Psychonomic Bulletin & Review

1998,5(4), 710-715

Presentation modality affects false memory

REBEKAH E. SMITH and R. REED HUNT
University ofNorth Carolina, Greensboro, North Carolina

Roediger and McDermott (1995)rejuvenated interest in Deese's (1959) paradigm for producing re-
liable intrusions and false alarms. Using this paradigm in three experiments, we demonstrated that
visual study presentation dramatically reduces the rate of false memories. Only auditory study pre-
sentation resulted in equal production of studied and critical items. Correct recall and recognition
were unaffected. The suggestion that visual presentation provides a means for discriminating be-
tween false and true memories was supported by Experiment 3: Pleasantness rating of study items
significantly reduced the creation of false memories regardless of modality.

False memories can be created reliably by presenting the literature supported our suspicions that visual pre-
lists of words that are all strong associates of a nonpre- sentation was the cause of this reduction.'
sented word. The strong associate usually is recalled All published reports of created memories obtained
and recognized at the same rate as that for presented with the Roediger and McDermott/Deese paradigm, with
items. This laboratory technique for invoking predictable one exception, have used an auditory list presentation
intrusions or false alarms has received considerable at- (Johnson et aI., 1997; Mather, Henkel, & Johnson, 1997;
tention in the literature following Roediger and McDer- McDermott, 1996; Norman & Schacter, 1997; Payne,
mott's (1995) replication and extension of Deese's (1959) Elie, Blackwell, & Neuschatz, 1996; Read, 1996; Schac-
work. ter et aI., 1996; Schacter, Verfaellie, & Pradere, 1996).
The experiments that we report here were motivated The one exception to the use of auditory presentation is
by an interest in the effects of item-specific processing of a study by Robinson and Roediger (1997) in which only
list words on recall of the critical, nonpresented associ- visual presentation was used. The rates of critical item
ates. Item-specific processing is assumed to affect a dis- intrusions were noticeably lower in this study than the
criminative process of distinctiveness at retrieval. Thus, rates in other published reports. For instance, Robinson
item-specific processing of list words may facilitate dis- and Roediger reported probabilities of false recall of.31
crimination of those items from nonpresented items and and .30 for 15-word lists as compared with a false recall
thereby reduce the level of false memory. We began this rate of .55 for 15-word lists reported by Roediger and
research by using a powerful item-specific study task McDermott (1995).
that requires subjects to report one thing that is different Although the primary goal of our paper is to report the
about an item from all other items in the list (Hunt & empirical finding of a modality effect on false memories,
Smith, 1996). Unfortunately, we were unable to evaluate we shall speculate about a possible explanation for this
the proposed role of distinctive processing because we effect. This speculation is grounded in the framework
failed to create false memories at a rate equivalent to true that explains memorial phenomena in terms of the inter-
memories in the control condition of that experiment, a action of distinctive and relational processing. Specifi-
condition comparable to the critical conditions ofRoedi- cally, we propose that visual presentation provides a bet-
ger and McDermott (1995). As it happened, we had in- ter means for discriminating between studied items and
advertently changed an apparently important aspect of the related critical items than does auditory presentation.
the experimental paradigm: modality oflist presentation. If this is the case, we should be able to enhance discrim-
Our items were presented via a computer monitor, and ination through some other mechanism even when audi-
the results showed unusually low levels of intrusions in tory presentation is used. In the third experiment, we in-
conditions that normally yield critical intrusions at the vestigated this possibility and obtained favorable results.
same rate as that for actually studied items. A search of
EXPERIMENT 1

We thank Robert Crowder, Arthur Glenberg, David Payne, and Henry In the first experiment, we used a modified version of
Roediger for helpful comments on a previous version of the manu- the Deese paradigm, in which the different lists of asso-
script. This research was supported in part by funds from the Research ciated words were presented blocked in one list followed
Council of the University of North Carolina at Greensboro granted to by a recall test. Given previous research showing that false
the second author. Correspondence may be addressed to R. E. Smith
or R. R. Hunt, Department of Psychology, University of North Carolina
memories can be created at a rate equal to that of true
at Greensboro, Greensboro, NC, 27412 (e-mail: [email protected]. memories using such a technique (McDermott, 1996), we
edu or [email protected]). expected to replicate the effect with auditory presenta-

