Journal of Educational Psychology1971, Vol. 83. No. 2,123-131

INDIVIDUAL DIFFERENCES IN VISUALAND AUDITORY MEMORY^

ARTHUR R. JENSEN1

University of California, Berkeley

Two groups were given visual and auditory forma of a digit memorytest in a counterbalanced order (auditory-visual group and visual-auditory group) under conditions of immediate and 10-second de-layed recall. Two control groups were given exclusively either thevisual or auditory test. Auditory memory was better than visual forimmediate recall; the reverse was true for delayed recall, the inter-action being significant beyond p < .001. The correlations betweenindividual differences in auditory and visual memory, after correctionfor attenuation based on reliabilities obtained from the control groups,did not significantly differ from unity for either immediate or delayedrecall. Thus there was no evidence for individual differences as a func-tion of sensory modality of the input. On the other hand, there was asignificant (p < .001) interaction between subjects and time of recall(immediate versus delayed).

Are there visual learners and auditorylearners? Do some individuals learn and re-member better in the one sensory modalitythan in the other? It is at least part of thegeneral folklore of psychology that suchauditory and visual types exist. Wechsler(1958), for example, in discussing the digitmemory span test, states: "It is certainlytrue that a good auditory memory does notgo with a good visual memory [p. 131]."

Interestingly enough, despite the fact thatthere is a venerable literature pertaining tothis question, an effort to find any real em-pirical support in psychological research forWechsler's statement has come to naught.Though the issue has been often discussedand even researched, it has apparently neverbeen subjected to a proper experiment—oneso designed as to be capable of answeringthe key question, at least for one type ofmemory. We know, of course, that no singleexperiment could answer the question for alltypes of learning and memory, since thereare large task-specific sources of variancewhich interact with individual differences.Therefore, the present paper is concerned

'The research reported in this paper was per-formed under a contract (Project No. 1867) withthe Cooperative Research Program of the Office ofEducation, United States Department of Health,Education, and Welfare.

* Requests for reprints should be sent to ArthurR. Jensen, School of Education, University of Cali-fornia, Berkeley, California 94720.

with this question only with respect to short-term memory, specifically memory for dig-its, a traditional part of standard psycho-metric tests such as the Stanford-Binet andthe Wechsler intelligence scales.

The problem really comes down to twoquestions: (a) Do people in general learnor remember more effectively when the ma-terial is presented visually or aurally? (b)Do some individuals favor one sensorymodality over another in learning and re-membering?

Most studies have tried to answer the firstquestion; few have tried to answer the sec-ond. The results are far from unanimous.Some investigators have found auditorymemory superior to visual (Binet, 1894;Koch, 1930; Miinsterberg & Bingham,1894). Others have found visual superior toauditory (Hawkins, 1897; Henmon, 1912;Kirkpatrick, 1894; O'Brien, 1921; Wor-chester, 1925). Simultaneous presentation ofthe material visually and aurally alwaysproduced the best recall (Binet, 1894; Koch,1930; Mimsterberg & Bingham, 1894). Dayand Beach (1950), who reviewed the entireliterature up to 1950, concluded that, interms of a box score, visual memory is gen-erally better than auditory. But a box scoreof experimental outcomes is, of course, sci-entifically a confession of ignorance. Itmeans the problem is complex and the con-ditions that produce one or another outcomemust be precisely specified. Interactions of

123

124 ARTHUB R. JENSEN

TABLE ISTUDIES OF THE CORRELATION BETWEEN VISUAL AND AUDITORY MEMORY

Investigator

Gates (1916o)Gates (1916b)Hao (1924)Wissler (1901)

Kitson (1917)Gates (1916a)Carey (1914)

Material

DigitsDigitsDigitsDigits (written)Digits (placed)VariousVariousVarious (pooled)

N

17216583

14414440

197150

Reliability

rAA

.52

.68

rw

.83

.64

CorrelationrAV

.57

.62

.39

.29

.39.09-. 54.07-. 41

.44

Note.—The author is indebted to John J. Geyer for compiling the information in this table.

the key variable (sensory modality) withother experimental variables could accountfor seemingly contradictory results in differ-ent studies, as could the interaction of thekey variable with subject variables or indi-vidual differences. For example, althoughthe majority of studies show that adults dobetter on visual than on auditory memorytasks, it appears that the reverse is true forchildren (Abbot, 1909; Hawkins, 1897;MacDougall, 1904). And visual short-termmemory has been found to decline more rap-idly with age (over 40) than auditory mem-ory (McGhie, Chapman, & Lawson, 1965).In brief, the question of the superiority ofvisual or auditory memory cannot be givena general answer that covers all conditionsand types of subjects. An experimentalanalysis of the specific conditions that makefor the superiority of one or the other isneeded.

