The Language of Studies

1
The Language of Studies

• Lecture 8
• Secs. 3.1 3.3
• Wed, Sep 14, 2005

2
Observation vs. Experimentation

• Observational study A study in which the
  subjects are not manipulated or controlled in any
  way, but simply observed.
• Designed experiment A study in which the
  researchers actively impose some treatment on the
  subjects and observe the response.

3
Think About It

• Think About It, p. 147.

4
Example

• Suppose researchers wish to determine whether
  there is a relationship between cocaine usage by
  pregnant women and birth defects.
• Should this be an observational study or a
  designed experiment?
• Why?

5
Example

• A traffic engineer is studying a particular
  intersection to determine the traffic flow.
• He needs to know the average number of cars that
  turn left, turn right, and go straight.
• Should this be an observational study of a
  designed experiment?
• Why?

6
Explanatory and Response Variables

• Response variable The variable that measures
  the outcome. It is believed to be affected by
  the explanatory variable.
• Explanatory variable The variable or factor
  that is thought to explain the observed outcome.
  It is believed to affect the response variable.

7
Lets Do It!

• Example 3.1, p. 149 Explanatory Variable and
  Response Variable.
• Lets do it! 3.1, p. 151 Explanatory Variable
  versus Response Variable.
• Lets do it! 3.2, p. 151 Food Myths.

8
Levels and Treatments

• The explanatory variable is typically denoted X.
• The response variable is typically denoted Y.
• Values of the explanatory variable are called
  levels.
• If there is more than one explanatory variable,
  then combinations of their values are called
  treatments.

9
Confound It!

• Confounding variable A variable whose effect
  cannot be separated from the effect of the
  explanatory variable.
• If a study has one or more confounding variables,
  then the researchers cannot attribute changes in
  the response variable to any one explanatory
  variable.
• Normally, researchers try to control for all
  confounding variables (that they can think of).

10
Example

• Example 3.4, p. 152 Treating Multiple Sclerosis
  with Protein.
• Lets Do It! 3.3, p. 154 Invasive versus
  noninvasive Cancer Treatments.

11
Do Explanatory Variables Explain?

• A study cannot prove that variations in the
  explanatory variable really were the cause of
  variations in the response variable.
• The study can only give evidence supporting that
  belief.
• It may be the case that there is a third variable
  that is affecting both the explanatory and
  response variables.
• It may be conceivable that the response
  variable affected the explanatory variable!