Copyright 1998 Psychonomic Society, Inc. 710

 FALSE MEMORIES AND STUDY MODALITY 711

tion. However, if presentation modality affects the dis- the visual condition [t(18) = 2.76]. The same trend was
crimination of presented and critical items, we should evident in the noncritical intrusions, but insufficient
not find equivalent production of the studied items and power was available to detect differences. The differ-
critical items in recall following visual presentation, and ences in intrusion rates were not due to shifts in criteria,
the production of critical lures should interact with modal- given the lack of difference between visual and auditory
ity of presentation. conditions on studied items [t(18) = -.72].
The interaction of modality and item type could occur
Method because visual study presentation reduces the likelihood
Subjects and Design. Twenty introductory psychology students par- of thinking of the critical item either at study or at test or
ticipated for credit. The between-subjects manipulation of study modal-
ity, visual or auditory, was combined with the within-subjects manip- because visual presentation increases the subjects' abil-
ulation of item type, which also had two levels: studied (items actually ity to discriminate items that they thought of at study
seen or heard at study) or critical (the high associates of the studied from those that had actually been presented. If the inter-
items that were not actually presented at study). action of modality and item type is solely due to a re-
Materials and Equipment. We randomly chose 10 critical items
from the lists presented in Roediger and McDermott (1995). The 5
duced likelihood of production of an item at retrieval,
highest associates of each critical item made up the 50-item study list. then the interaction of modality and item type should
The studied items were presented blocked during study, always going disappear when a recognition test is used. Alternatively,
from the Ist to the 5th associate. The words were read aloud by an ex- if recognition performance yields the same interaction,
perimenter or presented on a computer monitor. the result would implicate a contribution ofstudy modal-
Procedure. The experiment consisted of two phases: a study phase
and a test phase, separated by a brief period in which the subjects com- ity to the discrimination between critical and studied items.
pleted credit forms. Study presentation rate was one word/I.5 sec in
both the auditory and the visual conditions. At test, 3 min were allowed EXPERIMENT 2
for free recall, with the instructions that the subjects should be rea-
sonably sure that the words were in fact ones that they had either heard
or seen. The primary purpose of the second experiment was to
generalize the finding of a modality effect to the use of
Results and Discussion longer lists of associates and a recognition test. In order
The mean proportions of words produced on the free to do so while also making a direct comparison with pre-
recall test are shown in Table I as a function of modality vious reports of created false memories, we replicated Ex-
and item type. Modality had no reliable effect [F( 1,18) = periment 1 of Roediger and McDermott (1995) with the
1.42, MS e = .01], and the effect of item type [F(1, 18) = additional manipulation of modality. This design in-
20.84, MS e = .01] was qualified by a reliable interaction cludes separate free recall tests after each individual list
of the two variables [F(I,18) = 6.43, MS e = .01] (ex = presentation and a final recognition test. The use oflonger
.05). Individual comparisons indicated that critical and lists is important, because false alarm rates increase
studied item recall did not differ reliably following audi- when more items related to the lures are presented at
tory presentation [t(9) = -1.63], but following visual study (see, e.g., Robinson & Roediger, 1997). The inter-
presentation, studied items were twice as likely to be re- action of modality and item type in our first experiment
called as critical items [t(9) = -4.54]. Moreover, criti- could be an artifact of the relatively short lists of associ-
cal intrusions were reliably higher in the auditory than in ates used.