What about the second question, that ofindividual differences? It should be answer-able for a specified type of memory task.Kay (1958) used a paired-associate testconsisting of familiar words and found that45% of the subjects performed about equallyin both modes of presentation, 12% did sig-nificantly better in the visual, and 4% didbetter in the auditory. If the same kind ofmaterial is presented to a group of subjectsin each of two modalities independently, allaspects of the presentation being held con-stant except sensory modality, then thecorrelation between subjects' performancein the two modalities should be significantlyless than 1.00 (after correction for attenua-tion) , if the hypothesis of individual differ-

enences in the relative efficacy of the visualand auditory modalities in short-term mem-ory is to be upheld. Several studies havebeen conducted in this vein. They are sum-marized hi Table 1. The raw correlations(TAV) between auditory and visual perform-ance, being generally quite low, would givethe impression that individual differences inauditory and visual memory are not highlycorrelated. But in most cases, we do notknow the reliabilities (FAA and rvv) of theauditory and visual memory tests, andtherefore cannot interpret the magnitude ofthe correlation rAV. In the two instanceswhere the reliabilities (rAA and rw) aregiven, we do not know if they were obtainedby a method that parallels the method forobtaining rAV. Split-half reliability, for ex-ample, is different from test-retest reliabil-ity in theory and usually in fact, and so itwould make little sense to correct rAy forattenuation by using either type of reliabil-ity coefficient arbitrarily. The lack of anyreal answer to our question revealed bythese data thus highlights the minimum re-quirements of any experiment addressed tothe problem.

Essentials of Experimental Design

A proper test of the hypothesis must bedesigned to answer the question of whetherthe correlation (corrected for attentuation)between auditory and visual memory issignificantly less than 1.00, that is,

fAV < 1.00

INDIVIDUAL DIFFERENCES IN VISUAL AND AUDITOBT MEMORY 125

This implies: (a) Reliability estimates(»AA and rvv) that are at least as dependableas the correlation between auditory and vis-ual performance (rAv) must be obtained.(6) Reliability coefficients must be in everyway parallel to the rAV correlation, that is,both the total amount of practice and theintervening time interval in the subjects'performance for obtaining the auditory-auditory and the visual-visual correlations(PAA and rw) must be identical to the pro-cedures for obtaining the auditory-visualcorrelation (rAV). This can be achieved onlyby using independent groups of subjects ran-domly sampled from the same pool for de-termining the reliabilities of the visual andauditory tests, and still another group fordetermining the correlation between the vis-ual and auditory tests, (c) The possibilityof practice in one modality transferring dif-ferentially to the other must be counterbal-anced by using two groups to obtain thecorrelation between modalities, one whichpractices auditory first and then visual, andone which does the reverse. The two correla-tions that result are designated rAv and rvA,respectively, (d) The testing procedureshould be identical in all respects except forthe sensory modality of the presentation.

The present experiment was designed tofulfill these requirements. In addition tosensory modality, one other experimentalvariable was introduced: immediate recallversus a 10-second delay in recall. Sincetime of recall cannot be precisely controlledin individual subjects, it is preferable to at-tempt to assess the effect of the variable bymanipulating it and determining if it inter-acts with the subjects and with sensorymodality.

METHOD

Subjects

The subjects were 150 undergraduate universitystudents, 19 to 23 years of age, in an introductoryeducational psychology course. They were ran-domly assigned to four groups.

Group AA (n = 50). Subjects were given equiva-lent forms of the auditory digit memory test oneach of 2 days.

Group VV (n = 60). Subjects were given equiv-alent forms of the visual digit memory test oneach of 2 days.

Group AV (n = 25). Subjects were given the

auditory digit memory test on the first day (Day1) and the visual test on the second day (Day 2).

Group VA (n = 25). Subjects were given thevisual test on Day 1 and the auditory test on Day2.

