16. Experimental Research Designs

1
16. Experimental Research Designs

• Pelham Blanton
• Ch. 6

2
Random Assignment

• Every participant has an equal chance of being
  assigned to any particular condition of an
  experiment
• Compare with matching
• Participants in each conditions are matched
  according to a variable(s).
• Problem?

3
Random Assignment

• Equalizes groups
• Individuals within the groups may be different
  but the group averages on all variables (except
  the one of interest) should be the same
• Interested in group averages not individual
  performance
• Because it is random there will occasionally be
  instances in which the groups are not equal but
  it is the best we have (particularly if we
  average across many experiments converging
  evidence)
• Statistics rule out purely chance factors

4
Strengths of True Experiments

• Eliminate individual differences
• Through random assignment
• Differences average out across groups

5

• Eliminate confounds
• Person confounds experiments rule these out
• Procedural confounds unknowingly manipulate
  more than one thing
• Operational confounds influences construct
  validity (not internal)
• Three ways to avoid confounds
• Randomly assign participants
• Eliminate procedural and operational confounds
• Move studies to the lab

6

• Allow researchers to observe the unobservable
• fMRI
• implicit memory

7

• Provide information about interactions
• The effect of one independent variable on the
  dependent variable is different at different
  levels of the other independent variable
• Allows you to qualify your theories (boundary
  conditions)

8

• Minimize noise
• Extraneous variables that influence the dependent
  variable, but is evenly distributed across all
  conditions (compare with confounds)
• Noise increases variability so that it is
  difficult to see what is really going on
• Trade-off between reducing noise and maintaining
  external validity

9
Artificiality

• Mundane realism the degree to which the physical
  setting in an experiment is similar to the real
  world you want to extend to
• e.g., Marlatts bar at UW
• Focus on surface features
• Make lab as similar to real world as possible

10

• Experimental realism the degree to which a study
  is psychologically meaningful to participants
• e.g., Asch conformity, computer games
• Do subjects experience the intended psychological
  effect?
• Do subjects behave naturally spontaneously?
• usually require some deception (or at a minimum
  omission of truth)
• Can occur in the absence of mundane realism

11

• Mundane looks like real world
• Experimental feels like real world

12
Example Can jurors be influenced by the
attractiveness of the defendant?
13
(No Transcript)
14
Demo

• Were going to do an experiment in class
• Taste test can people taste fat in food?
• They should be able toentire theories of feeding
  behaviour have centred around maintaining lipid
  levels
• Yet, modern science does not include fat on the
  list of things humans can taste . . .

15
The Human Tongue

• Apparently, we can taste
• Sweet
• Salty
• Sour
• Bitter
• Not fatty

16

• If you choose to participate, you will be asked
  to try two identical products, that vary only in
  their amount of fat
• Participants with be randomly assigned to two
  groups (1) product A then B, (2) product B then
  A
• After eating both, decide which is high and which
  is low in fat (i.e., can you taste the
  difference?)