Table I
Recall and Recognition of Studied and Nonstudied Items
Item Type
False Alarms" Noncritical
Intrusionst
---
Experiment Task Modality M SD M SD M SD M SD M SD
Auditory .26 .10 .21 .10 2.30 2.67
Visual .29 .09 .11 .06 1.l0 .88

2 Recall Auditory .65 .08 .42 .14 1.l0 2.47

Visual .72 .07 .22 .24 .04 .60
Recognition Auditory .88 .11 .82 .27 .11 .15 01 .26
Visual .85 .12 .45 .27 .10 .32 .01 .26

3 Auditory SE .29 .08 .33 .19 .45 .83

Auditory PR .32 .06 .20 .21 .15 .37
Visual SE .33 .09 .18 .13 .20 .41
Visual PR .32 .08 .10 .11 .20 .52
*Mean proportions. "Mean absolute numbers.
712 SMITH AND HUNT

Method following visual presentation [t( 18) = 3.04]. False alarms

Subjects and Design. Twenty volunteers participated in this experi- to related lures were also greater for auditory than for vi-
ment for course credit. The design was the same as that of Experiment 1: sual presentation, but unrelated lures did not differ. Thus,
a between-subjects manipulation of modality and a within-subjects
manipulation of item type. the data clearly show that production of false memories
Materials. We used the same six critical items selected by Roediger in recognition is sensitive to mode of original list encoding.
and McDermott (1995, Experiment I), because these items produced One objection to the preceding conclusion is that the
the highest levels of false remembering in Deese's report. The first 12 modality effect on recognition was an artifact of the
associates of the 6 critical items made up the study list. Subjects either
saw the words on slides or heard a tape of the words.
modality effect on the intervening free recall tests. An
The 42 words on the recognition test were blocked into six sets. Each analysis of false recognition conditionalized on failure
block began with the highest associate and ended with the critical to recall the items suggested that the concern was un-
word. The other 5 words were as follows: an additional study item ran- warranted. Considering only the critical items that were
domly selected from Positions 2-6 on the list of associates; 2 related
not recalled on the intervening free recall tests, the mean
lures, selected from Positions 14 and 15 of the associate list, but not
shown during study; and 2 unrelated lures selected from Position 14 or probability offalsely recognizing the nonrecalled critical
15 of two different non studied associate lists. The construction of test items following auditory presentation (M = .73, SD =
lists matches that used by Roediger and McDermott (1995). .35) was significantly higher than the mean probability
Procedure. Six study/free-recall phases were followed by a final followingvisual presentation (M = .31, SD = .31) [t( 18) =
recognition test. Subjects were instructed that they would hear/see a
series of word lists, and that at the end of each list, they would hear/see
2.81]. Thus, regardless of whether or not items were re-
the word "recall" and should write as many words as they could re- called, false recognition of critical items was signifi-
member from the previous list, but not to guess. As in Roediger and cantly reduced by visual presentation relative to auditory
McDermott's (1995) study, they were instructed to write the last few presentation at encoding.
words first and then recall the rest in any order. Items were presented
The modality effect on false memories in recognition
at a rate of one word every 1.5 sec, and a maximum of2.5 min was al-
lowed for recall. Following the last free recall test, the subjects were suggests that visual presentation does not simply reduce
given a self-paced old/new recognition test. the likelihood ofproducing the critical items at retrieval.
When presented with the critical items on the recogni-
Results and Discussion tion test, subjects in the visual group were still less likely
Recall. Recall scores, averaged across the six tests, are to say that the critical items were part of the list. There-
shown in Table 1. As in the first experiment, modality ex- fore, visual presentation must either prevent the critical
erted no main effect [F(1,18) = 2.39, MSe = .02], but item words from coming to mind at study or provide a better
type was reliable [F(1,18) = 53.43, MSe = .02]. Most im- means of distinguishing between things that are inter-
portantly, item type interacted with modality of presen- nally generated and things that are externally presented.
tation [F(1,18) = 7.13, MS e = .02]. The interaction was a
result of the much larger difference between critical and EXPERIMENT 3
studied items following visual presentation [t(9) = - 6.31]
than following auditory presentation [t(9) = - 3.79], al- The first two experiments demonstrate that visual
though both differences were reliable. Not only is the dif- study presentation of associatively related lists dramati-
ference between critical and studied items larger follow- cally reduces the creation of false memories, relative to
ing visual presentation, but a comparison of critical item auditory study presentation in both recall and recogni-
recall for visual and auditory groups shows a reliable tion. Although subjects did falsely recall and recognize
modality effect on the creation offalse memories [t(18) = some critical items following visual presentation, the
2.30]. The noncritical intrusions followed the same pat- rate of production and recognition of critical items was
tern, but again the small number ofthese intrusions pre- far below the rate for studied items. Equivalent response
cluded meaningful analysis. Thus, the recall data repli- rates across item types occurred only when subjects heard
cate the findings of Experiment 1. the list.
Recognition. The recognition data are presented in The influence of modality could be due to changes in
Table 1, and analysis of these data confirmed that the what comes to mind at study, changes in what comes to
modality effect also occurred in recognition. Item type re- mind at test, or changes in subjects' ability to discrimi-
liably interacted with modality [F(1, 18) = 6.21, MS e = nate internal and external sources. Although our data do
.05], qualifying separate main effects of item type [F(1,18) not provide definitive evidence, it seems unlikely that
= 11.17, MSe = .05], and modality [F(1,18) = 10.12, the modality effect would be due solely to what comes to
MS e = .04]. The interaction of modality was even more mind at test. There is no reason to suspect differential
dramatic on the recognition test than on the recall tests; strategy use in free recall as a function of presentation
subjects in the visual group were much more likely to modality. Furthermore, visual presentation also reduced
call a studied item old than to call a critical item old false memories on the recognition test, where each item
[t(9) = - 3.97], but subjects in the auditory group were must be considered. Likewise, it seems unlikely that pre-
equally likely to call studied and critical items old [t(9) = sentation modality would alter what comes to mind at
- .627]. As with the recall results, false recognition of study. We are not aware of any evidence that associations
critical items was reliably higher following auditory than are modality sensitive.
 FALSE MEMORIES AND STUDY MODALITY 713

