Limited Magic of the Number Seven

7/28/2019 Limited Magic of the Number Seven

http://slidepdf.com/reader/full/limited-magic-of-the-number-seven 1/9



340 B R U C E M . R OSS

audio-speech system (Neisser, 1967; Sperl-

ing, 1967). The speech that hearing 5s per-f o r m in encoding items is usually carried o u ts u b v o c a l l y , b u t with congenitally deaf indi-v i d u a l s such a system obviously cannot f u n c -

t ion in the same w a y a s i t does fo r hearingi n d i v i d u a l s . T h e r e f o r e , whether or not

deaf 5s perform in the same manner as hear-ing 5s in retention o f item sequences ha s im -portant theoretical implications.

For material exhibited in a single presen-tation and discrimination learning requiringm e m o r y , there is evidence that deaf 5s re-

t a i n items as well or better than hearing 5s

(Blanton, 1968; Putnam, Iscoe, & Young,1962; Youniss, Feil, & Furth, 1965). But

in visual span tests, a type of task in whicha more crucial role c a n b e posited f o r audi-t o r y - v o c a l storage, the f inding of inferior

p e r f o r m a n c e on the part of deaf 5s has beenf r e q u e n t (Blair, 1957; Blanton. 1968).

Olsson and Furth (1966) found a mixedresult in that d e a f 5s were i n f e r i o r to hear-in g 5s with digit span performance but not

w i th a sequence composed o f nonsense f o r m s .For these authors this finding suggests that

s u p e r i o r i t y f o r digit span o n t h e part o fhearing 5s was due to the f a c t that they had

a great deal more everyday practice withnumber sequences than deaf 5s had. The

present study chose a procedure where, in

d i s t i n c t i o n f r o m nonsense forms, labeling w a seasy. However, the symbols used—plus,

m i n u s , circle, square—were unlikely to havebeen previously rehearsed as a sequence as

is the case for digit span.

The first experiment administered both

sequen t ia l ly patterned and unpatterned two-sy m b o l series to deaf a n d hearing childreno v e r an 8-yr. age range. It was hypothesizedthat deaf children would be inferior to hear-ing children with patterned series but not

w i t h th e more randomly ordered unpatternedseries . Deaf children receive considerabled r i l l o n rote memory in school b u t very littlepract ice in encoding novel material economi-c a l l y . In contrast, hearing children b y e x -e rcis ing their language skills have muchmore practice in retaining items in terms o forganized patterns. The second experimenta d m i n i s t e r e d patterned a n d unpatternedthree- a n d f o u r - s v m b o l series to deaf a n d

hearing children, a l l o f whom were at thechronological a ge (CA) o f 1 5 . Fulfillmento f th e prediction o f similar results for un-patterned series with deaf and hearingc h i l d r e n in both experiments would not in-

dicate that audio-speech coding cannot be af r e q u e n t a n d efficient method o f dealing with

v i s u a l information. It would, however, pro-v ide good evidence that an auditory-vocalsystem, in any literal sense, i s n o t crucial fo rretention o f sequentially presented items.

E X P E R I M E N T I

This experiment administered series com-posed o f only tw o d i f f e r e n t symbols to deafand hearing children at each 2-yr. age in-

terval from CA 7 to 15 . Two unpatterned

a nd three patterned series o f nine symbolsc o n s t i t u t e d the five two-symbol series per-f o r m e d by each 5. In the patterned series,

three permutations o f t h e same basic pat-t e r n were administered so that th e main dif-fer en ce among patterned series was in the

l o c a t i o n of the single r u n o f three symbolsoccurring in each series.

