Experimental Methods 626

1
Experimental Methods 626

• Science, Theory, Research Application

2
Objectives

• Recall goals of science
• Describe role of theory
• Recall types of theories
• Recall criteria for evaluating theory

3
Some questions

• What is science?
• What is purpose of science?
• Is there an omni-everything God?
• What is the relationship between religion and
  science?
• Is there free will?
• Is there a reality?
• Is science the description of that reality?
• What is the essential tension?

4
Goals of Science
5
Role of theory (nothing so practical as a good
theory)

• Identify predictors
• Make predictions
• Identify interventions
• Identify transference of interventions (e.g.,
  antibodies)
• Heuristic value serving to discover or to
  stimulate investigation

6
The problem-solving paradigm

• identify problem (discrepancy from some goal)
• hypothesize source of problem
• generate alternative solutions
• evaluate alternatives (compare anticipated
  results if solution adopted)
• pick alternative
• implement alternative
• Compare results with goal, revisit if discrepancy
  persists or arises again

7
Role of research

• Used to test theories
• Used to confirm effectiveness of
  interventions/predictions

8
Types of theories

• Extrapolation (folk remedies)
• intelligence as predictor of school performance.
• e.g., sleep deprivation as predictor of mono
• Abstraction (constructs and relationships)
• one level of explanation down from extrapolation
• e.g., intelligence as construct, self-esteem
  Extrapolations
• e.g., immune compromising interventions effect on
  mono
• Models
• Representation of how physical or function
  components fix together
• e.g., intelligence as organization of brain . . .
• e.g., germ theory of mono
• Typology or taxonomy (e.g., periodic table of
  elements)

9
Criteria for evaluating theory (all relative to
competition)
10
Criteria (contd) and Epistemological paradigms
11
All knowledge is tentative

• Alternative explanations !!!
• General
• Chance
• Reverse causality
• Demand characteristics
• Etc.
• Specific
• Two or more theories can explain data

12
Validity of interpretations of results (not of
study)

• Construct validity
• Measuring/manipulating less than you think
• Measuring/manipulating more than you think
• I.e., confounding
• Can it be measured/manipulated?
• Statistical conclusion validity
• Chance or misapplication of analysis are sources
  of alternative explanation (generally subject of
  statistics classes)

13
Internal validity

• Extent causal inferences can be drawn
• _____________
• _________________
• No alternative explanations
• Reverse causality (problem with passive
  observation designs if weak theory)
• 3rd variable (confounds w/ manipulations,
  unmeasured variables)
• Alternative conceptualizations of cause
• e.g., horse and fence
• e.g., open v. closed

14
External validity

• Degree findings can generalize across
• Samples/Populations
• Contexts
• Behaviors
• Time
• Operationalizations of variables
• (overlaps construct validity)

15
Role of application

• Its the point, stupid
• Prediction
• At this point cause does not matter
• Program evaluation research
• Cause is everything
• Should we put our money in this again? On a
  grander scale?
• Does less work?
• Generally use kitchen sink method, therefore
  confounding is a problem

16
Next class

• What are the techniques of the social sciences?