More plausible is the possibility that the critical lures The proportion of studied items correctly recalled was
come to mind during presentation of the study list, much examined first. The manipulations ofmodality [F( 1,76) =
in the spirit ofUnderwood's (1965) concept of implicit as- 1.49, MS e = .01] and study task [F(1,76) = 0.18, MS e =
sociative responses. In the absence of differential process- .01] did not influence the production of studied items,
ing ofthe studied and critical items, the critical items will nor was the interaction of the two variables significant
not be discriminated from studied items if they come to [F(1,76) = 1.59, MS e = .01]. Similarly, critical item re-
mind at test. Modality of study presentation provides one call was not affected reliably by the interaction of modal-
such dimension of differential processing. Visual process- ity and study task [F(1,76) = 0.45, MSe = .03]. In con-
ing differs from auditory processing, and on the speculative trast, both modality [F(1,76) = 11.26, MS e = .03] and
assumption that visual perception is less like thought than study task [F(1,76) = 7.21, MS e = .03] influenced the
is auditory perception, critical items would be more eas- likelihood of producing critical items on the recall test.
ily discriminated from list items following visual pre- Thus, while studied items were unaffected by our ma-
sentation than following auditory presentation of list nipulations, critical item production was influenced by
items. The implication is that other forms of differential modality of presentation and by the type of instructions
study processing should reduce the level of false recall. given to the subjects at encoding.
We tested his hypothesis in the third experiment, by com- The exact nature of this influence was investigated
paring the effects of a pleasantness rating orienting task with one-tailed independent t tests, which showed that
with standard instructions. critical item production was significantly decreased by
In the third experiment, we followed the same proce- visual presentation relative to auditory presentation for
dures as those of the first experiment, but with the addi- both the pleasantness rating [t(38) = 1.85] and standard
tion of a pleasantness rating study condition to both the encoding [t(38) = 2.93] groups. Likewise, pleasantness
auditory and visual conditions. Since pleasantness rat- rating decreased critical item production relative to stan-
ing requires consideration of the specific meaning of dard encoding, regardless of whether presentation was
each word, and since that meaning is different for each auditory [t(38) = 1.95] or visual [t(38) = 1.97]. Intru-
word, pleasantness rating encourages item-specific pro- sion of related distractors also was higher following au-
cessing ofa form that should be useful for discrimination ditory presentation with standard instructions than in
among items at retrieval. Thus, we predicted a lower other conditions, but again the small number of related
level of false recall following pleasantness rating of the and unrelated item intrusions precluded meaningful
study list than following standard instructions. analysis.
Finally, within-group comparisons contrasted critical
Method item production and studied item production. No reliable
Subjects and Design. Twenty volunteers, participating for course difference in recall of studied and critical items [t( 19) =
credit, were randomly assigned to each of four experimental groups re-
sulting from the manipulation of study modality (auditory, visual) and
0.75] occurred following auditory presentation with stan-
study task (standard encoding, pleasantness rating). dard encoding instructions. In contrast, both the visual
Materials and Equipment. The word lists from Experiment 2 were presentation pleasantness rating group [t(19) = -7.04]
either shown on slides or read aloud by an experimenter. and the visual presentation standard encoding group
Procedure. The procedure matched that of Experiment I, except for
[t( 19) = -4.35] recalled reliably more studied items
the additional instructional manipulation and an increase in presenta-
tion time to 2 sec. Half of the subjects were given intentional study in- than critical items. Collapsing over modality, critical
structions and the other half were given incidental instructions. The items were less likely to be produced than studied items
incidental instructions requested that each word be rated for pleasant- following pleasantness rating than following standard in-
ness on a scale of I = unpleasant to 5 = pleasant. Recall instructions structions [t(19) = -2.73].
were the same as in Experiment I for all conditions.