ProcedureApparatus. The subjects were tested alone in a

sound-protected room which was barren except forthe simple apparatus essential for the testing. Theexperimenter could observe the subject from anadjoining room through a one-way window. On thevisual test, digits (2 inches high) were presentedon an in-line display unit placed approximately 2feet from the seated subject, at eye level. Rate ofdigit presentation, stimulus selection, and all timeintervals were preprogrammed on a punched tapeand were automatically controlled by an apparatusdescribed in detail elsewhere (Jensen, Collins, &Vreeland, 1962). For the auditory test, the in-linedisplay unit was replaced by a speaker which pre-sented the digits (and other signals involved inthe procedure) in a normal, clear female voice.The tape recording was made to precisely the samepacing rate and time intervals as the visual presen-tation by having the speaker read aloud the entireautomatically presented visual test into a taperecorder. In front of the subject, who was seatedat a table, there was a response console tilted at a30° angle. Directly in the center of it was a clip-board holding the specially prepared forms onwhich the subject would write his response aftereach series of digits had been presented. Threeinches away from each side of the clipboard weretwo pushbuttons (94-inch in diameter) protrudingfrom the console; the right-hand button was la-beled with a red plus (+) sign and the left-handbutton was labeled with a red minus (—) sign. Toprevent the subjects' writing during the presenta-tion of the digit series and during the delay inter-val prior to recall, the subject was required tokeep his index fingers poised on the pushbuttons.(Failures to do so at the prescribed times wereautomatically recorded by the apparatus.)

MaterialsThe digit series, comprised of numerals 1

through 9 (zero was never used), were originallyproduced by a computer; the sequence was alwaysrandom, with two exceptions: (a) no digit wasever repeated in the series, and (b) no two digitsever occurred in the normal numerical order in theforward direction, such as 3-4 or 7-8.

There were two conditions of recall, immediateand delayed; in the delayed condition the subjectreceived the signal to write his response 10 sec-onds after the last digit of the series had beenpresented. The delay interval was filled by a ran-dom sequence of red plus and minus signs, at a1-second rate, on the screen (in the visual presen-tation) or spoken (in the audio presentation); thesubject was required to press the correspondingplus and minus pushbuttons as these symbols ap-peared on the screen or issued from tie speaker.The subject's errors and omissions of pushbutton

126 ARTHUR R. JENSEN

responses were automatically tabulated on coun-ters in the control unit, so it was possible to tell ifthe subjects were obeying instructions.)

Each digit series was preceded by the sound ofa "bong" as a signal to pay attention, and after thedigit presentation a "bong" signaled for the sub-ject to write his response. (The same "bong" wascommon to both the visual and auditory presenta-tions.)

InstructionsThe task was explained to the subject with the

aid of a wall chart which showed the two condi-tions (immediate and delayed) and the design ofthe whole procedure, as follows:

Immediate RecallEvents Time

bong (ready signal) 1 secondblank 1 seconddigits (2 to 9) 2-9 secondsbong (signal to write) 1 secondblank (for writing response) 13 secondsbong (ready signal) 1 secondetc.

Delayed RecallEvents Time

bong (ready signal) 1 secondblank 1 seconddigits (2 to 9) 2-9 secondsblank 1 second+ and — (random order) 8 secondsbong (signal to write) 1 secondblank (for writing response) 13 secondsbong (ready signal) 1 secondetc.

There were five replications of the test in eachday's session. Thus there were 8 (series lengths) X2 (immediate or delayed recall) X 5 (replications)= 80 digit series presented in each test.

To insure that all subjects understood the in-structions and procedure, the test proper was pre-ceded by a practice test consisting of eight series(four of each condition, immediate and delayed);the digit series was never larger than five in thepractice test.

The equivalent forms (1 and 2) of the test wereidentical in. all respects except that different digitseries were used in each. The visual and auditorytests also were identical (for both Forms 1 and 2)except in mode of stimulus presentation.

The subjects were tested on Forms 1 and 2 oneach of 2 days, Day 1 and Day 2, 1 week apart, atthe same hour.

MeasuresThe subjects were instructed to write down in

correct sequence as many of the digits as theycould recall. The answer sheet contained 80 seriesof 9 boxes for the subject to write his responses.The subject's score is the total number of digitsrecalled in the correct positions.

The possible total number of correct responsesfor 1 day's test session (of approximately 50 min-utes) was thus the sum of series lengths 2 + 3 +• • • + 9 = 44 X 2 (delayed or immediate) X 5(replications) = 440. The subjects' answer sheetswere directly punched on IBM cards and scoringwas done by computer.

