Full text of DOI:10.3758/BF03195773

Memory & Cognition
2002, 30 (6), 871-884
Prospective and retrospective memory
in normal aging and dementia:
An experimental study
ELIZABETH A. MAYLOR
University of Warwick, Coventry, England
and
GEOFF SMITH, SERGIO DELLA SALA, and ROBERT H. LOGIE
University of Aberdeen, Aberdeen, Scotland
Two experimentsinvestigatedthe effectsof normal aging and dementia on laboratory-basedprospec-
tivememory (PM) tasks. Participantsviewed a film for a laterrecognitionmemory task.In Experiment 1,
they were also required either to say “animal” when an animal appeared in the film (event-based PM
task) or to stop a clock every 3 min (time-based PM task). In both tasks, young participants were more
successful than older participants, who were, in turn, more successful than patients with Alzheimer’s
disease (AD). For successful remembering in the time-based task, older participants and AD patients
checked the clock more often than did young participants. In Experiment 2, participants were asked
to reset a clock either when an animal appeared in the film (unrelated cue–action) or when a clock ap-
peared in the film (related cue–action). Responses were faster in the related condition than in the un-
related condition. Again, there were differences in PM performance between young and older partici-
pants, and between older participants and AD patients. The observed deficits were not due to the
forgetting of the PM task instructions in either experiment. Retrospective memory (RM) tasks (digit
span, sentence span, free recall, and recognition) were more impaired by AD than were the PM tasks.
Factor analysis revealed separate factors corresponding to RM and PM.
Studies of memory in both normal aging and dementia lar, both naturalisticand laboratory studies have addressed
havelargelyfocusedon testsof retrospectivememory (RM). the effects of normal aging on PM, although with some-
These usually involvethe presentationof information that what mixed results (see Maylor, 1993b, 1996b; Maylor,
the participant or patient is subsequently required to rec- Darby, Logie, Della Sala, & Smith, 2002, for reviews).
ognize or recall in response to some prompt from the ex-
A useful framework for considering age deficits across
perimenter. Thus RM decline has been well documented different memory tasks was suggested by Craik (1986).
in both normal aging (e.g., Craik, Anderson, Kerr, & Li, Hisproposalwas thatperformanceisdeterminedbyan inter-
1995; Hultsch, Hertzog, Dixon, & Small, 1998; Zacks, action between external factors such as cues and context
Hasher, & Li, 2000) and dementia (e.g., Morris, 1996; (or, more generally, environmental support) and by the
Nebes, 1992; Spinnler, 1999). In contrast, studies of type of processing operation required by the task. Craik
prospective memory (PM), requiring participants to re- argued thatself-initiatedmental activities,such as retrieval,
member to do somethingin the future withoutany prompt- are particularly difficult in old age, but that age-related
ing such as explicit instructions to recall, have been rela- deficits can be reduced in situations in which environ-
tively neglected. However, because of the importance of mental support is high. In the resulting hierarchy of mem-
the successful functioning of PM to living independently ory tasks, PM tasks were placed at the top since, by defi-
in the community (Cockburn & Smith, 1988; Sinnott, nition, they are low in environmental support and require
1989), there has recently been a growing interest in PM a high degree of self-initiated activity; consequently,they
(Brandimonte, Einstein, & McDaniel, 1996). In particu- should be the most impaired by old age. Below PM tasks
in the hierarchy came RM tasks of free recall, cued recall,
recognition, and priming, in descending order of age
deficits. Thus, the clear prediction from Craik’s analysis
This research was funded by The Wellcome Trust (Grant 047065)and
was that age-related memory impairments should be
greater in PM tasks than in RM tasks.
Since Craik’s (1986) suggestion,numerous studies have
investigated the effects of normal aging on performance
in PM tasks conducted both inside and outside the labora-
tory. In some cases, age deficits in PM performance have
the Medical Research Council (Grant G9606610N). We are grateful to
Sarah MacPherson for conducting the vigilance experiment and to Fer-
gus Craik, ChristopherHertzog, and an anonymousreviewer for helpful
comments on earlier versions of the paper. Correspondence should be
addressed to E. A. Maylor, Department of Psychology, University of
Warwick, Coventry CV4 7AL, England (e-mail: elizabeth.maylor@
warwick.ac.uk).
871
Copyright 2002 Psychonomic Society, Inc.