An important consideration in this and the

tw o following experiments is the relativelyl o n g exposure time fo r each symbol. Apartf r o m th e procedural desirability o f pacing

the task slowly for the younger children,there was also a theoretical reason for using

long symbol exposures. A clear distinctionis of ten drawn between th e "span o f a p -prehension" and the "span of immediatememory" (Hunter, 1929). The former in-

volves fast, even tachistoscopic item ex-

posures, while immediate memory exposures

are considerably longer, but just how muchlonger is not clear, particularly with youngc h i l d r e n . More importantly, th e span o f a p -prehension is said to have a limit o f four o rfive i te ms. A recent interpretation is thatth e reason only f o u r o r f i v e rather than sevenor eight items are recalled in the span of

apprehension is because of the high rate o fitem encoding required (Neisser, 1967). As

th e present study maintains that only fouror five items are ordinarily recalled inde-p e n d e n t l y o f exposure time, it seemed wiseto make exposure time so long that there

c o u l d b e n o possible overlap with th e spano f apprehension. A long exposure o f each



VISUAL M E M O R Y IN THE DEAF AND H E A R I N G 341

item allows 5s time for rehearsal, but if only

four or five items are retained, even with re-

hearsal, it can only strengthen the claim that

seven has no special magic.

Method

Subjec t s .— T h e 5 s were 1 0 0 hearing and 100deaf children divided into five groups o f 2 0 5 saccording to age. Each group w as approximatelyevenly divided as to sex. The hearing childrenwere sampled f r o m Grades 2, 4, 6, 8, and 10 of aparochial grade school and public high school.Selection w as m a de to eliminate atypical studentsand those no t close to the modal ages fo r thesegrades of CA 7, 9, 11, 13, and IS . Th e deaf chil-dren were drawn f r o m tw o state schools for thedeaf and were close to the same ages as the hearing

5s. Children were selected w ho were either con-genitally deaf o r wh o became deaf at an age beforethey could have acqu ired language. Teacherscooperated to exclude partial ly hearing children.

P r o c e d u r e .— A l l 5 s were ru n individual ly andgiven the same treatment. The five two -symbolseries presented to a l l 5s were preceded by threedifferent practice series, four, five, and five symbolslong. If 5 m a de a n error in responding to a prac-tice series, i t was repeated until he got i t correct.Th e symbols were projected one at a time a s al ight figure on a dark ground with th e viewingscreen 8-10 ft. in f r o n t of 5s . For 8 of the 10

groups, exposure time per symbol was 2.75 sec.with a 1-sec. interval between exposures. For the

C A 7 and 9 deaf groups, exposure time was 3.75sec. with a 1-sec. interval between exposures. U seo f a different time interval fo r these groups w asan unintended mistake. A blank expo sure pre-ceded and fo llo wed each series. After the blankexposure appeared at the end of each series, 5 w a sinstructed to draw in consecutive order the symbolshe had just seen, using a pencil on an unlinedsheet o f paper and putting symbols in a top tobot tom order. Some 5 s drew fewer and some morethan nine symbols as they were no t told th e series

length o f either practice o r test series. In a ddi -tion to the time taken in drawing the symbols, a10-sec. interval was given before the ready signalfo r the next series. Special care was taken w ithdeaf 5s to make sure they understood th e instruc-tions. Few of the deaf 5 s understood signs, but ablackboard and pencil and paper were used to giveinstructions and examples in addition to the prac-tice series.

S y m b o l s e r i e s .—Series were composed solely o fthe symbols + and — and w ere nine symbols lo ng.A length o f nine symbols w as used as a standardseries length because o f Miller's (19S6) statement

that people ca n repeat back nine binary digits andeight decimal digits. The first and fifth serieswere considered unpatterned, and the second, third,a nd fourth , pat terned. The first series was, f r o mbeginning to end, - i 1 - + h +. Thefifth series was the same as the first with symbols

t ransposed: 1 1 h H . T h e second,third, a nd fourt h series a ll contained a run o fthree symbols with th e other symbols a l t e r n at i n g :- + + + - + - + -, + - + —+ +,and 1 h + H 1 , respectively. Thesepatterned series are na m ed by the placement o f

their triple of identical symbols as the 2-3-4 series,th e 6-7-8 series, and the 4-5-6 series. Hal f of the5 s received th e above order and half th e sameseries in b a c k w a r d s o r d e r f r o m the fifth to thefirst series.