RESULTS

Effects of Sensory Modality

Table 2 presents the means and standarddeviations of digit recall scores for thefour groups in the study. All mean differ-ences between the main effects of auditoryand visual, between immediate and delayedrecall, and between Day 1 and Day 2 (i.e.,the general practice effect) are revealed byanalysis of variance to be significant beyondthe .001 level. Day 2 performance is signifi-cantly superior to Day 1, indicating a prac-tice effect. Immediate recall is nearly alwayssuperior to delayed recall; only 2 of the 150subjects showed either no overall differenceor a difference in favor of delayed recall.Overall, visual was superior to auditory,but this fact must be viewed in relation to

TABLE 2SUMMARY STATISTICS ON THE FOUR GROUPS IN THE EXPERIMENT

Measure

Immediate (I)Delayed (D)Difference

(I- D)Relative

difference"

Visual (Group VV)

Dayl

M

172.22152.66

19.56

11.38

SD

22.2530.16

17.32

10.78

Day 2

M

182.94171.26

11.68

6.10

SD

20.6925.41

17.64

9.29

Auditory (Group AA)

Day 1

M

180.37136.00

44.37

24.20

SD

20.0625.68

19.44

10.62

Day 2

M

184.39U6.29

38.10

20.31

SD

20.0228.32

21.48

11.44

Visual-auditory (Group VA)

Day 1

If

182.84163.00

10.84

11.00

SD

21.6029.61

16.62

9.90

Day 2

M

181.88137.60

44.28

24.72

SD

22.3331.48

13.60

9.91

Auditory-visual (Group AV)

Day 1

M

169.28129.48

39.80

23.24

SD

23.4627.84

19.02

11.81

Day 2

M

177.96156.60

21.36

12.40

SD

26.3834.59

18.64

12.55

Note.—re = 60 for Group VV and for Group AA, n — 26 for Group VA and for Group AV."• The formula for the relative difference is ((I - D)/I] X 100.

INDIVIDUAL DIFFERENCES IN VISUAL AND AUDITORY MEMORY 127

the marked interaction between sensorymodality and time of recall.

The interaction between sensory modality(auditory versus visual) and time of recall(immediate versus 10-second delay) is sig-nificant beyond the .001 level. The form ofthis interaction is shown in Figure 1. Forimmediate recall, auditory was superior tovisual (p < .05) and for delayed recall,visual was markedly superior to auditory(p < .001).

Transfer or practice effects from visual toauditory and from auditory to visual prac-tice unfortunately cannot be properly as-sessed from these data. Differences in diffi-culty level of the auditory and visualconditions on Day 1 are confounded withDay 2 performance, such that comparisonsof Day 2 scores between Groups VA and AAand between Groups AV and W cannot berigorously interpreted with respect to therelative amounts of transfer from one sen-sory modality to the other. This transfer orpractice effect could be properly assessedonly if the auditory and visual tasks weresomehow experimentally equated in diffi-

80

a:<o% 70

Xo

<§60

£.

I I

ImmediateRecall

10" Delay

FIG. 1. Recall (as percentage of maximum possi-ble recall) of auditory (A) and visual (V) digitseries, under two conditions of recall: (a) immedi-ately following presentation of the digit series, and(b) following a 10-second delay after the presenta-tion of the series. (Series varied in length from twoto nine digits presented at a rate of 1 digit persecond.)

culty. (Statistical equating, as by analysisof covariance, would be an incorrect pro-cedure in this case.)

Serial Position Effects

Since serial position effects (i.e., proba-bility of correct recall of a digit as a func-tion of its position in the series) have somebearing on the theoretical interpretation(see Discussion section) of the interactionshown in Figure 1, the serial position dataare given in Tables 3 and 4 as probabilityof correct recall of each digit at each posi-tion in the series.

Individual Differences

The principal question this study was de-signed to answer is, Do some persons havebetter short-term memory in one modalitythan in another? If the true-score correla-tion between modalities is significantly lessthan unity, the answer to this question isyes. If the true-score correlation does notdiffer significantly from unity, it cannot beconcluded that individual differences in re-call interact with sensory modality of thepresentation.

