Overview of Methods

1
Overview of Methods
2
Objectives

• Understand two approaches to research
• Understand types and properties of research
  strategies

3
(False?) Dichotomies
4
The Focus Dichotomy

• Problem focused
• Find a problem (e.g., violence, poor decision
  making, marital conflict, stereotyping)
• Look at it (find sources of problem often
  qualitative approach)
• Look at theories that might be relevant
• Test adequacy of theories/actions taken
• Predictive power
• Mediating and moderating variables
• Effect of interventions

5
Theory Focused

• Have theory
• Look for tests (comparisons with chance, with
  other theories) to validate theoretical
  propositions (mediators)
• Look for better, cheaper measures of constructs
• Look for boundary conditions (moderators)
• Look for phenomenon explained by theory (increase
  comprehensiveness/generalizability of theory)
• At some point, this is the end game (i.e., X is a
  problem of type Y, therefore theory Z can be
  applied to make W prediction or V invention AND
  WE DO NOT NEED TO CONFIRM THAT W HAPPENED OR THAT
  V WAS EFFECTIVE)

6
(No Transcript)
7
Tradeoffs and Triangulation

• Think program of research
• No single study will answer a question
• but it should take a reasonable swipe at an
  aspect of the question
• No single design will address a problem
• The greater your repertoire, the more you can do
  to promote your brainchild
• Others are, or have, worked on related problems
• review the literature
• For any particular study, design depends on
  purpose!

8
Research Strategies

• Taxonomies of methods
• Experimental/correlational/case
• Runkel McGrath
• Properties
• Obtrusiveness
• Universal to particular
• Control (precision)
• Realism
• Validities (internal/external/construct/statistica
  l conclusion)
• ?

9
Quasi-experiment (equivalent groups)
10
Behaviors
Observations
11
Control Over Relevant Variables (Experimental V.
Subject)

• Three choices
• ______________
• Control group v. Experimental group (treatment
  group)
• Drug dosage (none, 5mg, 10mg, 15mg)
• Placebo training v. Training program you are
  testing
• Defining quality of experiment (mode X)

12
Control Over Variables (2rd)

• _____________
• Exclusion
• Equalization based on measurement
• But what is relevant?
• Can you really measure, exclude or equalize all?
• If equalizing, does it reflect natural
  distribution? Is measurement perfect?
• So how do we assure the variables are controlled?
  
• Randomization!

13
Control Over Variables (3rd)

• ____________________
• Defn randomly assigned to the various levels of
  the independent variable.
• For some, defining quality of experiment.
• Not the same as random selection (random
  sampling).
• Assures proportionally equalizing all variables
  except the manipulated variable.
• Sample must be large enough (consider sample of
  2) or enough experiments.
• Thus, we are not confounding manipulated variable
  with some other(s) relevant variable(s).

14
Problems with __________________________
15
Other Methods

• Case
• Idiographic issue (next slide)
• Evaluation
• Just field experiment
• Tend to throw in kitchen sink, so lots of
  confounds
• Action research
• Really describes science nicely (circular)

16
Idiothetic

• Combination of nomothetic and idiographic
• Nomothetic
• General principles that apply to most/all
• Justifies averaging?
• Idiographic
• Specific case (clinical)
• Particularistic (v. universal) research?
• Testing specimen v. casting nets
• Modal theories v. theories of individual
  differences

17
Next class

• Project
• Think about research area that interests you and
  come up with three (3) research questions related
  to it.