Name:
PSYCHOLOGY 435
Spring, 1999: Test #3

Part I. Please answer all three questions in this section. Questions 1 and 2 are worth 30 points each, and Question 3 is worth 20 points. (Distribute your time accordingly!) As always, you should CITE EXPERIMENTAL EVIDENCE (including names!), where possible, to support your answers.
 

(Ok, here is a sample answered test, though I've only outlined the answer for Question 1 to show you what material was relevant.  A lot of people mentioned the notion that Amsel claimed lack of a RF was frustrating, but provided no specifics on the learning mechanisms involved, and in some cases, failed to mention anything about partial reinforcement!  In answering questions like these, starting out with a pseudo-outline like the one I've provided is not a bad idea.  With that type of outline, the asnwers ought to write themselves.)
1. Discuss the extinction theories of Amsel, Capaldi, and Hull. For each, tell me about what mechanisms or processes are important in the theory, and at least two findings that fit the theory (be specific about experiments, where appropriate), and at least two findings/experiments that go against it. Briefly discuss, too, whether Amsel's or Capaldi's theory is more similar to Hull's, and why.
 

Main Ideas/Mechanisms                                    Some Evidence For                        Some Against

AMSEL
    Drive component of frustration                            Amsel & Roussel                                Hulse
    Competing R component                                     Daly                                                  Godbout et al.
    Discriminability                                                    Roberts                                             Capaldi & Waters
    rf....sf                                                                  Mowrer & Jones                                Hill & Spear
    partial vs. continuous RF
    classical conditioning

CAPALDI
    N-length                                                             Capaldi & Waters                               Hill & Spear
    MR,  MN                                                            Godbout et al.                                     Hulse
    short term memory                                              ITR: Capaldi, Hart, & Stanley             Pavlik
    discriminability                                                     ratio & interval length effects
    classical conditioning                                            variability effects
    3 principles
        variability
        length
        transitions

HULL
    IR,   SIR                                                               spontaneous recovery                          PREE
    Drive Reduction                                                   relearning                                            Seward & Levy
    exitatory & inhibitory associations                        spacing effect                                       Heyes et al.
    effective reaction potential                                    reminiscence                                        competing R studies, etc.
 
 
 
 

2. How are the following relevant to evaluating the Hull-Spence Algebraic Summation Theory (Continuity Theory) and the Sutherland & Mackintosh Attentional Hypothesis Testing Theory (Non-Continuity Theory)? (Note: You can briefly describe the findings of the relevant experiments, and then tell me why they are relevant to one or the other theory.)

(a) The work by Gibson and her colleagues on Differentiation

This involved a type of observational learning that did not require a reinforced R to a given S;  so against Hull.
(b) The work by Thomas and his colleagues on peak shift, and on what the animal pays attention to.
Thomas & Jones:  Peak Shift is due to adaptation level rather than inhibition;  you can get shifts to the same side as the S- (so against Hull)
Thomas, Freeman et al.:  Animals can learn about several dimensions at once (this could fit Hull, or you could relate this to what we learned about global focusing)
 (c) Two studies on errorless discrimination technique.
We discussed Terrace's study, which supports Hull by claiming to find no peak shift when this technique is used (presumably because it does not cause any inhibition to be associated with the S-).  We also talked about Rilling's evidence that there in fact was inhibition to S- (in which case EDT poses a problem for Hull, since the expected peak shift was missing).
  (d) Two studies on relational learning.
One of the three we talked about could be explained by peak shift, but the other two could not:  Kohler;  Lawrence & DeRivera; Gonzales et al.
 (e) The work on IDS & EDS
Mackintosh & Little's studies showing that a second problem will be learned more easily when it involves the same dimension that was important in an earlier problem.  That fits Sutherland and Mackintosh, particularly because the stimuli in the two problems are different enought to prevent any account of the results as being due to generalization.
(f) Levine's results with H-Theory in humans
You could have talked about Levine's evidence that people use hypotheses instead of trial-and-error responding (fits attentional theory), or you could have discussed the 'sudden' one-trial learning involved in the win-stay lose-shift strategy (incompatible with a continuity approach), or you might have chosen to discuss global focusing as a type of incremental learning.
 
 

3. Which model of human attention is best supported (or contradicted) by each of the following, and why? (Note: You can pick different models for each; I'm not asking you to choose just one model from all the different ones you studied.)

(a) dual-task studies such as Johnston & Wilson, and Tyler et al. (And you may bring in other dual-task studies, if you wish)

Kahneman's capacity model or the flexible filter model can handle these results.  So can many others.  But, Broadbent's blocking filter approach would definitely have problems with evidence for late selection, as in the J&W study.
  (b) studies on the cocktail party effect, and switching experiments like Treisman & Geffen
Again, Broadbent is out.  Others can handle these findings, but they were specifically discussed in concord with Treisman's attenuator model.
  (c) focused attention studies
I expected you to bring up several sample studies:  Moray, Cherry, Neisser, etc.  Depending on what findings you concentrated on, you could have tied these to a number of models.
 (d) studies using consistent mapping
This, of course, asked about studies & theories in automaticity, and that's what you ought briefly to have brought up.
   
 

Part II. Briefly identify five of the following (each is worth 4 points):

1. Fountain & Hulse

Can rats learn a 'rule' about how RF amounts change from trial to trial?
2. Kerr, Ostapoff, & Rubel
Raised birds in single-tone environments to test the Lashley-Wade claim that generalization gradients reflect learning (which is what they found)
  3. Mowrer & Jones
The discriminability theory of extinction claiming that resistance to extinction will depend on the similarity of learning & extinction contexts
  4. Easy-to-hard effect
Marsh:  An easy discrimination problem on a given dimension will aid the learning of a later hard problem on the same dimension
 5. Adelman & Maatsch
Study showed that response competition speeded extinction (the R competition group involved rats who had to back out of the goal box)
6. DRL schedule
Schedule in which animals are RFed for waiting between responses.
 
 

BONUS (for up to three points): Identify the one remaining term above (but let me know which term you want to count as your bonus!!!)