The appropriate reliability coefficients forthe visual and auditory tasks are obtainedfrom the correlations (rAA and rvv) betweenDay 1 and Day 2 for Group AA and GroupW, for immediate recall and delayed recallseparately. These reliability coefficients arethen used to correct the correlations (rAvand TVA) between auditory and visual tasksfor attenuation to obtain the true-score cor-relation between modalities. The true-score(i.e., corrected) correlations (/) are rlv =rAv/VVAA»Vv and rvA = rvn/VrA*rvv. Theresults are shown in Table 5. Since none ofthe corrected correlations is significantlyless than unity, these data do not supportthe hypothesis that individual differencesin memory for digits are a function of thesensory mode (visual versus auditory) ofpresentation. This is true for both immedi-ate and delayed recall. Even though themean difficulty level of auditory and visualtasks is reversed in immediate and delayedrecall, the rank order of individuals' recallscores remains the same in both modalities.Another way of stating this, in terms of the

128 ARTHUR R. JENSEN

TABLE 3PROBABILITY OF CORRECT RECALL IN VISUAL DIGIT SERIES AS A FUNCTION OF LENGTH OF SERIES,

SERIAL POSITION OF DIGIT, AND TIME OF RECALL (IMMEDIATE VERSUS DELAYED)

Serieslength

2

3

4

5

6

7

8

9

Time ofrecall

ImmediateDelayed

ImmediateDelayed

ImmediateDelayed

ImmediateDelayed

ImmediateDelayed

ImmediateDelayed

ImmediateDelayed

ImmediateDelayed

Serial position

I

982976

983981

982954

976932

982971

942916

896872

914856

2

987962

981965

979937

956850

936899

907775

778764

728733

3

985951

977919

940830

907843

861691

724683

623597

4

975833

949794

890800

795644

662604

605523

5

961826

906772

779627

594536

552464

6

921794

749558

599499

505396

7

786519

550428

494359

8

557444

418311

9

458340

M

985968

983966

978911

956846

924847

831676

670604

589508

Note.—Decimals are omitted. Differences greater than .040 are significant beyond the .01 level.

analysis of variance, is that there is no sig-nificant Subjects X Modality interaction.

Interaction of Subjects X Recall Condition

Time of recall (immediate versus de-layed), on the other hand, interacts signifi-cantly with subjects, both for visual (GroupW: F = 4.09, df = 49/196, p < .001) andfor auditory (Group AA: F = 5.62, df =49/196, p < .001) presentation. Individualdifferences in immediate and delayed recallhave between 60% and 70% of their true-score variance in common. While virtuallyall subjects show poorer delayed than im-mediate recall, some subjects show a con-siderably greater memory decrement afterdelayed recall than do others. Thus, subjectsdo not maintain the same rank order ofability in immediate and delayed recall,while they do maintain the same rank orderin the aural and visual tests.

DISCUSSION

Mean Differences

The present findings are consistent withearlier studies that generally reported su-perior recall for auditory than for visualpresentation. All these studies used immedi-ate recall, which favors the auditory. Theimportant question, however, is why doessensory modality interact with time of re-call? Clues to the answer are provided byrecent theory and research on short-termmemory.

Sperling (1963) has hypothesized that aprimary function of verbal rehearsal invisual memory tasks is to convert informa-tion from visual to auditory storage. That isto say, the memory traces are transformedand encoded in an auditory form, regardlessof the sensory channel of reception. Re-hearsal, therefore, should be less necessary

INDIVIDUAL DIFFERENCES IN VISUAL AND AUDITORY MEMORY 129

TABLE 4PROBABILITY OF CORRECT RECALL IN AUDITORY DIGIT SERIES AS A FUNCTION OF LENGTH OF

SERIES, SERIAL POSITION OF DIGIT, AND TIME OF RECALL (IMMEDIATE VERSUS DELAYED)

Serieslength

2

3

4

5

6

7

8

9

Timeolrecall

ImmediateDelayed

ImmediateDelayed

ImmediateDelayed

ImmediateDelayed

ImmediateDelayed

ImmediateDelayed

ImmediateDelayed

ImmediateDelayed

Serial position

1

997984

998981

996976

997917

986940

980893

921817

900830

2

997973

997954

982926

978796

932805

922700

794631

733710

3

998957

918878

962744

876698

870594

708617

588509

4

929818

969672

844628

805466

639459

589409

5

983689

896568

761428

592376

469361

6

950636

750340

588320

465284

7

858373

615255

473247

8

700272

508227

9

670257

AT

997979

998964

956900

978764

914712

849542

695456

599426

Note.—Decimals are omitted. Differences larger than .040 are significant beyond the .01 level.