872
MAYLOR, SMITH, DELLA SALA, AND LOGIE
been observed that are, at least in some respects, more is becausepatients’PM failures are more frustrating to car-
striking than age deficits in RM performance (e.g., Cock- ers (although no more prevalent) than are their RM fail-
burn & Smith, 1991; Mäntylä & Nilsson, 1997; Maylor, ures.
1993a,1996a). However, in othercases, equivalentPM per-
To ourknowledge,theonlyexperimentalstudyto attempt
formance in young and older adults has been found in the a comparison between PM and RM deficits in dementia
context of age-related decline in RM performance (e.g., was by Huppert and Beardsall (1993), although, as noted
Einstein & McDaniel, 1990; Maylor, 1990). Age deficits below, whether their PM task was a true test of PM has
are therefore not always greater in PM tasks than in RM been questioned. They focused in particular on a specific
tasks, contrary to Craik’s suggestion.
item on the Rivermead Behavioral Memory Test (Wilson,
These mixed findingsin the literature may be partly ex- Cockburn,& Baddeley, 1985)—namely, “remembering to
plainedby the fact that PM tasks themselvesvary as much delivera message.” The procedureis as follows. The tester
as RM tasks in terms of external factors and processing asks the participant to “watch what I do, and when I have
demands(seeKvavilashvili,1992).In otherwords, itisprob- finished, do the same thing.” The tester then traces a short
ably not the case that all PM tasks are low in environmen- path around the room, describing each action en route: “I
tal support and high in self-initiatedactivity. For example, am going to start from this chair, and take this envelope
age deficits are often absent in PM tasks conducted out- with me. From here I am going over to the door. And from
side the laboratory, where participants are able to set up the doorto the window. From the window to the table. I am
appropriate external cues (see Maylor, 1993b, for a sum- going to leave this envelope on the table, and from here I
mary), and age deficits in the laboratory are greater when am going back to the chair.”
processingdemands increase (e.g., Einstein,Holland, Mc-
Huppert and Beardsall (1993) scored performance sep-
Daniel, & Guynn, 1992; Einstein, McDaniel, Manzi, arately for remembering to pick up the envelope at the
Cochran, & Baker, 2000; Einstein, Smith, McDaniel, & start and place it down at the end (defined as PM), and re-
Shaw, 1997; Kidder, Park, Hertzog, & Morrell, 1997; membering each stage of the route in the correct order
Maylor et al., 2002). Craik’s (1986) framework therefore (defined as RM). They found that for the PM component,
remains useful in understanding different effects of nor- those with minimal dementia were similar to those with
mal aging across PM studies.
mild or moderate dementia, whereas for the RM compo-
One way in which PM tasks differ is in terms of whether nent, those with minimal dementia were more similar to
the trigger for action is a particular event or a particular the normal controls. Huppert and Beardsall therefore con-
time (event based vs. time based, respectively;Einstein & cluded that PM impairment may be an earlier indicator of
McDaniel, 1990). For example, compare remembering to dementia than RM impairment. However, other interpre-
make a telephonecall when you next see a telephonewith tationsare possible (see Maylor, 1995, for discussion).For
remembering to make a telephone call at 1600 hours. A example, it could be argued that the PM component does
number of laboratory studies have found smaller age not correspond with the usual definition of remembering
deficits in event-based PM tasks than in time-based PM in the future withoutbeing prompted because participants
tasks (e.g., Einstein, McDaniel, Richardson, Guynn, & were specificallyasked by thetester to repeat the sequence
Cunfer, 1995; Park, Hertzog, Kidder, Morrell, & May- of actions just observed. Thus, there seems to be no logi-
horn, 1997; but see also d’Ydewalle, Luwel, & Brunfaut, cal basis for defining certain elements of the sequence as
1999). These findings can be readily interpreted within prospective and others as retrospective. Instead the task
Craik’s (1986) framework by supposing that age deficits couldbe describedas a RM test of serial recall for actions.
are less apparent in event-based tasks because the target The fact that some elements in the sequence are more
event itself provides a strong external cue to support per- likely to be omitted than others in the early stages of de-
formance, and therefore self-initiated retrieval is unnec- mentia could reflect differences in difficulty across the
essary. In time-basedtasks, however,environmentalsupport various elements. Clearly, further empirical investigations
is low and self-initiated retrieval is high because partici- are required to address Huppert and Beardsall’s sugges-
pants have to remember to monitor the time effectively tion that PM tasks are particularly sensitive to the early
(for evidence that older adults check the clock less often stages of dementia.
