Experiments

• Experiment

  In the natural sciences, scientists set out to discover scientific laws of cause and effect by manipulating variables and observing the results
• Laboratory experiment

  • Researcher manipulates variables and measures the effects
  • Allows establishing cause-and-effect relationships
  • Highly reliable as experiments can be replicated
• Conducting a laboratory experiment

  1. Take a set of identical subjects

  2. Divide into experimental and control groups

  3. Vary the independent variable for experimental group

  4. Keep independent variable constant for control group

  5. Measure and compare the dependent variable
• Independent variable


  The causal factor that is manipulated by the researcher
• Dependent variable


  The effect that depends on the independent variable
• Sociologists have occasionally used the laboratory experiment to study human behaviour
• There are three types of experimental methods used in sociology: laboratory experiments, field experiments, and the comparative method
• Reliability of laboratory experiments

  • Precise specification of steps allows replication
  • Detached researcher with no influence on results
• Positivist sociologists favour the laboratory experiment in principle, but recognize its shortcomings
• Interpretivists reject the laboratory experiment as it fails to achieve validity by producing unnatural behaviour
• Hawthorne Effect


  When people behave differently because they know they are being studied, rather than in response to the variables being manipulated
• The Hawthorne Effect can lead to invalid data in laboratory experiments
• Free will


  Interpretivists argue that humans have consciousness and choice, so their behaviour cannot be explained by cause and effect
• Field experiments

  • Take place in natural surroundings rather than a lab
  • Subjects are generally unaware they are being studied, avoiding the Hawthorne Effect
  • Researcher manipulates variables to observe effects
• Field experiment example

  • Rosenhan's pseudopatient experiment where researchers presented themselves as mentally ill at hospitals
• Field experiments are more natural and valid than laboratory experiments, but have less control over variables
• Comparative method


  • A 'thought experiment' where the researcher compares two similar groups that differ on one variable
  • Allows studying past events and avoids ethical issues of experiments
• The comparative method has less control over variables than field experiments, so it is less certain in establishing cause-and-effect