Experimental Psychology PSY 433

1
Experimental PsychologyPSY 433

• Chapter 9
• Conditioning and Learning (Cont.)

2
Maze Times (Both Labs)
3
Maze Errors (Both Labs)
4
ANOVA (Repeated Measures)
Tests of Within-Subjects Effects Measure
MEASURE_1 Source Type III Sum of
Squares df Mean Square F Sig. Time Sphericity
Assumed 21082.344 14 1505.882 9.128 .000 Greenho
use-Geisser 21082.344 1.236 17062.838 9.128 .004
Huynh-Feldt 21082.344 1.284 16420.288 9.128 .00
4 Lower-bound 21082.344 1.000 21082.344 9.128 .
008 Error(Time) Sphericity Assumed 39263.785 238
164.974 Greenhouse-Geisser 39263.785 21.005 186
9.285 Huynh-Feldt 39263.785 21.827 1798.892 L
ower-bound 39263.785 17.000 2309.634
5
More Talented Animals

• Sweet Sundance Gong Show act
• https//www.youtube.com/watch?vYO7lAHsn84Q
• Cirque de Sewer Gong Show act
• https//www.youtube.com/watch?vSGxuEKcE_94

6
Choosing an Experimental Design

• Between vs within subjects designs offer
  different tradeoffs, but there are more than
  practical considerations at stake.
• Choice of design may affect the actual outcome of
  the research.
• Sometimes using a between-subjects and a
  within-subjects design produces different
  results.
• Carryover effects may exist without the
  experimenters knowledge.

7
Order Effects

• Order effects (practice effects) experiencing
  one level affects behavior in another level
• Effects of practice, boredom, fatigue
• Example Does content (biology text vs. novel)
  affect proofreading speed? Order is Biology-Novel
• Order effects are controlled in within-subjects
  designs by randomizing or counterbalancing the
  presentation orders.

8
Order Effects in Proofreading
9
Differential Carryover Effects

• Carryover effects, differential/asymmetrical
  transfer effects occur when experiencing one
  level affects performance on the next.
• The effect of the first level on the second level
  differs depending on which comes first.
• Effect of B following A ? effect of A following B
• Confound occurs when one level consistently
  precedes the other.

10
Differential Carryover Effects in Problem Solving
11
Classical Conditioning Example

• Grice and Hunter (1964) - human conditioning
• UCS is air puff UR is blinking
• Vary CS intensity (loud or soft tone)
• Done between-subjects or within-subjects
• 500 ms CS 500 ms ISI US
• Varying CS intensity in BS design has no effect
  on CR
• Varying CS intensity in WS design has large
  effect.

12
(No Transcript)
13
Contrast Between Stimuli

• In a WS design, subjects can compare two levels
  of a stimulus in the same experiment and may
  respond to the two stimuli differently.
• This effect occurs despite the randomization and
  counterbalancing that were used to control for
  differential order effects
• First, choose a random order (LSSLSLLLS)
• Then counterbalance the random order ½ got it,
  ½ got the reverse (SLLSLSSSL)

14
(No Transcript)
15
Instrumental Conditioning Example

• Bower (1961) 3 groups of rats trained to run
  down an alley (maze) for food.
• Two IVs kind of reward color of maze.
• Reward
• Constant 8 got 8 pellets per maze run
• Constant 1 -- got 1 pellet per maze run
• Contrast got 1 pellet in one colored maze, 8 in
  the other colored maze (black/white)
• All 3 groups ½ got black ½ got white maze.

16
(No Transcript)
17
(No Transcript)
18
ABBA Counterbalancing

• Define 2 conditions A and B or S and L
• In Bowers case, 1 pellet (S) 8 pellet (L)
• Present in order ABBA (SLLS)
• If order effects are linear, they will then be
  distributed evenly across conditions.
• If nonlinear, do not use ABBA, or give practice
  trials first.
• Bower used both ABBA and BAAB.
• Use Balanced Latin Square with gt2 groups.

19
Small-n Designs

• Behavior often cannot be studied in large groups
  (large-n designs)
• Small-n frequently used in therapeutic situations
• Reversal designs ABA or ABAB
• A baseline recording of behaviors
• B introduction of treatment
• IV is essentially treatment / no treatment.

20
Example Crying Behavior
Reinforcement

Removing the positive reinforcement (attention)
extinguishes crying behavior.
21
FIG Kanto7e 9-8, Hart data
22
Multiple Baseline Designs

• Observe different behaviors, before and after
  learning.
• Can be done two ways
• Observe multiple behaviors in one individual --
  like a within-subject design
• Observe a single behavior in different
  individuals between-subjects
• Treatments are introduced at different times.

23
(No Transcript)
24
(No Transcript)
25
Example Siblings Compared

• 3 pairs of autistic vs normal siblings
• Baseline -- observed target behaviors (counting,
  letter ID, etc)
• Treatment -- trained normal sib to reinforce
  behaviors of autistic sib
• DV -- number of correct performances of behaviors.

26
(No Transcript)
27
Changing Criterion Design

• Instead of comparing different people or
  different behaviors, progress in shaping behavior
  over time is measured.
• The target behavior needed for reinforcement is
  changed as a behavior is acquired.
• Range-bound changing criterion instead of a
  target, a range for reinforcement is established.
• Distributed-criterion design targets are spread
  across several behaviors.