than young adults in time-based PM tasks, see Einstein
et al., 1995; Hohaus, Shum, White, & Caird, 2000; Park of both normal aging and dementia across a range of PM
et al., 1997). and RM tasks. Young adults, older adults, and patients di-
The aim of the present study was to examine the effects
In contrast to the extensive literature on normal aging agnosedas possiblyorprobablysuffering from Alzheimer’s
and PM, there have been surprisingly few studies of PM disease (AD) were compared in two experiments. In Ex-
in dementia(see Mayloret al., 2002, for discussion of pos- periment 1, we were concerned with the distinction be-
sible reasons for this). Anecdotal evidence suggests that tween event- and time-based PM tasks. Participants were
people initially consult their doctors because of their rel- asked to view a short film for a subsequent memory test
atives’ PM rather than RM problems (e.g., Camp, Foss, and also to indicateeither whenever an animal appeared in
Stevens, & O’Hanlon, 1996; McKitrick, Camp, & Black, the film (event-based PM task) or whenever 3 min had
1992). From a recent questionnaire study by Smith, Della elapsed (time-based PM task). In the latter case, a clock
Sala, Logie, and Maylor (2000), it would appear that this was positioned out of view (to one side) of the participant

PROSPECTIVE AND RETROSPECTIVE MEMORY
873
so that a head movement was required to see the time. that after a task involvingan envelope, they were to turn it
Thus, we were able to record clock-checking behavior over, seal it, and write their initials on the back. Thus pa-
(see Ceci, Baker, & Bronfenbrenner, 1988; Ceci & Bron- tients may perform poorly in a PM task, not necessarily
fenbrenner, 1985; Einstein et al., 1995; Harris & Wilkins, because of impaired PM, but because of impaired RM. In
1982; Hohaus et al., 2000; Park et al., 1997). Our main the present study, therefore, we questioned participants at
prediction with respect to normal aging was that deficits the end of testing to ensure that both patients and controls
would be less apparent in the event-based PM task than in had successfully retained the PM task instructions.
the time-based PM task because of the greater environ-
The present PM tasks were designed to provide reason-
mental support provided by the target event occurring in ably naturalistic simulations of everyday PM tasks such
the film relativeto the (not directly visible)clock showing as remembering to look out for a post box while shopping
that 3 min had elapsed. In general, the RM literature sug- (see our event-based task in Experiment 1) or remember-
gests that there are similarities between the effects of nor- ing to return to the kitchen after 3 min when boiling an
mal and abnormal aging (e.g., priming is less impaired by egg (see our time-based task). To anticipate,Experiment1
both normal aging and dementia than is free recall; see was more successful than Experiment 2 in producing sig-
Prull, Gabrieli, & Bunge, 2000). We therefore expected a nificant effects of the various task manipulationsin terms
greater deficit due to dementia in the time-based PM task of the numbers of correct PM responses. Nevertheless, the
than in the event-based PM task.
two experiments together provided the opportunity to
The comparison in Experiment 2 was between two combine the data to produce sufficiently large numbers of
event-based PM tasks in which the relationship between participants to test Huppert and Beardsall’s (1993) pro-
the PM cue event and the corresponding action was ma- posal that PM is more prone to decline than RM in the
nipulated. Thus participants reset a clock in response ei- early stages of dementia. Thus we included a number of
ther to the appearance of a clock in the film (“see a clock– RM tasks in both experiments for comparison—namely,
stop the clock”; related cue–action) or to the appearance digit span, sentence span, free recall, and recognition. At
of an animal (“see an animal–stop the clock”; unrelated least some of these RM tasks were approximately compa-
cue–action). The related conditionwas assumed to be less rable with the PM tasks in terms of retention interval.
demanding than the unrelated condition in terms of re- From Craik’s (1986) framework, they were expected to
trieval processes and environmental support (see Craik, produce a range of aging and dementia deficits, with free
1986). We therefore predicted that the effects of both nor- recall as the most impaired. Note that since PM tasks also
mal aging and dementia would be smaller in the related differ in many ways (see Kvavilashvili, 1992), our aim
than in the unrelated condition.
was not to compare PM and RM in general (as if they were
In both experiments,we recorded not only whether par- single entities); rather, we were interested in whether at
ticipantssuccessfully responded to the PM cues for action least some PM tasks are less prone to decline in dementia
when
but also
they responded. As West and Craik (1999) than at least some RM tasks.