Results and Discussion

Scoring rationale.-—Correct responses were

scored in two ways, initial span and nine-

place span. The initial span was the num-

ber of symbols correct f r o m the beginning

o f the series to 5"s first error. Thus if 6"got the first symbol wrong, his score was

zero. In the nine-place span all nine sym-

bols were scored regardless of intervening

errors. For both the initial and nine-place

span, only the first nine symbols were scored,

even if 6" wrote down more than nine sym-

bols. Mean results for initial span are shown

in Table 1 and for nine-place span in Table

2. Means in both tables are based on data

from the same 5s. No d i f f e r e n c e was f o u n dbetween the two orders of administration so

results were combined. Also combined in

both tables, since they produced quite simi-

lar scores, were the results f r o m the two

T A B L E 1

MEAN N U M B E R O F S Y M B O L S I N I N I T I A L M E M O R YS P A N F O R T H E T W O - S Y M B O L S E R I F . S

I N EXP. I

G r ou p

CA 7 hearingCA 7 deaf

C A 9 hearingCA 9 deaf

CA 11 hearingC A 11 deaf

CA 13 hearingCA 13 deaf

CA 15 hearingC A I S deaf

Unpat-ternedseries

2 . 852 . 7 8

2 . 7 03 . 7 0

3 . 4 53 . 1 0

3 . 5 03.80

3.053.68

Ru n position in

patterned series

2-3-4

3.703.55

4 . 6 52 . 1 5

4 . 4 53 . 4 5

4 . 9 03 . 2 0

5.35*5 . 0 5

4-5-6

2 . 6 02 . 9 5

3.603.95

4 .5 02 . 6 5

4 . 9 56.00**

4.75*4 . 8 5

6-7-8

3 . 5 04 . 2 0 *

5.35**5 . 1 0

5 .30*4.85*

6.35**4 . 8 5

6.55**5 .75*

* p < .05.** p < .01,



342 B R U C E M . R O S S

T A B L E 2

M E A N N U M B E R O F S Y M B O L S I N N I N E - P L A C EM E M O R Y S P A N F O R T H E T W O - S Y M B O L

S E R I E S IN E X P . I

G r o u p

C A 7 h e a r i n gC A 7 d e a f

C A 9 h e a r i n gC A 9 d e af

C A 1 1 h e a r i n gCA 11 deaf

CA 1 3 h e a r i n g

CA 13 deaf

C A IS he ar i ngC A IS deaf

U n p a t -ternedseries

4 . S 7

488

4.505.48

4 .985.05

5 .05

5.505.185 . 6 5

R u n posi t ion in

patterned series

2-3-4

5 . 105 . 6 5

5 . 7 04 .90

5.80*4.95

6 . 2 0 *

5.156 . 5 5 *6.95**

4-5-6

S.105 . 2 5

5 . 0 55 .45

5 . S O

4.90

5 . 9 0

7.05*

6.15*6.25

6-7-8

5.005 . 6 5 *

6 . 1 5 * *6 . 6 5 *

6 . 5 5 *5 . 7 0

7 .00**

6.75*7.30**7.10**

* p < .05.** p < . 01 .

unpatterned series for each 6". These seriesw e r e put first and last in each adminis t ra -t ion to c o n t r o l fo r possible learning o r f a -t igue effects. N o t e t h a t in Tables 1 a n d 2 ,

patterned series are presented in a left- to-right order which is not the same as the

a c t ua l o r de r o f presenta t ion .

