Full text of DOI:10.3758/BF03195268

Memory & Cognition
2002, 30 (1), 89-96
Top-down processing and the suffix effect in
young and older adults
MAURA PILOTTI, TIM BEYER, and MARIYA YASUNAMI
Washington University, St. Louis, Missouri
The suffix effect has generallybeen viewed as reflectingan automatic component involving the termi-
nal item and a component susceptible to top-down influences involving the preterminal items (Balota
& Engle, 1981). Recent findings have raisedquestions about this widely acceptedview of the suffix effect
by showing that the terminal component is not insensitive to top-down influences (Bloom & Watkins,
1999). In the present study, we attempted to uncover the source of these contradictory findings by as-
sessing whether the terminal item’s susceptibility to top-down influences might depend on subjects’
first-hand experience of the extent to which recall is affectedby the suffix. Furthermore, we examined
whether age differences in the suffix effect could be attributed to age-related declines in inhibitory
processes. Our findings supported both predictions.
In thestandardsuffix procedure,subjectshearlistsof ran- sources that otherwise couldbe devotedto the preterminal
domly organized digitsor letters, followed by either a suffix items, further depressing their recall.
go
(e.g., theword ) or a silentintervalof equivalentduration.
The burden of proof for the two-componentaccount re-
Their taskisto recallthedigitsorlettersintheorderinwhich sides in the demonstration that factors that can modulate
theywere presentedandtoignorethesuffix.Whentherecall the allocation of attention affect the preterminal compo-
of the suffix lists is compared with the recall of the con- nent of the suffix effect but have little or no influence on
trol lists, there is a reliable increase in errors for the items theterminal component.Among thesefactors, presentation
in the proximity of the suffix, with the terminal item, rate and practice afford two of the most compelling tests
which enjoysnear-perfect performance in the controlcon- for this account,because increased experiencewith the re-
dition,displayingthe largest incrementin errors. This phe- call task and slowing the rate of presentation can promote
suffix effect
nomenon is called the
.
attentionto the recall lists, increasing the opportunitiesfor
One of the most widely known accounts of the suffix top-down processing such as rehearsal and list organiza-
effect postulatesthat this phenomenonarises from two dis- tion. Accordingly, the two-component account predicts
tinct memory mechanisms, each influencingprimarily, but that practice and slowing the presentation rate should re-
not exclusively, different portions of the serial position duce the preterminal component of the suffix effect but
curve (Balota & Engle, 1981; Greenberg & Engle, 1983; should have less influence (if any) on the terminal compo-
Penney,1985).Accordingto this account(hereafter named nentif thiseffect arises from thesuffix’s maskingthe echoic
two-componentaccount
), the effect of the suffix on the trace of the last item.
the
terminal item reflects, for the most part, an automaticform
The few studies that have examinedthe influenceof rate
of processing typical of the functioning of echoic mem- and practice on the suffix effect in young adults have re-
ory. That is, the suffix is thoughtof as masking the echoic ported contradictory results. For instance, Bloom and
memory trace of the terminal item, thereby impairing its Watkins (1999) found that slowing the presentation rate
recall (Cowan, 1984; Crowder, 1971, 1978; Crowder & and practice reduced the terminal, but not the preterminal,
Morton, 1969; Greene & Crowder, 1984). In contrast, the componentof the suffix effect, whereas Balota and Engle
effect of the suffix on the preterminal items, which are the (1981) reported that these factors influenced the preter-
items most difficult to recall, is assumed to reflect, for the minal component only. The most striking methodological
most part, subjects’ failure to excludethe suffix from atten- difference between these studies involves the administra-
tion. As a result, the suffix is thought of as taking up re- tion of suffix and control lists. Bloom and Watkins used a
within-subjectsdesignwith random presentation,whereas
Balota and Engle adopted either a between-subjectsdesign
(Session 1) or a blockedpresentation(Session 2). If the rea-
This work was supported by Grant F32 DC00342 from the National
son for focusing attention on specific list items is the dif-
ficulty with which they are recalled, as was postulated by
Balota and Engle, can these contradictory findings reflect
the allocationof attentionto different list items, shaped by
subjects’experiencewith thesuffix? Of course, for subjects
exposedtobothsuffix and controllists, as in theBloomand
Institute on Deafness and Other Communication Disorders. We thank
Ben Chi for his help in conductingthe experiment. Correspondencecon-
cerning this article should be addressed to M. Pilotti, Department of
Psychology, HunterCollege, 695Park Avenue,New York,NY 10021(e-mail:
mpilotti@hunter.cuny.edu).