Results Discussion
The mean proportions of words produced on the free The results of Experiment 3 replicated the modality effect of Ex-
periments I and 2. In comparison with auditory presentation, visual
recall test are shown in Table I as a function of modality, presentation significantly decreased the likelihood of producing criti-
study task, and item type. The reliable effects of modality cal items on a recall test. Furthermore, subjects who performed a pleas-
[F(1,76) = 6.00, MSe = .02], study task [F(1,76) = 4.78, antness rating task at study were less likely to produce critical items on
MSe = .02], and item type [F(1,76) = 34.15, MSe = .02] the recall test than subjects who intentionally tried to remember the
words. Only the control group, auditory presentation with standard in-
were qualified by reliable interactions of modality and
structions, produced critical items at a rate that was equivalent to the
item type [F(1, 76) = 13.52, MSe = .02] and of study task rate of correct recall of studied items. Both visual presentation and the
and item type [F(I,76) = 7.29, MSe = .02]. Neither the pleasantness rating task resulted in a critical intrusion rate below the
interaction between modality and study task [F(1,76) = rate of studied item production.
.0 I, MSe = .02] nor the three-way interaction [F(1, 76) = The lack of an effect of pleasantness rating on study list items
should not be seen as inconsistent with the enormous literature show-
1.39, MSe = .02] was significant. Separate analyses were ing benefits of semantic orienting tasks. Note that the comparisons in
conducted for the two item types to investigate the sig- this experiment were between pleasantness rating and intentional
nificant interactions. memory instructions. Previous research has found comparable perfor-
714 SMITH AND HUNT