With a chance level o f 5 0 % giving am e a n score o f 4.50 s y m b o l s if 5" r e p r o d u c e dnine symbols, i t is the a u t h o r ' s o p i n i o n t h a tgreater dependency should be placed o n t h einitial span as a measure of m e m o r y p r o -ficiency fo r two -symbol series . Ho w ever ,th e nine-place span does o f f e r a possibi li tyfo r f ur t h e r c o n f i r m a t i o n o f r e s u l t s o b t a i n e dwith th e initial span. This s up p l e m e n t a r ycheck is desirable since results do not lendthemselves well to an overall parametric

ana lysis wi th scores tha t varied o n l y f r o mzero to nine a n d a n u n e v e n n u m b e r o f u n -pat terned a nd patterned series. In b o t htables it is nevertheless possible to carry o u ta la rge number o f s ta t is t ica l tes ts unfavor-able to the hypotheses that deaf and hearingi"s a re f u n d a m e n t a l l y a l i ke a nd t h a t m e m o r y

span for pa t terned series wi l l exceed tha tfo r unp at terned series. Suppo rt f o r th ec la im that memory span for unpat ternedseries is o n l y f o u r or f ive s y m b o l s is a di r e c tfunct ion of the obtained mean scores.

U n p a t t e r n e d s e r i e s .— T h e range o f m e a nscores fo r unpatterned series in Table 1 isf r o m 2.70 to 3.50 symbols fo r h e a r i n g 5"sand f rom 2.78 to 3.68 symbols for deaf 5"s.T he ranges in Table 2 wh i c h , o f course, per-

mit greater leeway for guessing are f r o m4.50 to 5 .18 for hearing 5"s and f r o m 4.88to 5 .65 for deaf .9s. When the extremes ofthese ranges were compared in significancetests, n o n e approached significance. Deafa n d h e a r i n g 5"s were so c lose together inpe r f o r m a nc e t h a t no tests were made exceptat CA 9 wh e r e de a f 6"s p e r f o r m e d signifi-cantly better than hearing 5"s in initial spanby th e Man n - W h itn e y U test (p < . 0 5 ) ,

but as deaf v S s were a l lowed a longer sym-bo l e x p o s ur e t i m e a t this age, this resultm ust be disco unted. Significance fo r thisresul t was not conf irmed with the nine-placespan.

The hypothesis that naive 5s o r d i n a r i l yretain o n l y f o ur or five items at a singleviewing w a s well supported fo r un p a t t e r n e dseries using initial span, a s a l l m e a n s we r eless than f o u r i tems. With th e nine-placespan, several means slight ly higher than five

i te m s w e r e o b t a in e d . H o w e v e r , th e ques-tion arises whether the increase in n um b e ro f i t e m s s h o wn for the nine-place span is ac o n t r i b ut i o n o f s uc h f a c t o r s a s m e m o r y r e -cency or only the natura l increase tha t w o u l dbe expected f r o m count ing guesses as correc t .Accordingly , predic ted nine-place meansbased on guessing expectancies were calcu-lated fo r unpat terned series. These entriesw ere ca lcula ted by subtrac t ing both th e m e a ninitial span and mean number o f missingitems per series f r o m 9.00 a nd then mult i -plying this di f fe re nce by the chance level o f.5. This p r o duc t was then added to the ini-tial span to give th e predic ted nine-placemean fo r a given group. T h e predic ted u n-pat terned series means corresponding t o t heeight entries in this column in Table 2 f r o mC A 7 h e a r i n g t o C A 1 5 deaf w e r e , in o r d e r ,5.20, 5.46;5.15,6.12; 5.71, 5.61; 5.71,6.15;

5.08 , 6 .05 . Only the es t ima ted mean for

C A 1 5 hearing, 5.08, w as l o we r t h a n th ea c t u a l n i ne - p l a c e m e a n . There is, therefore ,no evidence t h a t 5"s we r e r e m e m b e r i n g a nya d d i t i o n a l i tems in the unpat terned series bynine-place scoring since Ss' performance was



V I S U A L M E M O R Y I N T H E D E A F A N D H E A R I N G 343

o n th e average a l i t t le be low what would beexpected on a chance basis. Results for

patterned series in nine-place spans were less

predictably related to initial spans so t h a tinclusion of nine-place spans as a separatetype o f me a s u re me nt is just i f ied .