—Accepted by previous editorial team
89
Copyright 2002 Psychonomic Society, Inc.

90
PILOTTI, BEYER, AND YASUNAMI
Watkins study, the terminal item, which displays near- stimulus suffix paradigm. First, evidence of age-related
perfect recall in the controllists, displaysthe largest decre- declines in inhibitory processes (see Carlson, Hasher,
ment in recall in the suffix lists. Consequently, one would Connelly, & Zacks, 1995; Connelly, Hasher, & Zacks,
expect subjects to direct attention to the terminal item as 1991;Hasher & Zacks, 1988;Zacks & Hasher, 1994) led to
a way of improving its recall, making the effect of the suf- the prediction that excluding the suffix from attention
fix on this item susceptible to top-down processing (as wouldbeparticularlyarduousforolderadults.Consequently,
was reported by Bloom & Watkins). In contrast, for sub- older adults were expected to exhibita larger suffix effect
jects exposed to onlyone type of list, as in Session 1 of the on the items whose processing depends on attentionalre-
Balota and Engle experiment, the difficulty of recalling sources. The two-component account predicts that the
the preterminal items (either with or without the suffix) is resource-dependent nature of the preterminal suffix effect
likely to be the most salient aspect of their performance. would make this effect susceptibleto the influence of age-
Consequently, one would expect subjects to focus atten- related declines in inhibitory processes, whereas the auto-
tion on these items to ameliorate their recall (as was re- matic nature of the terminal suffix effect would shield this
ported by Balota & Engle in Session 1). Prior practice with effect from such declines. In contrast, the hypothesis that
only one type of list may also be expected to preserve this direct knowledgeof the effect of the suffix on recall might
strategy even though subjects recall both types of lists (as modulate the allocation of attention to specific list items
was reported by Balota & Engle in Session 2).
led to the prediction that age differences in the pretermi-
In lightof theseassumptions,thegoalofthepresentstudy nal and terminal components of the suffix effect would
was to test the hypothesis that direct knowledge of the ef- depend on the administration of list type. Second, evi-
fect of the suffix on recall can modulate the allocation of dence of age-related declines in temporal memory (Fabi-
attention in the stimulus suffix paradigm. To this end, we ani & Friedman, 1997; Lowe & Rabbitt, 1997; Maylor,
examined the effect of the suffix on subjects’ recall as a Vousden, & Brown, 1999; Parkin, 1997) led to the predic-
function of practice, presentation rate, and administration tion that older adults would yield an overall lower recall
of suffix and controllists. With respect to the latter factor, rate, with age-related declines in inhibitory functioning
between-
the subjectsreceivedeithercontrolor suffix lists (
further depressingrecall in the suffix lists. Lastly, evidence
subjectscondition
) or bothlistspresentedinseparateblocks that olderadults’recall improves withpractice(Taub, 1973;
blockedcondition
). The subjectswere warned, via instruc- Taub & Long, 1972) and slowing the rate of presentation
(
tions and training trials, as to the type of lists they were (Wingfield,Poon,Lombardi,& Lowe, 1985)suggestedthat
required to recall prior to each block or during the entire these factors might attenuatethe effect of the suffix on the
experimentalsession,so thattheoccurrenceofthesuffixwas items that are sensitive to shifts in attentional resources.
predictable in both experimental conditions. This design As was discussed earlier, the two-component account and
permitted us to assess whether the effects of top-down the hypothesis proposed here made quite different predic-
processing, which Bloom and Watkins (1999) found to be tions as to which items might be sensitive to such shifts.
limited to the terminal item with random presentation of
METHOD
suffix and control lists, could be attributedto subjects’ at-
tempts to predict the occurrence of the suffix. Specifically,
we examined whether subjects exposed to both suffix and
control lists would extend attention to the preterminal
items when the occurrence of the suffix was entirely pre-
dictable.