mance following these two study conditions (e.g., Hyde & Jenkins, However, one result reported by Israel and Schacter directly contra-
1969). dicts our data. Subjects in a control condition were given visual pre-
sentation of words, and the false memory effect was not reduced. Im-
GENERAL DISCUSSION portantly, visual presentation was accompanied by simultaneous
auditory presentation. Auditory presentation may have truncated visual
processing of the word analogous to the reduction of sensory process-
The three experiments reported here demonstrate the existence and ing in the generation paradigm when semantic context is provided
reliability ofa modality effect on false memory in the Roediger and Me- (Blaxton, 1988; Jacoby, 1983). Pictures, with their more elaborate sen-
Dermott (1995) paradigm. In comparison with auditory presentation, sory processing requirements, would be immune to the influence ofsi-
visual presentation of study items reduces the level of false memory.
multaneous auditory presentation.
The fact that the modality effect occurs in recognition suggests that In summary, the Roediger and McDermott (1995) paradigm is de-
false memory is not just a product of associative activation at retrieval
signed to bring critical nonpresented items to mind at study. The cues
in recall, but rather that the critical item probably comes to mind at
provided by free recall instructions are then as likely to reinstate the
study as well as at test (Roediger & McDermott, 1995; Underwood,
critical items as the presented items unless some basis for differentia-
1965). If the critical item does come to mind at study, it then is part of
tion occurs at study. The existence of a modality effect suggests that
the encoding experience, much as categorical information is part of the
even subtle differences in processing of critical and presented items at
experience of encoding a block of category instances. Thus, mere pres-
study can service the distinctive processing in retrieval that is neces-
ence on the study list is not sufficient for subjects to discriminate be-
sary to reduce false memory.
tween a list item and a critical lure. Successful discrimination requires
processing along some dimension that differentiates the study item
from the critical item, and speculatively, visual processing of the study REFERENCES
items provides a better basis for differentiation than does auditory pre-
sentation. In support of this general idea, the results of Experiment 3 BLAXTON, T. A. (1988). Investigating dissociations among memory
indicate that other dimensions of differential processing-in this case, measures: Support for the transfer appropriate processing frame-
pleasantness rating of the study list-also will reduce the rate offalse work. Journal of Experimental Psychology: Learning. Memory, &
memory. Cognition, 15,657-668.
Our interpretation of the data derives from the more general view of BLAXTON, T. A. (1996). Distinguishing false from true in human mem-
memory as determined by the processing of similarity and differences ory. Neuron, 17,191-194.
among elements of an event (Hunt & McDaniel, 1993). In retrieval, DEESE, J. (1959). On prediction of occurrence of particular verbal in-
differences are presumed to be important for discriminating among el- trusions in immediate recall. Journal ofExperimental Psychology,
ements of an event, because the event is defined by similarity among 58, 17-22.
the elements. For example, critical items and studied items in this par- HASHTROUDI, S., FERGUSON, S. A., RAPPOLD, V. A., & CHROSNIAK,
adigm share the property of having been experienced at encoding, and L. D. (1988). Data-driven and conceptually driven processes in partial-
therefore, reinstating the similarity processing with the cue "Recall the word identification and recognition. Journal of Experimental Psy-
study items" is insufficient for distinguishing critical and studied chology: Learning, Memory. & Cognition, 14, 749-757.
items. Differential processing of the items at study facilitates discrim- HUNT, R R, & McDANIEL, M. A. (1993). The enigma of organization
ination at retrieval and reduces the recall and recognition of critical and distinctiveness. Journal ofMemory & Language, 32, 421-445.
items. HUNT, R. R., & SMITH, R E. (1996). Accessing the particular from the
Existing evidence is congruent with various aspects of our interpre- general: The power of distinctiveness in the context of organization.
tation. The assumption that visually driven processing is more dis- Memory & Cognition, 24, 217-225.
criminable from thought than is auditorially driven processing has not HYDE, T. S., & JENKINS, J. J. (1969). Differential effects of incidental
been directly tested in the reality monitoring literature, but Johnson, tasks on the organization of recall of a list of highly associated
Nolde, and De Leonardis (1996) have reported that source monitoring words. Journal ofExperimental Psychology, 82, 472-481.
is more highly correlated with visual detail than with auditory detail in ISRAEL, L., & SCHACTER, D. L. (1997). Pictorial encoding reduces false
the original event. Notice, however, that our assumption is that modal- recognition of semantic associates. Psychonomic Bulletin & Review,
ity of presentation will be useful only to the extent that it differentiates 4,577-581.
among items. In the Roediger and McDermott (1995) paradigm, we JACOBY, L. L. (1983). Remembering the data: Analyzing interactive
would not predict a modality effect on studied items. This prediction processes in reading. Journal of Verbal Learning & Verbal Behav-
is not only consistent with our results but also with those of Hash- ior, 22,485-508.
troudi, Ferguson, Rappold, and Chrosniak (1988), who found that au- JOHNSON, M. K., NOLDE, S. E, & DE LEONARDIS, D. M. (1996). Emo-
ditory study increased false alarms on a recognition test without influ- tional focus and source monitoring. Journal of Memory & Lan-
encing hits. guage,35,135-156.
Our assumption about the importance of processing similarities and JOHNSON, M. K., NOLDE, S. E, MATHER, M., KOUNIOS, J., SCHACTER,
differences among items is also consistent with existing evidence. Me- D. L., & CURRAN, T. (1997). The similarity of brain activity associ-
Dermott (1996) and Mather et al. (1997) found increased levels of crit- ated with true and false recognition memory depends on test format.
ical item recall with blocked list presentation relative to random list Psychological Science, 8, 250-257.
presentation at study. Blocked presentation encourages relational pro- MATHER, M., HENKEL, L. A., & JOHNSON, M. K. (1997). Evaluating
cessing, and in this case the relationship among items is the critical characteristics of false memories: Remember/know judgments and
item. Therefore this item is brought to mind more readily than with ran- memory characteristics questionnaire compared. Memory & Cogni-
dom presentation. tion, 25, 826-837.
Read (1996) reported that instructions to remember the order of pre- McDERMOTT, K. B. (1996). The persistence of false memories in list
sentation reduced the recall of critical items. Because information cor- recall. Journal ofMemory & Language, 35, 212-230.
responding to serial order is different for each item, successful serial NAIRNE, J. S., NEATH, I., SERRA, M., & BYUN, E. (1997). Positional dis-
recall has been attributed to distinctive processing (Nairne, Neath, Serra, tinctiveness and the ratio rule in free recall. Journal of Memory &
& Byun, 1997). Thus, Read's data are consistent with the importance Language,37,155-166.
of distinctive processing in reducing false recall. NELSON, D. L., REED, V. S., & McEvoy, C. L. (1977). Learning to order
Israel and Schacter (1997) discovered that false memory is less likely pictures and words: A model of sensory and semantic encoding.
ifthe study materials are pictures rather than words. Nelson, Reed, and Journal ofExperimental Psychology: Human Learning & Memory,
McEvoy (1977) argued that the picture superiority effect in memory is 3,484-497.
due to more elaborate sensory processing required by pictures, which NORMAN, K. A., & SCHACTER. D. L. (1997). False recognition in younger
in turn enhances the distinctiveness of pictures at retrieval. Thus, Israel and older adults: Exploring the characteristics of illusory memories.
and Schacters data can be seen as congruent with our interpretation. Memory & Cognition, 25, 838-848.
 FALSE MEMORIES AND STUDY MODALITY 715