P a t t e r n e d s e r i e s .— Me a n c o mbi ne d s c o re sfo r th e three patterned series were comparedbetween hearing and deaf 5s at all ages by

Mann-Whitney U tests. At no age w e rehearing or deaf 5s significantly di fferen tf r o m each other either in the initial or thenine-place span. Only for the 2-3-4 seriesat CA 9 was the hearing 5s' pe rfo rma nc esignificantly better by initial span than that

o f deaf 5s ( / > < .05), bu t this significanced id no t h o l d for the nine-place span.

It is apparent in Tables 1 and 2 t ha t the

patterned series are not o f e q u a l difficulty.Each patterned series in both tables w a stested against the mean of the two unpat-terned series by use of the Wilcoxonmatched-pairs s igned-ranks test. Significantresults f rom these tests are indicated byasterisks in the tables . Using both spanmeasures, these tests indicate that th e serieswith runs in the 6-7-8 posi t ion are most fre-quent ly significa ntly better retained . At a l lbu t C A 7 t h e 6-7-8 series a re significant fo rhearing 5s by both span measures . In a d d i -tion, the two other patterned series are sig-nificant by bo t h me a s u re s a t C A IS , a nd t he2-3-4 series is significant by nine-place spanat CA 11 and 13 . The deaf gro u ps s ho wa similar t r e n d ; for the 6-7-8 series, CA 7,11, and 15 are significant by initial span,

and a l l but CA 11 by nine-place span. Also ,fo r deaf groups th e 4-5-6 series is significantby bo t h me a s u re s a t CA 1 3 , an d th e 2-3-4series at CA IS by nine-place span.

In the results for patterned series there iso n l y o ne example o f a 2-3-4 o r 4-5-6 seriesthat is significant without the 6-7-8 series ofthe same group a lso being significant ; thisis the 4-5-6 series fo r initial span with C A 1 3deaf 5s. A po s t h o c e xpla na t i o n a s t o w h y

th e 6-7-8 series is best retained can be madein t e rms of the o c c u r r e n c e o f a n initial stringo f a l ternat ing symbols in this series. Itw o u l d appear to be m o r e dif f icul t to detecta l ternat ing symbols that occur af ter a t r iple

— t h e 2-3-4 series—or when th e alternatingpattern is disrupted—the 4-5-6 series.

It was pointed out that there were no

a ge differences fo r unpatterned series byeither span measure . Ho w ever, there w ere

some age differences i n pe r fo rm a nc e o n pa t-terned series. The measure used to c o m -pare adjacent ages on both span measuresw a s M a n n - W h i t n e y U tests ba s e d o n me a ncombined errors for the three pat ternedseries. For hearing 5s the CA 9 5s w e r ef o u n d t o b e significantly better than C A 7 5 swith initial span (p < , 0 5 ) b ut n ot w i t hnine-place span. For deaf 5s the CA 13 5s

were significantly better than CA 11 5s w i t h

nine-place span bu t no t with initial span.Thus since even those significant differencest ha t w e re o bt a i ne d w e re no t c o n f i r m e d , noconsistent developmenta l t rend is apparent .

A l t h o u g h i t can be sat isfactori ly concludedthat th e results fo r u n p a t t e r n e d series a requi te accurate estimates, the values in Tables1 a n d 2 pro ba bly a re s l ight underest imateso f re tent ion capaci ty fo r p a t t e r n e d series.In o t he r w o rd s , t he re s ho u ld be mo re pa t -terned series with 7 to 8 retained items.

