Experiments

• Random assignment ensure ? difference in response
  due to either
• Treatment
• Chance in the assignment of treatments
• If an experiment finds a difference among groups,
  we ask whether this difference is (real) or due
  to the chance assignment
• If the observed difference is larger than what
  would be expected just by chance, then it is
  statistically significant

18
Illustrative example

• Consider an experiment of weight gain in
  laboratory rats
• There would be differences in weight gain even if
  both groups received the same diet
• Just by luck, some faster-growing rats would end
  in one group or the other
• If we assign many rats to each group, the effects
  of chance will balance out
• Key concept use enough controls to balance out
  chance differences