In the present study, we examined the suffix effect not
only in young adults, but also in older adults, as an addi-
tional test of the two-component account. Surprisingly,
the handful of studies that have investigatedthe suffix ef-
fect in different age groups not onlyhave neglectedto pro-
vide a direct test for this account, but also have yielded
Subjects
The subjects were 180 young adults and 160 older adults. The
young adults were Washington University undergraduate students,
who participated in the experiment for course credit. Their mean age
was 19.40 years (16–31). The older adults were community mem-
bers from the Aging and Development Subject Pool, who were paid
for their participation. Their mean age was 75.61 years (60–86). All
the subjects reported themselves as being in a healthy condition for
their age, and none wore a hearing aid. They were given the vocab-
ulary subsection of the Shipley test (Shipley, 1940) to ascertain their
cognitive status. The mean vocabulary score of young adults was
33.60 (SD = 2.61), and the mean vocabulary score of older adults was
contradictoryfindings.Forinstance,ManningandGreenhut- 35.48 (SD = 2.60). Sixty-eight young adults and 64 older adults par-
ticipated in the between-subjects condition (fast-rate, 16 young and
Wertz (1990) asked young and older adults to recall 16
16 older adults per list type; slow-rate, 18 young and 16 older adults
six-letter lists, half with the suffix and half without it, all
per list type). Fifty-six young adults and 48 older adults participated
auditorilypresentedat a rate of one letterper second. With
in the blocked condition (28 young and 24 older adults per presen-
tation rate). The same number of subjects participated in a blocked
suffix and control lists in separate blocks, Manning and
Greenhut-Wertz found no age differences in the terminal
component of the suffix effect, but a larger preterminal
componentin the aged. In contrast, Wiseman (1998), who
used a similar experimentaldesign, found age differences
in both the terminal and the preterminal components. In
lightofthesecontradictoryfindings,in thepresent study, we
relied on evidencefrom theliteratureon aging to formulate
condition involving eight digits (described below). Three older
adults who either repeatedly recalled a digit that was not part of the
stimulus material (seven) or displayed difficulties in understanding
the instructions were excluded from the experiment.
Stimuli, Procedure, and Design
The stimuli of this experiment consisted of the digits 1–9 (ex-
cluding the bisyllabic digit 7), and the word go, all spoken by a male
three predictionswith respect to older adults’ recall in the talker, digitized at a 20-kHz sampling rate, and equated in amplitude

SUFFIX EFFECT AND AGE
91
Table 1
The Influence (in Percentages) of Age, Practice, and Rate on the Terminal and Preterminal
Suffix Effects, in Each Experimental Condition
Factor
Terminal Item
Between Subjects
Blocked
Blocked (8 digits)
Age (young vs. old)
Practice (before vs. after)
Rate (fast vs. slow)
16 vs. 19
18 vs. 16
18 vs. 17
10 vs. 27 *
22 vs. 15 **
20 vs. 16
17 vs. 27*
28 vs. 16 *
27 vs. 17*
Preterminal Items
Age (young vs. old)
Practice (before vs. after)
Rate (fast vs. slow)
4 vs.15 *
12 vs. 7 *
7 vs. 12
6 vs. 12*
13 vs. 5*
10 vs. 8
6 vs. 13*
12 vs. 6 *
10 vs. 9
*Effect reliable at p < .05. **Effect reliable only in young adults.
and length (400 msec). Prior to the experiment, the loudness of the rect serial positionand a 0 was given in any other case. For
stimulus material was adjusted separately for each subject via a brief
the purpose of our study, the statistical analyses focused
identification test to ensure audibility.
In the experiment, the subjects received 96 experimental lists,
1999). The four serial positionsprior to the lastare referred
constructed by sampling without replacement either seven or eight
on the last five serial positions (see Bloom & Watkins,
preterminal positions
to as
. Therefore, any reliable decre-
digits, within the constraints that no consecutive digits could be
neighbors. There were two experimental conditions, each unique
with respect to the way suffix and control lists were administered. In
the between-subjects condition, the subjects heard two blocks of 48
seven-digit lists, all of one type (either suffix or control). In the blocked
condition, the subjects received two blocks of seven-digit lists, one
with 48 suffix lists, and the other with 48 control lists, with the order
in which blocks were presented counterbalanced across subjects.
The blocked condition was replicated with eight-digit lists, to assess
whether increasing the demands of the recall task would preserve
the results observed with seven-digit lists. In all the conditions, prac-
tice effects were measured as recall rate changes from the first block
of 48 lists to the second block of 48 lists (before practice vs. after
practice). Furthermore, in all the conditions, the rate at which digits
were presented was manipulated between subjects to ensure the de-
velopment of listening strategies specific to the prevailing presenta-
tion rate. Thus, half of the subjects heard digits at a slow rate (1 digit/
sec) and half at the fast rate (2 digits/sec).
