Three Kinds of Research

1
Three Kinds of Research

• Descriptive, observational, survey
• Correlational compare existing groups
• Experimental methods (experiment ? study)
• For any design, key question
• Does study put theory in question to risky test
  by adequately ruling out alternative explanations
  for result?

2
Observational Studies

• Common applications
• Some aspects of functional (operant) analysis of
  behavior
• Ethology
• Norming of tests
• Strengths can be cheap, quick can suggest
  unanticipated hypotheses
• Weaknesses very hard to rule out alternative
  explanations, especially to infer causation,
  measurement problems

3
Correlational Methods

• Can be used when there are ethical problems with
  doing controlled experiments
• Used when there are practical problems with doing
  experiments
• Inferring causality from correlation not
  impossible (geology, cosmology), but generally
  very difficult in psychology
• Third variable Z may explain association
  between X Y, without X causing Y

4
Correlational Methods, contd

• Uses
• Compare two naturally occurring groups on
  variable that may be (a) cause (of) group
  membership
• See how strongly you can predict X from Y
• Strengths widely applicable, cheap, usually
  ethical
• Weaknesses third variable issues, measurement
  problems

5
Experimental Methods

• Analogy poke nature in specified way, then
  watch selected aspect(s) of response
• Poking manipulated (experimenter-controlled)
  intervention/variable
• Uses
• animal models
• cell studies in neuroscience
• psychosocial studies (e.g., social psychological
  studies in diagnosis)

6
Experimental Methods, contd

• More uses
• treatment outcome studies
• experimentally induced (low-grade)
  psychopathology
• aptitude X treatment interactions
• experimental psychopathology studies
• Strengths strongly testing causation
• Weaknesses ethical, generalizability,
  measurement problems statistical power if N small