PAYNE, D. G., ELIE, c.J., BLACKWELL, J. M., & NEUSCHATZ,J. S. (1996). SCHACTER, D. L., VERFAELLIE, M., & PRADERE, D. (1996). The neuro-
Memory illusions: Recalling, recognizing, and recollecting events psychology of memory illusions: False recall and recognition in am-
that never occurred. Journal of Memory & Language, 35, 261-285. nesic patients. Journal ofMemory & Language, 35, 319-334.
READ, J. D. (1996). From a passing thought to a false memory in 2 min- UNDERWOOD, B. J. (1965). False recognition produced by implicit ver-
utes: Confusing real and illusory events. Psychonomic Bulletin & bal responses. Journal of Experimental Psychology, 70,122-129.
Review, 3, 105-111.
ROBINSON, K. J., & ROEDIGER, H. L., III (1997). Associative processes in NOTE
false recall and false recognition. Psychological Science, 8, 231-237.
ROEDIGER, H. L., III, & McDERMOTT, K. B. (1995). Creating false I. Blaxton (1996) compared several neuroimaging studies and de-
memories: Remembering words not presented in lists. Journal ofEx- tected a pattern suggestive of the importance of modality of presenta-
perimental Psychology: Learning, Memory, & Cognition, 21, 803-814. tion.
SCHACTER, D. L., REIMAN, E., CURRAN, T., YUN, L. S., BANDY, D., Mc-
DERMOTT, K. B., & ROEDIGER, H. L., III (1996). Neuroanatomical
correlates of veridical and illusory recognition memory: Evidence (Manuscript received October 14, 1997;
from positron emission tomography. Neuron, 17, 267-274. revision accepted for publication February 24, 1998.)