An experimental trial consisted of a sequence of digits, followed
by either the suffix go in rhythm with the digits or a silent interval
of equivalent duration (400 msec). A 400-msec tone serving as a cue
for the subjects to start their recall was placed 4 sec after the last
digit for both the suffix and the control sequences (see Balota &
Engle, 1981). The tone was followed by a 15-sec silent interval, dur-
ing which the subjects were instructed to report the digits in strict se-
rial order, without making corrections or backtracking on a response
sheet containing either seven or eight rows. In the blocked condition
and in the between-subjects condition involving suffix lists, the sub-
jects were informed that the word go would occur at the end of some
(or all) sequences of digits. They were instructed to ignore the word
and not recall it along with the digits. All the subjects initiated each
trial by pressing the Enter key of the computer keyboard. The sub-
jects were given 10 training trials to familiarize themselves with the
assigned presentation rate and type(s) of list. Depending on the ex-
perimental condition, the training trials contained either suffix and
control lists or only one type of list. The experiment involved the
following factors: experimental condition (between subjects vs.
blocked), age (young vs. old), serial position (preterminal vs. terminal),
list type (suffix vs. control), practice (before practice vs. after prac-
tice), and presentation rate (fast vs. slow).
ment in the recall of the preterminal items (averaged
across these items) owing to the presence of the suffix rep-
resentsthepreterminalsuffix effect. Position7 (or 8 ineight-
terminal position
, and any
digit lists) is referred to as the
decrement in the recall of the last item owing to the pres-
ence of the suffix represents the terminal suffix effect.
Figures 1–3 display the mean percentage of correct recall
in each experimentalconditionas a function of age, serial
position (1–7 or 1–8), list type, practice, and presentation
rate, whereas Table 1 displays changes in the preterminal
and terminal suffix effects owing to age, practice, and rate.
The findings of this experiment can be summarized in
three points(see Table 1). First, when the subjectsrecalled
onlyone type of list (between-subjectscondition),age dif-
ferences emerged in the preterminalcomponentof the suf-
fix effect. However, when the subjects recalled both suffix
and control lists (blocked condition), age differences in-
volvedthe terminal and the preterminalcomponentsof the
suffix effect. Second, practice reduced the terminal and
preterminalsuffix effects in theblockedcondition,whereas
practice reduced only the preterminal suffix effect in the
between-subjects condition. Third, the effects of slowing
the rate of presentation were weak overall across all the
experimentalconditionsandagegroups.Theseobservations
3
were supported by a 2 (young vs. old) 2 (suffix vs. con-
3
3
trol list) 2 (slow vs. fast) 2 (before vs. after practice)
analysisof variance (ANOVA) on the percentage of correct
responses at the terminal and preterminal positions in each
experimental condition. Note that the results reported here
are significant at the .05 level, unless otherwise indicated.
Between-Subjects Condition (Seven-Digit Lists)
Terminalitem. In bothagegroups,thesuffix depressed
F
MS
=
e
the recall of the terminal item [ (1,124) = 54.63,
355.79], but young adults were overall better at recalling
MS
F
RESULTS
this item than were older adults [ (1,124) = 24.27,
=
e
F
355.79]. No other effects reached significance ( s < 2.3),
A strict scoring procedurewas utilizedfor the analyses, suggesting that age, practice, and rate did not reliably af-
so that a 1 was given for a correct digit recalled in the cor- fect the terminal component of the suffix effect.

92
PILOTTI, BEYER, AND YASUNAMI
Figures 1. Mean percentages of correct recall in young and older adults as a function of serial position, list type, practice, and pre-
sentation rate in the between-subjects condition.
