experiments

Class Notes: Experiments

Experiments are one of many methods available to the social scientist. Unlike all other methods, however, experiments allow the researcher to specify cause and effect. What are the basics of experiments?

What defines an experiment?
- Experimental treatment
  - procedure or substance to investigate
- Use of at least two groups
  - control group
    - does not get treatment
  - experimental group
    - gets treatment
- An experiment may have many groups
Toothpaste example
- Does toothpaste and brushing lead to fewer cavities?
- How can we test?
  - Set up two groups, and randomly assign to:
    - control group
      - brush with water three times/day for six months
    - experimental group
      - brush with toothpaste three times/day for six months
  - Measure effect of toothpaste and water brushing with:
    - number of cavities after six months
  - Make both groups have zero cavities
- Time out for some terminology
  - Independent Variable (IV)
    - toothpaste or no toothpaste
  - Dependent Variable (DV)
    - number of cavities
  - Extraneous Variables
    - must be controlled
- Back to the experiment
  - After six months count cavities
  - If fewer in experimental group, you can say:
    - "Toothpaste has been proven to be an effective dentifrice when used as part of a regular brushing program."
- Recall hypotheses
  - Here hypothesis is:
    - Brushing with toothpaste will result in fewer cavities
  - More generally:
    - The IV led to some change in the DV

Bad experiments
- Nox-Out©
  - Imagine you are a researcher for a pharmaceutical company and you are asked to test an new drug, Nox-Out©, a new sleeping pill. For now, here are the variables:
    - IV: Nox-Out©
    - DV: some way(s) of measuring Nox-Out©'s effects
      - hours of sleep
      - quality of sleep
      - latency of sleep
    - Extraneous:
      - age of participants
      - fatigue level of participants
      - hours since last sleep
      - location of experiment
- Consider the following situations in the Nox-Out© experiment:
  - Age
    - Control group has average age of 25
    - Experimental group has average age of 75
      - Results are not valid graph
  - Fatigue
    - Control group members all work as ditch diggers
    - Experimental group members all work in offices
      - Results are not valid
  - Hours since sleep
    - Control group gets Nox-Out© at 11:00 a.m., or five hours since waking
    - Experimental group gets Nox-Out© at 11:00 p.m., or 16 hours since waking
      - Results are not valid
  - Location of experiment
    - Control group sleeps at the local Hilton
    - Experimental group sleeps at local dive motel
      - Results are not valid
      - How to fix? diagram

Back to Main Page