for auditory than for visual presentation,since no effort is needed on the subject'spart to transform the auditory input to getit into auditory storage. (Yet, strangelyenough, unilateral temporal lobe damage orlobectomy seriously impairs auditory butnot visual memory—Meyer, 1961, p. 545.)Rehearsal serves not only to convert the vis-ual to auditory storage, but also tostrengthen the memory trace, whether pres-entation is visual or auditory. If subjectstransform the visual input into an auditoryform, then it should be less susceptible tovisual retroactive interference (the seriesof red pluses and minuses projected on thescreen) than to auditory forms of retroac-tive interference (a spoken series of "plus"and "minus"). The subjects often reportassuming a more passive attitude in theauditory than in the visual presentation;they experience an "echo chamber" effectfor immediate recall of auditory digits, andsimply read back the "echo" in writing down

their answer. But the echo fades rapidly andalso is highly susceptible to interference byany other intervening auditory input, andthus there is a greater decrement in the de-layed recall of auditory as compared withvisual digits.

This interpretation is consistent withGates' (1916b) finding that subjects testedwith lists that exceeded their immediatememory span remembered 25.9% less than

TABLE 5RAW CORRELATIONS BETWEEN DAY 1 AND DAY 2

FOR THE FOUR GROUPS AND CORRELATIONS COH-BECTED FOB ATTENUATION FOB THE CHANGED-

MODALITY GROUPS (AUDITORY-VISUAL ANDVISUAL-AUDITOBY)

Recall

ImmediateDelayed

Raw correlations

'AA

.82

.80

TV

.70

.79

'AV

.86

.79

TA

.78

.79

Correctedcorrelations

r'AV

1.13.99

r'VA

1.031.00

130 ARTHUR E. JENSEN

their span on visual and 36.6% less on audi-tory tasks. If we assume that the amountrecalled is less than the subject's span whenthe series presented exceeds the span be-cause the subsequent items in the seriesretroactively inhibit the earlier items, it ap-pears from Gates' results that auditory pres-entation is more retroactively inhibitingthan visual presentation. In the present ex-periment it is interesting that the digit posi-tion with the lowest probability of correctrecall (in series of three to nine digits) is,on the average, 2.14 positions from the endof the series for visual, and 2.71 for auditory(see Tables 4 and 5). This is consistent withGates' (1916b) finding.

Waugh (1960) has suggested that subj ectsrehearse a digit series (or any other kindof serial list) "cumulatively"—that is, dur-ing presentation, subjects repeat to them-selves the items from the beginning of theseries up to the item being presented. Theearlier items are therefore more thoroughlyrehearsed, and this would account for thepronounced primacy effect (i.e., superior re-call of the first items as compared with thelast) typically found in short-term serialmemory. Corballis (1966) tested this hy-pothesis for immediate recall of digits sepa-rately under visual and auditory presenta-tion. The experiment also included twoconditions of presentation—one favoredcumulative rehearsal, the other did not. Itturned out there was cumulative rehearsalfor visual digits but little, if any, for audi-tory. Corballis concluded that rehearsal,particularly cumulative rehearsal, is morerestricted by auditory than by visual pres-entation.

The Waugh and Corballis hypothesesshould predict a greater primacy effect forthe visual than for the auditory digit series,due to greater cumulative rehearsal from thebeginning of the list for the visual than forthe auditory series. And we should expectthis to show up even more for delayed thanfor immediate recall, since rehearsal shouldlessen decay of the memory trace. This pre-diction is not entirely borne out by the pres-ent data. An index of the primacy effect wasderived for every length of list; it consistedsimply of the ratio of first-position correctresponses to last-position correct responses.

In immediate recall, the mean index was1.24 and 1.12 for visual and auditory tasks,respectively; but delayed recall showed theopposite: 1.47 for visual versus 1.83 forauditory. Since delay of recall increases theprimacy effect for both visual and auditorytasks, it is puzzling, and also contrary to theWaugh-Corballis hypothesis, that thestrongest primacy is found for auditory de-layed recall.