Preterminalitems. Althoughthe suffix affected the re- 191.22], but older adults recalled this item less well than
F
F
MS
call of the preterminalitems inbothage groups [ (1,124)= did young adults [ (1,96) = 39.78,
e = 361.06]. Older
e = 302.72], recall was higher overall in young adults also yielded a larger terminal suffix effect than
MS MS
e = 191.22]. Although,
MS
19.73,
F
F
adults [ (1,124) = 59.40,
groups, recall increased with slowing the rate of presenta- in both age groups, overall performance was higher after
MS MS
= 302.72]. In both age young adults [ (1,96) = 19.33,
e
F
F
tion [ (1,124) = 10.608,
e = 302.72] and improved as practice [ (1,96) = 5.89,
e = 361.06], the influence of
= 83.10]. practice on the terminal suffix effect was only marginally
F
MS
a function of practice [ (1,124) = 15.93,
The preterminalsuffix effect increasedwith age (1,124) = significant [ (1,96) = 3.20,
MS
e
F
F
MS
p
e = 191.22, = .08; other
s < 1.3]. Tests of simple effects indicated that practice
e = 83.10]. No other effects were reliable, ( s < reliably reduced the terminal suffix effect in young adults
MS
F
F
5.44, e = 302.72,anddecreasedwith practice [ (1,124)=
MS
F
5.49,
1.4).1
[ (1,54)= 5.79, e = 58.61;13% vs. 6%], but not in older
F
F
adults( = 1; 30% vs. 23%), even thoughthere was a sim-
ilar reduction in the overall size of this effect.
Blocked Condition (Seven-Digit Lists)
Terminal item. The suffix affected the recall of the ter-
Preterminal items. In both age groups, the suffix af-
F
fected the recall of the preterminal items [ (1,96)= 48.37,
F
MS
minal item in both age groups [ (1,96) = 89.44,
=
e

SUFFIX EFFECT AND AGE
93
Figure 2. Mean percentages of correct recall in young and other adults as a function of serial position, list type, practice, and pre-
sentation rate in the blocked condition.
MS MS
F
e = 84.67],andpracticehada marginallybeneficialeffect was higher after practice [ (1,96) = 19.39, e = 352.05]
F
MS
p
F
MS
on overallperformance [ (1,96) = 3.18, e = 436.08, = and at the slow rate [ (1,96) = 6.43, e = 352.05]. Older
.08]. Young adults recalled more preterminal items than adults yielded a larger terminal suffix effect than did
MS MS
e = 209.82], and practice
F
F
older adults [ (1,96) = 67.14,
= 436.08] and had a young adults[ (1,96) = 6.02,
e
F
MS
F
smaller preterminal suffix effect [ (1,96) = 5.46,
=
reduced theterminalsuffix effect in bothgroups[ (1,96)=
84.67]. Practice reduced the preterminal suffix effectein 9.36,
= 209.82]. Interestingly, although slowing the
MS
e
F
MS
F
both age groups [ (1,96) = 11.46,
e = 84.67;other s < rate of presentationalso reduced the terminal suffix effect
MS
e = 209.82], a test
F
1.1; see note 1].
in both age groups [ (1,96) = 6.29,
of simple effects illustrated that the influence of rate was
MS
e = 168.89;
F
Blocked Condition (Eight-Digit Lists)
Terminal item. Although the suffix affected the recall other s < 1.3].
of the terminal item in both age groups [ (1,96) = 122.34,
reliableonly after practice [ (1,48) = 4.19,
F
F
Preterminal items. The preterminal items displayed a
MS
e = 209.82], young adults were better overall at recall- main effect for all the factors involved:list type, age, prac-
F
ing this item than were older adults [ (1,96) = 22.94, tice, and rate. These effects indicatedthat althoughthe suf-
MS
e = 352.05]. In both age groups, overall performance fix affected the recall of the preterminal items in both age

94
PILOTTI, BEYER, AND YASUNAMI
Figure 3. Mean percentages of correct recall in young and other adults as a function of serial position, list type, practice, and pre-
sentation rate in the blocked condition with eight-digit lists.
F
MS
= 65.80], overall perfor- rate blocks, top-down processing (via practice) influenced
e
groups [ (1,96) = 60.51,
mance was higher in youngadults[ (1,96)= 26.70,
593.66]. In both age groups, performance increased with suffix effect. In contrast,when thesubjectsrecalledonetype
MS
F
MS
=
e
both the terminal and the preterminal components of the
F
slowing the rate of presentation [ (1,96) = 5.47,
=
of list, top-down processing influenced only the pretermi-
593.66]and improved as a function of practice [ (1,96) = nal component.These results are consistentwith two basic
MS
e
F
14.14, e = 593.66]. Older adultsyieldeda larger preter- assumptions regarding subjects’ perceptions and corre-
F
MS
minal suffix effect [ (1,96) = 9.52,
= 65.80], and sponding reactions to the effects of the suffix on recall.
practice reduced the preterminal suffix efefect in both age First, attention to specific list items is driven by the diffi-
F
MS
F
groups [ (1,96) = 7.80,
e = 65.80; other s < 1].
culty with which they are recalled (Balota & Engle, 1981).