This re tent ion underest imation for pat ternedseries o ccu rred in two wa ys. Fo r an easyseries such as the 6-7-8 series there were to om a n y perfect series so t ha t ma ny 5 s d id n otreach their performance ceiling. The largest

n u m b e r o f perfect s e r i e s w a s 8 a t C A 15 .The second way that the entries in Tables 1

a nd 2 u nd e re s t i ma t e d 5s' p e r f o r m a n c e c a m ea b o u t because 5s sometimes remembered a

substant ia l segment of a patterned series butreversed th e s ymbo ls . In a few cases, sym-bo l reversal w a s s h o w n by the result that a l lnine symbols in a series were wrong. This

f i nd i ng pro vo ke d a n e rro r a na lys i s i n t e rmsof ru n length o f consecut ive errors. It wasf o u n d fo r bo t h he a r i ng a nd deaf 5s at everyage that on a per series basis , runs of threeo r m o re c o ns e c u t i ve e r ro rs ( bu t no t o fs ho r t e r lengt h) w e re mo re f r e q u e n t f o r pat-terned than unpat terned series even though,as Tables 1 and 2 show, to ta l errors were

greater fo r u np a t t e rne d t ha n p a t te rne d s e ri es .This result suggests th e possibili ty that withthe present highly repetitive series, patternf o r m is s o me t i me s re me mbe re d direct lyra t he r t ha n i nd i vi d u a l s ymbo ls t ha t c a n b e





V I S U A L M E M O R Y I N T H E D E A F A N D H E A R I N G 345

Thus the two span measures appear to be

somewhat less predictably related to each

other than was the case for two-symbol series

in Exp. I.

It can be observed in Tables 3 and 4 that

fo r hearing -$8 th e means fo r three-symbolunpatterned series overlap those fo r two-

sym bo l unpatterned series in Exp. I, but val-

u e s f o r four-symbol series a re consistently

smallest. Fo r deaf 6"s there is no overlap,

th e means fo r three-symbol unpatterned

series are smaller than the means for two-

sym bo l unpatterned series, and the means for

four-symbol unpatterned series are smaller

than the means for three-symbol unpatterned

series.

M e a n s fo r unpatterned series with dif fer -ent numbers o f symbols were tested against

each other by Mann-Whitney U tests, using

CA 15 results f r o m Exp. I for two-symbol

series. The two-symbol unpatterned mean

fo r hearing 5s was not significantly higher

than a ny means fo r three- o r four-symbol

series by initial span, but the two-symbol

mean fo r nine-place span w a s significantlyhigher than th e means f o r both four-symbol

series (p < .02). Also, the two three-sym-bo l means were significantly higher than o nef o u r - s y m b o l mean by both spans (p < .02).