Individual Differences

The fact that there are no significant indi-vidual differences as a function of sensorymodality would seem consistent with thehypothesis that the stimuli, regardless ofsensory channel, are encoded in a singleauditory short-term memory system. Underthis condition, and assuming the subjectshad no primary visual or auditory defects,the only source of individual differences invisual and auditory memory would be in theeffectiveness with which subjects trans-formed visual input into an auditory mem-ory trace. This can be thought of as a kindof mediation process. It might well be thatyoung children, the mentally retarded, thesenile, and certain types of brain-damagedsubjects would show marked individualdifferences in auditory and visual memoryspan. A college population, on the otherhand, by its very nature has been thor-oughly, though indirectly, screened forthese conditions. Also, one would expect amore or less uniformly high level of develop-ment of the rudimentary transformationalor mediational skills involved in digit mem-ory among a young, intellectually superiorsegment of the population. (The presentsubjects were at least in the upper 10% ofall high school graduates.) The present re-sults, therefore, cannot safely be generalizedto a population that includes all ages, alllevels of intelligence, or organic brain ab-normalities.

REFERENCES

ABBOT, E. E. Memory consciousness in orthogra-phy. Psychological Monographs, 1909, 11, 127-158,

BINET, A. Introduction a la psychologie experimen-tal. Paris: Bailliere, 1894.

CAREY, N. Factors in the mental processes of school

INDIVIDUAL DIFFERENCES IN VISUAL AND ATJDITOBT MEMORY 131

children. British Journal 0} Psychology, 1914, 7,453-490.

COHBALLIS, M. C. Rehearsal and decay in immedi-ate recall of visually and aurally presented items.Canadian Journal oj Psychology, 1966,20, 43-51.

DAT, W. F., & BEACH, B. R. A survey of the re-search literature comparing the visual and audi-tory presentation oj injormation. (AF TechnicalReport No. 5921) Dayton, Ohio: Wright-Pat-terson Air Force Base, 1950.

GATES, A. I. Correlations and sex differences inmemory and substitution. University oj Cali-fornia Publications in Psychology, 1916, 1, 343-350. (a)

GATES, A. I. The mnemonic span for visual andauditory digits. Journal of Experimental Psy-chology, 1916,1,393-404. (b)

GATES, A. I. Variations in efficiency during theday, together with practice effects, sex differ-ences, and correlations. University oj CaliforniaPublications in Psychology, 1916, 2, 1-56. (c)

HAO, Y. T. The memory span of 600 Chinese schoolchildren in San Francisco. School and Society,1924, 20, 507-510.

HAWKINS, C. J. Experiments on memory types.Psychological Review, 1897,4,289-294.

HENMON, V. A. C. The relation between mode ofpresentation and retention. Psychological Re-view, 1912,19, 79-96.

JENSEN, A. R., COLLINS, C. C., & VBEELAND, R. W.A multiple S-R apparatus for human learning.American Journal of Psychology, 1962, 75, 470-476.

KAY, B. R. Intra-individual differences in sensorychannel preference. Journal of Applied Psychol-ogy, 1958, 42,166-167.

KIBKPATBICK, E. A. An experimental study ofmemory. Psychological Review, 1894,1, 602-609.

KITSON, H. D. The scientific study of the collegestudent. Psychological Monographs, 1917, 23(1,Whole No. 98).

KOCH, H. L. Some factors affecting the relativeefficiency of certain modes of presenting materialfor memorization. American Journal of Psychol-ogy, 1930,42,370-388.

MAcDouGALL, R. Recognition and recall. Journal ofPhilosophy, Psychology, and Scientific Method,1904,1, 229-233.

McGHiE, A., CHAPMAN, J., & LAWSON, J. S. Changesin immediate memory with age. British Journalof Psychology, 1965,56, 69-75.

MEYER, V. Psychological effects of brain damage.In H. J. Eysenck (Ed.), Handbook of abnormalpsychology. New York: Basic Books, 1961.

MTJNSTERBERGI, H., & BiNQHAM, J. Memory. Psy-chological Review, 1894,1, 34-38.

O'BsiEN, A. A quantitative investigation of theeffect of mode of presentation upon the processof learning. American Journal of Psychology,1921, 32, 249-283.

SPEBLINQ, G. A model for visual memory tasks.Human Factors, 1963, 5,19-31.

WATJQH, N. C. Serial position and the memoryspan. American Journal of Psychology, 1960, 73,68-79.

WECHSLBB, D. The measurement and appraisal ofadult intelligence. (4th ed.) Baltimore, Md.:Williams & Wilkina, 1958.

WISSLER, C. The correlation of mental and physicaltests. Psychological Review Monograph Supple-ment, 1901, 3(6, Whole No. 16).

WOBCHESTEB, D. A. Memory by visual and audi-tory presentation. Journal of Educational Psy-chology, 1926,16, 18-27.

(Received January 11,1970)