1. retroactive interference (RI)

2. proactive interference (PI)

- previous learning weakened by later learning

- eg. knowing French worsens after learning Spanish

- difficulty recalling old information due to newle learned information

- didn't conclude that sleep enhanced learning

- many current experiments on the positive effects of sleep on learning (memory consolidation and sleep) fail 

- new learning made more difficult by prior learning

- eg. knowing French hurts one's ability to learn Spanish

- difficulty in learning new information due to already existing information

Associationists' Methodology Retroactive Interference

- letters represent nonsense CVCs like DAX, ZIL)

Experimental Group

List 1 - A-B

List 2 - A-D

Test for 1st list - A-? Correct Response is B

Control Group

List 1 - DAX-ZIL

List 2 - DAX-WEP

Test for 1st list - DAX-? Correct Response is ZIL

-------

RI should cause > forgetting in the Experimental Group

RI = [%correct (control gr.) - %correct (experimental gr.)]/[% correct (control gr.]

ex: RI = [70% - 20%]/70% = .714

- letters represent nonsence CVCs like DAX, ZIL 

Experimental gr.

list 1 - A-B

list 2 - A-D

test for 2nd list - A-? Correct response is D

list 1 - DAX-ZIL

list 2 - DAX-WEP

test for 2nd list - DAX-? correct response is WEP

Control gr.

list 1 - A-B

list 2 - C-D

test for 2nd list - C-? correct response is D

list 1

PI should cause > forgetting in the experimental gr

RI = same formula

ex: PI = (80% -50%)/80% = .375

possibility #1

- response competition

people provide the response most strongly associated w/ the lefthand CVC

forgetting occurs bc the incorrect response is often more strongly associated, so it wins the competition

McGeoch 91932)

- interference effects

- forgetting due to entry of new info into memory

- operate on weak content only

- LTM uses semantic code

- similarity-based interference

- similarity of coding enhances interference effects

- response competition

______

- memories neither strengthen nor weaken over time

- memory strength established during conscious processing

RI

- list 1 - DAX-ZIL

- list 2 - DAX-WEP

test - DAX-? correct response is ZIL

forgetting of ZIL occurs bc the stronger association bet. DAX and WEP leads people to incorrectly respond 'WEP'

list 1 DAX-ZIL

list 2 DAX-WEP

test DAX-? correct response is ZIL

failure to correctly respond ZIL

must be due to incorrectly responding WEP

Melton and Irwin

RI

list 1 A-B

list 2 A-D

test - A? correct response is B

experimental gr. 

list 1 - A-B (5x)

list 2 - A-D (5,10,20, or 40x)

test - A-? correct response is B

RI = degree of forgetting of list 1 responses

prediction 1: RI should increase as List 2 Associations get stronger

prediction 2: forgetting due to RI should EQUAL the tendency to make list 2 responses

- higher retroactive interference is not matched by greater tendency to make the List 2 response

- more practice w/ list 2 actually reduces list 2 responses

the response competition hypothesis is WROng

The Unlearning hypothesis

melton and irwin (1940)

Retroactive I

list 1 - JEP-BIP

list 2 - JEP-RUL

test - JEP-? correct response is BIP

learning JEP-RUL might weaken the JEP-BIP association

A test of the unlearning hypothesis

barnes and underwood (1959)

list 1 - A-B (to perfection)

list 2 - A-D (1,5,10 or 20x)

test - A-? report both B and D responses

ex:

list 1 - KIP-TEK (until learned)

list 2 - KIP-POJ (1,5,10 or 20x)

test - KIP-? report both TEK and POJ

evidence for the unlearning hypothesis

conclusions

as list 2 responses increases, list 1 responses decreases

therefore, strengthening list 2 assoc. weakens list 1 assoc.?

evidence against the Unlearning hypothesis

postman and stark (1968)

list 1 - A-B (to perfection)

list 2 - A-D (1,5,10 or 20x)

test:

was A presented w/ B, E, or F?

and

was A presented w/ D, G, or H?

list 1 KIP-TEK (to perfection)

list 2 KIP-POJ (1, 5, 10 or 20x)

test

was KIP presented w/ TEK, HIF, or LOC?

and

was KIP presented w/ POJ, RIN, or WOX?

conclusions

as list 2 practice increases, recognition of list 1 responses is unaffected

therefore,

list 1 assoc. are not  weakened by strengthening list 2 assoc.

assoc. are not unlearned bc people recognize them if they are presented

maybe list 1 responses are suppressed?

the response set suppression hypothesis

postman, stark, and fraser (1968)

the list 1 assoc. are not weakened by learning list 2 assoc.

list 1 responses are suppressed by learning list 2 assoc (due to greater recency of learning)

Evidence for the response set suppression hypothesis

postman, stark and fraser 

list 1 A-B eg. JEP-WOP to perfection

list 2 A-D eg. JEP-SOV to perfection

gr. 1 test = list 1 recall

A? (JEP?) correct response is B (or WOP)

gr. 2 test = list 2 recall

A? (JEP?) correct response is D (or SOV)

evidence for the response set suppression hypothesis

postman, stark, fraser

both groups tested 

20 minutes and 48 hours after learning the lists

results:

list 1 responses were impaired when list 2 was learned relatively recently

list 1 responses were not impaired when list 2 was much less recent

conclusions:

recency of learning list 2 suppresses list 1 responses?

an alternative to the response set suppression hypothesis

a distinctiveness hypothesis

interference occurs bc due to their similarity, people cant separately access A-B and A-D pairings

prediction:

interference should be less if learning of A-b pairings is made distinct fr. learning of A-D pairings

Underwood & Freund

(distinct)

gr. 1

list 1 A-B eg. JEP-WOP - 32x and then on SAME DAY

list 2 A-D eg. JEP-SOV to perfection

gr. 2

list 1 A-B eg. JEP-WOP - 32x and then 3 days later

list 2 A-D eg. JEP-SOV to perfection

test for both gr. (1 hr delay after list 2)

list 2 response required

A? (JEP?) correct response is D (or SOV)

will separating the lists in time lead to less PI?

results:

interference is much reduced if time separates learning list 1 and list 2

conclusion:

interference occurs bc experiences are hard to distinguish fr. one another

gr 1

list 1 A-B eg. JEP-WOP to perfection

list 2 A-D eg. JEP-SOV to perfection

gr 2

list 1 A-B eg. JEP-WOP to perfection

list 2 A-D eg. JEP-SOV to perfection

and

while learning list 2 report list 1 responses

more evidence for the distinctiveness hypothesis

postman & gray

test for both grps: one week later

list 2 response required

A? (JEP?) correct response is D (or SOV)

what we know

- distinctiveness increases as the time bet. learning two similar assoc. is increased

- distinctiveness increases if people think about two similar assoc simultaneously

contradiction?

the underlying principle:

encoding variability

thinking about the same stimulus differently will create distinctive representations in memory

list 1 A-B eg. JEP-WOP to perfection

list 2 A-D eg. JEP-SOV to perfection

how to prevent interference bet. JEP-WOP and JEP

think of JEP as meaning something different when it is paired w/ WOP than with SOV

2 assoc:

BANK - VIOLIN

BANK - LAMP

how to avoid interference?

(RIVER) BANK - VIOLIN

(MONEY) BANK - LAMP