Fo r deaf -9s, means fo r two-symbol unpat-

te r ne d series were significantly higher than

a l l three- a nd f o u r - s y m b o l unpatterned series

T A B L E 3

M E A N N U M B E R OF S Y M B O L S IN I N I T IA L M E M O R YS P A N FOR THE T H R E E - S Y M B O L AN D FOUR-

S Y M B O L SERIES IN EXP. II

TABLE 4

M E A N N U M B E R O F S Y M B O L S I N N I N E - P L A C EM E M O R Y S P A N FOR THE THREE- AN D

F O U R - S Y M B O L S E R I E S IN EXP. II

G r o u p

T h r e e - s y m b o l hearingT h r e e - s y m b o l deaf

F o u r - s y m b o l h e a r i n gF o u r - s y m b o l deaf

T h r e e - s y m b o l heari ng

T h r e e - s y m b o l deaf

Four-symbol hearingFour-symbol deaf

Unpat-ternedseries

3 . 3 22 . 0 5

2151 . 9 2

3.38

2 . 1 5

1.881.85

R u n position inpatterned series

3-4

3 . 3 52 . 5 0

2 . 3 52 , 1 5

2-3-4

5 . 6 0 *

4.55*

3.053.95*

5-6

3 . 2 51 . 9 5

2 . 2 52 . 2 5

4-5-6

3 . 1 0

2 . 9 5

2 . 7 53.10

7-8

3 . 2 51 2 5

2 . 3 02 . 1 0

6-7-8

4 . 0 0

2 . 8 5

1 . 7 01 . 3 5

G r o u p

T h r e e - s y m b o l hearingThree-symbol dea f

F o u r - s y m b o l hearingF o u r - s y m b o l deaf

T h r e e - s y m b o l hearingT h r e e - s y m b o l deaf

F o u r - s y m b o l hearingF o u r - s y m b o l deaf

Unpat-ternedseries

4 . 6 53.68

4 . 0 83.38

4 . 5 53 . 8 4

3 . 1 83.48

R u n position in

patterned series

3-4

4.40

4 . 3 5

3.653.80

2-3-4

6.45**5 . 5 0 *

4.95**5.35**

5-6

4 . 8 54 . 6 0

3 . 9 53 . 4 0

4-5-6

4 . 9 04 . 5 5

3 . 6 54 . 0 5

7-8

4 . 6 03 . 2 5

3 . 3 03 . 2 5

6-7-8

5.304.90

3.853 . 1 0

* p < .01,

* t < .05.

**p < .01.

by both spans (p < .02). There were no

significant di f feren c es between three- a ndf o u r - s y m b o l means fo r deaf - S " s . It can bec o n c l u d e d that there is a general trend f o ra d d i t i o n a l symbols in a series to increase

re tention difficu l ty , although results in termso f significance tests a re somewhat irregular.

When deaf vs. hearing comparisons were

p e r f o r m e d , one series, the three-symbol series

with no repetitions (top o f Tables 3 a n d 4 ) ,s h o w e d significantly better performance o nth e part o f hearing 5 s f o r both spans byMann-Whitney U test (p < .05).

Patterned series.—With the same three-

s y m b o l series, hearing 5"s p e r f o r m e d signifi-cant ly better than deaf 5s on the patterned

series with a d o u b l e t o n in the 7-8 positions

b y both spans ( / > < . 0 5 ) . T h e c o m b i n e dpatterned series means f o r these groups were

significantly better fo r hearing 5 s by init ia lspan (p < .02) but not by nine-place span.

No other groups in Exp. II showed signifi-cant di f fe re nce s f a v o r i n g hearing 5s, al-

t h o u g h they p e r f o r m e d consistently better

than deaf 5"s w i t h th e other three-symbol

series.

As in Exp. I, Wilcoxon matched-pairssigned-ranks tests were performed to deter-

m i n e which patterned series were signifi-c a n t l y better than th e mean o f th e unpat-

terned series o f th e same group. Significant



346 B R U C E M. R O S S

results are denoted by asterisks in Tables 3

a n d 4 . Three o f th e four 2-3-4 series inTable 3 w e r e significant by initial span and

a l l four 2-3 -4 series in Table 4 by n ine - p l a c espan. No other ostensib ly "patterned" series

was significant in either table. A post hocexplanation o f these results is that in theabsence o f a discernible pattern for the ear lysymbols , such as the alternating symbols int w o - s y m b o l series, initial span c o u l d retain

o n l y th e abbb pattern present in the initialsymbols of the 2-3-4 p a t t e r n e d series.

Mann-Whitney U tests w e r e also carr iedo u t between th e m e a n s o f pa t te r ne d series o fd i f f e r e n t symbol composi t ion . There w a s asigni f icant d i f fe re nce favor ing th e combine dm e a n o f t h e three t w o - s y m b o l patterned

series a t C A 1 5 in Exp. I o v e r th e combinedm e a n s o f b o t h pat te rne d th r e e - sym bo l series

in Exp. II. This finding h o l d s for both

h e a r ing a nd d e a f 5"s with both spans. Pat-terns w e r e s o much simple r in t w o - s y m b o lseries t h a t this r e s u l t c a n b e d i s c o u n t e d . F o rth e m e a n s o f pat te rne d series with triples inTables 3 and 4 the th r e e - sym bo l series w e r esignificant ly better retained t h a n th e f o u r -

symbol series by h e a r i n g -Ss in both spans(p < .02), but th is result d i d n o t hold fo rdeaf 5s. T h e m e a n s o f series with double-

t o n s s h o w e d no significant differences be-twe e n three- a n d f o u r - s y m b o l series fo r eitherdeaf o r hearing 5"s. The unique n a t u r e o feach pa t te r ne d series implies that no c o m -pa r iso ns pu r e l y in terms of ser ies composi-t ion a re possible , bu t p r e s e n t results suggestthat, a s with u n p a t t e r n e d series, th e diffi-

c u l t y o f pa t te r ne d series increases as the

n u m b e r o f d i f f e r e n t s y m b o l s in a series

increases.