Second, subjects’ assessment of which items are difficult
to recall is shaped by their experiences with the effects of
the suffix on recall, which depend on the administrationof
GENERAL DISCUSSION
In the present experiment, we found that, when the sub- list type. We have argued earlier that when subjects recall
jects recalled both the suffix and the control lists in sepa- both suffix and control lists, the terminal item, which en-

SUFFIX EFFECT AND AGE
95
joys near-perfect performance in the control lists, be- help subjects, whose task is to recall both suffix and con-
comes relatively more difficult to recall in the suffix lists. trol lists, to encode the list items that are deemed to be in
In contrast, when subjectsrecall either type of list, they do need of additionalprocessing. Second, at the fast rate, the
not have the opportunity to experience how the recall of terminal suffix effect observed with eight-digit lists was
the terminal item changes as a function of the presence of relatively larger than the terminal suffix effect observed
the suffix. Thus, for these subjects, the preterminal items with seven-digit lists. A larger terminal suffix effect has
are likely to be perceived as the most difficult to recall, more latitude to be influenced by encoding strategies that
relativeto all the other items (includingthe terminal item). subjects may have developed during the first block of tri-
Our findingsindicatethat subjects’ direct experiencewith als. Thus, if a slow rate of presentation facilitates the ap-
the effect of the suffix on recall afforded by the within- plication of such strategies, it is not surprising that its in-
subjects design directs attention to the terminal item and fluence was detected on the terminal item of eight-digit
thus determines the terminal component’s susceptibility lists.
to top-down influences. Furthermore, the findings of our
With respect to aging, the present study yielded an in-
experiment and those of Bloom and Watkins (1999) sug- teresting pattern of results. Given the age-related declines
gest that, when list type is administered within subjects, in inhibitory processes documented in the literature, we
the predictability of the suffix plays a role in directing hypothesizedthat older adults would be more affected by
subjects’ attention to the preterminal items. Specifically, the presence of the suffix. Specifically, we expected age-
blocked presentation of suffix and control lists permits relateddeclinesininhibitoryprocessestomagnifysubjects’
subjectstodirectattentionto thepreterminalitems, whereas difficulties in excluding the suffix from attention, allow-
random presentation leads subjects to restrict attention to ing the suffix to take away a larger portion of the process-
the terminal item (as was shown by Bloom & Watkins, ing resources that could be allocatedto specific list items.
1999), perhaps as a result of their attempts to predict the In the present study, older adults not only performed less
occurrence of the suffix. However, even with a predictable well than young adults across all the experimental condi-
suffix, the influence of practice on the preterminal suffix tions and list types, but also displayed an increased effect
effect may require a large number of trials before it can of the suffix on those sections of the recall lists that pro-
actually be detected. In fact, Watkins and Sechler (1989, ducedreliablepracticeeffects. Two pointscanbe made with
Experiment 1), who also adopted a blocked design for list respect to these findings:First, because age differences in
type but used a smaller number of lists, found that prac- the suffix effect mirrored the influence of practice, such
tice only produced a trend in the direction of a decrement differences appear to reflect older adults’ failure to focus
in the preterminal component (see their Figure 1, Block 1 attention on specific portions of the recall lists while at-
vs. Block 3, p. 683).
tempting to exclude the suffix from attention. However,
Interestingly, the subjects of our study were not equally across all the experimental conditions, even young adults
sensitiveto the two factors that were assumed to modulate were not able to eliminate the disrupting effect of the suf-
top-down processing in the stimulus suffix paradigm: fix on recall, suggesting that age-related declines in in-
practice and rate. As was discussed earlier, the influenceof hibitory processes simply magnify subjects’ pre-existing
practice was robust. In contrast, slowing the presentation difficultiesin ignoringa task-irrelevant stimulus. Second,
rate had an overall beneficial effect on the recall of the because there were age differences in overall recall, older
preterminal items of the suffix and control lists in the adults’ performance in the stimulus suffix paradigm ap-
between-subjects condition and reduced the terminal suf- pears to reflect age-related declines not only in inhibitory
fix effect in the blocked condition involving eight-digit processes, but also in memory for temporal order (Fabiani
lists, but onlyafter practice. The overallbeneficialeffect of & Friedman, 1997; Maylor et al., 1999). Declines in tem-
slowing the presentation rate on the preterminal items in poralmemory havebeengenerallyattributedtoolderadults’
the between-subjects condition indicates that a slow rate failureto use contextualinformationefficiently(see Parkin,
may help subjects, who are recalling either suffix or con- 1997), arising from selective neural losses in the pre-
trol lists, to encode the list items that are the most in need frontal cortex (Lowe & Rabbitt, 1997). These difficulties
of processing in both lists. It also suggests that, for these can be assumed to produce more displacement errors (re-
subjects, slowing the rate of presentation is unlikely to calling a digitin the incorrect position),yieldingthe lower
contributeto the development of strategies directed at ex- recall rate observed in the present experiment.
