-BASIC PROCEDURE: HAVE A PERSON PAY ATTENTION TO ONE ASPECT OF INPUT WHILE IGNORING OTHER ASPECTS

BASIC QUESTION: WHAT IS THE FATE OF INFORMATION THAT WE DON'T PAY ATTENTION TO

THE CLASSIC WORK OF C. CHERRY

PROCEDURE: DICHOTIC LISTENING
    ALLOCATE ATTENTION VIA SHADOWING

TYPICAL FINDING:

1. THERE IS LIMITED AWARENESS OF INFORMATION THAT ONE DOES NOT PAY ATTENTION TO: LIMITED TO BASIC FEATURES (E.G., VOCAL QUALITIES, LOCATION)

2. MOST INFORMATION DOES NOT APPEAR TO BE PROCESSED DEEPLY. THAT IS, TO AN INTERPRETATION OR LEVEL OF MEANING EXTRACTION (AT LEAST SUCH KNOWLEDGE DOES NOT REACH CONSCIOUSNESS)

3. SOME UNATTENDED INFORMATION IS PROCESSED SUCH THAT THE MEANING REACHES CONSCIOUSNESS (E.G., ONE'S OWN NAME)

EARLY MODELS: "FILTERS" THAT BLOCK UNATTENDED INFORMATION AND FACILITATES THE PROCESSING OF THE ATTENDED TO INPUT

EXAMPLE OF FACILITATION

TASK: DECIDING WHETHER TWO TARGET LETTERS ARE THE SAME OR DIFFERENT

MANIPULATE: (1) WHETHER THE TWO LETTERS ARE THE SAME OR NOT
(2) EXPECTATION OF WHAT WILL BE COMING

EXPERIMENT
 

TRIAL N-1	CONTROL -WARNING	EXPERIMENTAL WARNING + INFORMATION POSITIVE            NEGATIVE TRIALS
	+	C                                C
TRIAL N	C C	C C                               C G

FIND:
COMPARE   CONTROL vs POSITIVE
 

• FASTER (A BENEFIT) WITH C C WHEN PRESENTED C ON TRIAL N-1 (RATHER THAN +)

• SLOWER (COST) WHEN PRESENTED A C ON TRIAL N-1 RATHER THAN +

AS AN EXAMPLE; CONSIDER HAVING 100 UNITS OF ATTENTION IF A GIVEN TASK REQUIRES 80 UNITS TO BE PERFORMED EFFICIENTLY, THAT MEANS WE ONLY HAVE 20 UNITS TO DO OTHER TASKS

SO HOW COULD THIS EXPLAIN THE COSTS AND BENEFITS WE JUST SAW?
LOGIC GOES AS FOLLOW==>IF WE EXPECT SOMETHING, WE ALLOCATE OUR RESOURCES TO PROCESSING THE EXPECTED STIMULUS (OR LOCATION ETC). IF INDEED THE EXPECTED DOES
OCCUR, THEN WE SHOULD GET A FACILITATION (BENEFIT IN PROCESSiNG, RELATIVE TO A CONTROL CONDITION IN WHICH ATTENTIONAL RESOURCES AREN'T ALLOCATED OR ALLOCATED TO THE SAME EXTENT)
HOWEVER, IF WE EXPECT SOMETHING (LIKE C C) BUT SOMETHING ELSE APPEARS WE, HAVING ALLOCATED OUR RESOURCES ELSEWHERE, DON'T POSSESS SUFFICIENT RESOURCES TO BE EFFECTIVE

EXPECTATION PRIMING: THE FINDING OF COSTS AND BENEFITS SUGGEST A LIMITED RESOURCE MODEL

NEELY'S STUDY

TASK: LEXICAL DECISION
MANIPULATIONS:
(1) SEMANTIC RELATEDNESS OF PRIME AND TARGET: EITHER UNRELATED: BODYPART (TRIAL N-1); ROBIN (TRIAL N) or RELATED (E.g., BIRD.... ROBIN)

(2) EXPECTATION: THE PROBABILITY AN ITEM WOULD BE FOLLOWED BY A GIVEN WORD TYPE
e.g.. IF ONE SEES THE PRIME BODYPART THERE WAS A .67 PROBABILITY THAT THE ITEM ON THE NEXT TRIAL WOULD BE A TYPE OF BIRD; AND ONLY A .17 PROBABILITY THAT IT WOULD ACTUALLY BE A BODYPART (SUCH AS THE LETTER STRING: LEG)

(3) THE DELAY BETWEEN PRIME AND TARGET

SHORT: 250 MSEC.
LONG:  700 MSEC

FINDINGS:

AT SHORT DELAYS
GET SEMANTIC PRIMING OF RELATED WORDS (BIRD->ROBIN) REGARDLESS OF EXPECTATIONS

-SUGGESTS A BOTTOM-UP, AUTOMATIC PROCESS

AT LONG DELAYS

GET EXPECTANCY BASED AND NOT SEMANTIC EFFECTS

SO ONE FINDS A BENEFIT IF THE LETTER STRING ON TRIAL N (EQ. ROBIN) IS EXPECTED BECAUSE IT HAD BEEN PRECEEDED BY BODYPART AS THE PRIME

BUT GET A COST IF IT IS THE UNEXPECTED LETTER STRING (EVEN IF SEMANTICALLY RELATED)
I.E.,  IF GET BODYPART (EXPECTING ON THE NEXT TRIAL A TYPE OF BIRD 2/3 TIMES) BUT LEG IS PRESENTED ON THE NEXT TRIAL

THUS: A FAST AUTOMATIC (UNCONSCIOUS) PRIMING AND A SLOWER DEVELOPING EXPECTANCY BASED FORM OF PRIMING

BUT CONSIDER THE FOLLOWING CASE

TASK: NAME THE GREEN LETTER, IGNORE THE RED LETTER