C O N C L U S I O N S

The two major hypothe se s of the study we rewell s u p p o r t e d by re su l t s which showe d that(a) the ave rage span fo r i tems in unpat te rne dseries w a s a b o u t f o u r a n d ( 6 ) results fo rh e a r i n g 5 s w e r e d i s t i n g u i s h a b l e f r o m those fo rdeaf 5s in both unpat te rne d and pat te rne d seriesfo r only o ne type o f three-symbol series.

O the r findings were that series regulari t iessuch a s symbol runs a nd symbol alternationssignificantly increased th e n u m b e r o f symbols inm e m o r y span beyond that f o u n d with unpat-terned series. How e ve r, re gu la r i t ie s we re pre s-

ent in some series that were ineffective in in-creasing- memory span.

In o n e respect , results were more favorablethan had been expected in s u p p o r t of the origi-nal hypo thesis of a l imited span. I t had been

t h o u g h t that in cont rast to multisymbol series,two-sy mb ol series might produce spans con-sistently longer than f o u r or five symbols.This d i d n o t p r o v e to be the case when 5s'results were evaluated against guessing proba-bilities.

It is conceivable that 5s older than C A 1 5w o u l d have a longer span for unpatterned series,bu t the norms given for the Wechsler AdultIntelligence Scale digit span suggest that a nyincrement wo uld be smal l . F or two -symbolunpatterned series in the present experiment no

significant increase in span w as shown betweenC A 7 and 15.

As stated in the in tr oduc t ion , other studieshav e s h o w n a t le ast e qua l pe r formance inr eten t ion on the part of deaf 5s w h o w e r e c o m -pared to matched hearing 5s. Partial originali tycan be c la ime d for the present study in showingt h a t deaf 5 s were general ly no t infe r ior in se-quential retention when fam i l ia r i ty a nd pre viousp r ac t ice with st imulus i tems were adequatelycont ro l le d .

The assu mptio n that the aud ito ry-voca l sys-tem is necessarily involved in rememberingvisual i tems has been challenged in experimentswith hearing 5 s . F o r example, Posner (1967)h a s shown that informat ion f r o m visua l d is-plays need no t invar iab ly be encoded verbally tobe well remembered. This do es no t me an, o fcour se , that hearing 5 s d o n o t f re que nt ly in -v o lv e th e auditory-vocal system in retainingvisual i tems. Co nversely, cases can also oc curwhere sequential auditory st imuli seem to b eencoded directly as visua l imagery (Wo o d-

w o r t h & Schlosberg, 1954, p. 2 2 1 ) . The trans-l a t ion o f visual sequences into auditory-vocals torage i s pre sumably a more common st ra te gy,bu t it is not a necessity.

R E F E R E N C E S

B L A I R , F. X . A study of the visual memory o fdeaf and hearing children. A m e r i c a n A n n a l s o fth e D e a f , 1957, 102, 254-263.

B L A N T O N , R. L. Language learning and perform-ance in the deaf. In S. Rosenberg & J. H. Kop-lin ( E d s . ) , D e v e l o p m e n t s in a p p l i e d p s ych o l i n -guis t ic r e s e a r c h . N ew Y o r k : Macmillan, 1968.

H U N T E R , W. S. Learning: II. Experimentalstudies o f learning. In C. M urchison (Ed . ) ,F o u n d a t i o n s o f e x p e r ime n ta l p s ych o log y . W o r -cester, Mass, : Clark University Press, 1929,



Limited Magic of the Number Seven

Documents