cluding the suffix from attention. However, how can we
The age differences uncovered in the present study are
explain the finding of the blocked condition (eight-digit consistent with those reported by Wiseman (1998), who
lists) in which a slower rate reduced the effect of the suf- administered suffix and control lists in separate blocks
fix, but onlyon the terminal item and after practice?There and found age differences in both the terminal and the
are two pointsto considerhere.First, inthiscondition,slow- preterminalcomponentsof the suffix effect. They conflict,
ingthepresentationrate had a beneficialeffect on therecall however, with those reported by Manning and Greenhut-
of the preterminal and terminal items of both suffix and Wertz (1990), who adoptedthe same designbut found age
control lists, as predicted by our hypothesis. Thus, as was differences only in the preterminal component. Unfortu-
discussed earlier, slowing the rate of presentation may nately, both experiments involveda limited number of ex-

96
PILOTTI, BEYER, AND YASUNAMI
perimental trials per list type and did not include training
trials. Thus, their findings may reflect differences in how
separate groups of subjects initially adjust to the stimulus
suffix procedure. Of course, this argument does not apply
to the study by Parkinson and Perey (1980), who com-
pared the performance of 10 older adults in the stimulus
suffix paradigm with that of a large group of young adults
with different digit spans. In this experiment, subjects re-
called 120 lists (60 per list type) organized in blocks of
lobe deficits. In P. Rabbitt (Ed.), Methodology of frontal and execu-
tive function (pp. 39-59). Hove, U.K.: Psychology Press.
Manning K. S., & Greenhut-Wertz, J. (1990). Visual and auditory
modality and suffix effects in youngand elderly adults. Experimental
Aging Research, 16, 3-9.
Maylor, E. A., Vousden, J. I., & Brown, G. D. A. (1999). Adult age
differences in short-term memory for serial order: Data and model.
Psychology & Aging, 14, 572-594.
Parkin, A. J. (1997). Normal age-related memory loss and its relation
to frontal lobe dysfunction.In P. Rabbitt(Ed.), Methodologyof frontal
and executive function (pp. 177-190). Hove, U.K.: Psychology Press.
30. With the young adults’ data collapsed across digit Parkinson, S. R., & Perey, A. (1980). Aging, digit span, and the stim-
ulus suffix effect. Journal of Gerontology, 35, 736-742.
Penney, C. G. (1985). Elimination of the suffix effect on preterminal
items with unpredictablelist length:Evidence for a dual model of suf-
fix effects. Journal of Experimental Psychology: Learning, Memory,
span, age differences emerged in boththe terminal and the
preterminal suffix effects. In the present study, we also
found such differences in the blocked condition, suggest-
ing that the subjects in the Parkinson and Perey study fo-
cused attention on both the preterminal and the terminal
items of the recall lists. However, Parkinson and Perey did
not alert subjects to the effect of the suffix on recall prior
to the experimental session and did not examine perfor-
manceas a functionofpractice.Therefore, it is possiblethat,
in thisexperiment,the age differences observedon the ter-
minal component emerged only after the first two blocks
of trials—that is, after the subjects had the opportunity to
be exposed to both suffix and control lists.
& Cognition, 11, 229-247.
Shipley, W. C. (1940). A self-administering scale for measuring intellec-
tual impairment and deterioration. Journal of Psychology, 9, 371-377.
Taub, H. A. (1973). Memory span, practice, and aging.Journalof Geron-
tology, 28, 335-338.
Taub, H. A., & Long, M. K. (1972). The effects of practice on short-
term memory of young and old subjects. Journal of Gerontology, 27,
494-499.
Watkins, M. J., & Sechler,E. S. (1989). Adaptingto an irrelevant item
in an immediate recall task. Memory & Cognition, 17, 682-692.
Wingfield, A., Poon, L. W., Lombardi, L., & Lowe, D. (1985). Speed
of processing in normal aging: Effects of speech rate, linguisticstruc-
ture, and processing time. Journal of Gerontology, 40, 579-585.
Wiseman, T. A. (1998). Frequency, intrusions, and suffix effects in im-
mediate recall anddelayed recall as a functionofagingandAlzheimer’s
disease. Unpublished doctoral dissertation, Graduate Center of the
City University of New York.
Zacks, R. T., & Hasher, L. (1994). Directed ignoring:Inhibitory regu-
lation of working memory. In D. Dagenbach & T. H. Carr (Eds.), In-
hibitory processes in attention,memory, and language(pp. 241-264).
San Diego: Academic Press.
REFERENCES
Balota, D. A., & Engle, R. W. (1981). Structural and strategic factors
in the stimulus suffix effect. Journal of Verbal Learning & Verbal Be-
havior, 20, 346-357.
Bloom, L. C., & Watkins, M. J. (1999). Two-component theory of the
suffix effect: Contrary findings. Journalof Experimental Psychology:
Learning, Memory, & Cognition, 25, 1452-1475.
Carlson, M. C., Hasher, L., Connelly, S. L., & Zacks, R. T. (1995).
Aging, distraction, and the benefits of predictable location. Psychol-
ogy & Aging, 10, 427-436.
Connelly,S. L., Hasher, L., & Zacks, R. T. (1991). Age and reading:
The impact of distraction. Psychology & Aging, 6, 533-541.
Cowan, N. (1984). On long and short auditory stores. Psychological
Bulletin, 96, 341-370.
Crowder, R. G. (1971). The sound of vowels and consonants in imme-
diate memory. Journal of Verbal Learning & Verbal Behavior, 10,
587-596.
Crowder, R. G. (1978). Mechanisms of auditory backward masking in
the stimulus suffix effect. Psychological Review, 85, 502-524.
Crowder, R. G., & Morton, J. (1969). Precategorical acoustic storage
(PAS). Perception & Psychophysics, 5, 365-373.
Fabiani, M., & Friedman, D. (1997). Dissociations for temporal order
and recognition memory in aging. Neuropsychologia, 35, 129-141.
Greenberg, S. N., & Engle, R. W. (1983). Voice changes in the stim-
ulus suffix effect: Are the effects structural or strategic? Memory &
Cognition, 11, 551-556.
Greene, R. L. & Crowder, R. G. (1984). Modality and suffix effects
in the absence of auditory stimulation. Journal of Verbal Learning &
Verbal Behavior, 23, 371-382.
Hasher, L., & Zacks, R. T. (1988). Working memory, comprehension,
and aging: A review and a new view. In G. H. Bower (Ed.), The psy-
chologyof learning and motivation:Advances in research and theory
(Vol. 22, pp. 193-225). New York: Academic Press.
NOTE
1. Young adults’ recall scores were relatively high, raising the ques-
tion of whether high- and low-performing youngadults might have been
differentially affected by the presence of the suffix. To address this issue,
we organized young adults in two groups, high- and low-scoring sub-
jects, and assessed the effects of list type, practice, rate, and subgroupon
recall.
Between-subjects condition. The terminal suffix effect was larger for
the low-scoring subjects [F(1,60) = 14.76,MSe = 185.92],butsuch a dif-
ference didnot interact with either practice or rate. In contrast, the preter-
minal suffix effect illustrated a complex pattern of interactions involv-
ing practice, rate, and list type. Specifically, before practice, slowing the
presentation rate did notaffect the preterminal suffix effect in either low-
or high-scoring subjects (Fs < 1). After practice, slowing the presenta-
tion rate increased the preterminal suffix effect in low-scoring subjects
[i.e., it improved recall in the control lists more than in the suffix lists;
F(1,30) = 5.70, MSe = 86.39] and decreased it in high-scoring subjects
[F(1,30) = 13.37, MSe = 31.04].
Blocked condition (seven digits). The terminal and preterminal suffix
effects were larger forlow-scoringsubjects[F(1,48)= 18.04,MSe = 46.04,
and F(1,48) = 7.56, MSe = 25.69, respectively]. However, these differ-
ences did not interact with either rate or practice (Fs < 1).
(Manuscript received May 12, 2000;
Lowe, C., & Rabbitt, P. (1997). Cognitivemodels of ageing and frontal
revision accepted for publication October 11, 2001.)