Finally, the issue of the relationship between PM and
have recentlyshown for event-based tasks, response times
to PM cues can provide sensitive measures of the accessi- RM has been much debated in the literature (e.g., see
bility or level of activation of the PM task. For time-based Brandimonte et al., 1996). In addition to addressing this
tasks, precise response times indicate how accurately par- issue by examining whether or not either normal aging or
ticipants are able to monitor the time and how well they dementia have differential effects on PM and RM, the
are able to keep the PM task in mind during an ongoing present study also provided the opportunityto explore the
task. In providing more sensitive assessments of PM per- PM–RM relationshipthroughfactor analysison measures
formance, response times may be particularly useful in from several different memory tests in the search for dis-
cases where some groups are close to ceiling in terms of tinct PM and RM factors.
the number of correct PM responses.
As arguedelsewhere(Maylor,1995,1996b;Mayloret al.,
2002), in studyingPM in normal aging and dementia, it is
particularly important to separate PM from RM failures.
Thus, in a PM task, participantsare required to remember
not only that something has to be done (PM component)
but also what has to be done, and when and where (RM
component). Studies of PM have consequently attempted
to keep the RM componentas simple as possibleto ensure
that performance is determined by PM rather than by RM
(e.g., Maylor, 1998). However, retaining the PM task re-
quirements may not be such a trivial task for dementia pa-
tients. Indeed, Huppert, Johnson, and Nickson (2000) ob-
served that 60% of a group of individuals with probable
dementia (mean age of 81.5 years) failed to show intact
EXPERIMENT 1
Method
Participants. There were three groups of participants: young
adults, older adults, and patients with AD. The young participants
(n 5 30) were aged between 19 and 40 years and were undergradu-
ate students at the University of Aberdeen. The older participants
(n 5 30) were aged between 58 and 75 years and were members of
a volunteer panel at the Department of Psychology, University of
Aberdeen. Both the young and older participants were paid for tak-
ing part in the study. AD patients were recruited through the Mem-
ory Clinic of the Mental Health Department at the University of Ab-
erdeen. Patients were not included in the present study if they scored
below 16 on the Mini-Mental State Examination (MMSE; Folstein,
Folstein, & McHugh, 1975). The patients were diagnosed as suffer-
ing from progressive cognitive deterioration due to possible or prob-
RM for a PM task that required participants to remember able AD on the basis of formal criteria (Spinnler & Della Sala,

874
MAYLOR, SMITH, DELLA SALA, AND LOGIE
1988), which broadly correspond to those proposed by NINCDS-
Apparatus and Stimuli. A 17-min silent color film of scenes
ADRDA (McKhann et al., 1984). All the patients fulfilled the crite- and landmarks in and around the city of Aberdeen was recorded
ria for inclusion in research studies as set forth by the Medical Re- using a hand-held camcorder. During the film, five different animals
search Council’s Alzheimer Disease Workshop Steering Committee (a cow, a horse, a dog, a pig, and a sheep) appeared after approxi-
(Wilcock et al., 1989). Two patients were excluded because they mately 3, 6, 9, 12, and 15 min, respectively. Each animal was on the
were unable to understand the task instructions. The remaining pa- screen for exactly 3 sec. The film was prepared so that the animal
tients (n 5 24) were aged between 53 and 84 years and had been at- scenes blended smoothly with the preceding and subsequent scenes.
tending the Memory Clinic for between 7 and 66 months (M 5 30.2, The film was recorded onto a videotape and played back on a color
SD 5 18.6). Their MMSE scores ranged from 16 to 27 with a mean television with a screen diagonal of 711 mm. Still color photographs
of 22.1 (SD 5 3.6).
of 12 scenes from the film and 12 scenes of other Aberdeen loca-
Background details for the three groups of participants are pro- tions not included in the film were produced for a 24-item two-
vided in Table 1. There was no significant age difference between choice visual recognition test. Responses in the event-based PM task
the older adults and the AD patients [t(52) 5 20.75, p . .10]. An were timed using a handheld stopwatch.
analysis of variance (ANOVA) on years of education revealed a sig-
For the time-based PM task, an analogue clock display was pro-
nificant effect of group [F(2,81) 5 9.98, MSe 5 9.19, p < .001]. Post duced on a black-and-white computer screen using Visual Basic.
hoc Newman–Keuls comparisons showed a marginally significant dif- The diameter of the clock was 145 mm. There were lines of 7 mm
ference between the young and older participants (p 5 .06), but a extending inward from the perimeter of the clock at 0º, 90º, 180º, and
significant difference between the older participants and the AD pa- 270º, and these were labeled 0, 1, 2, and 3, respectively. Each of the
tients (p < .01).
numbers was 10 mm high. There were second (length 5 145 mm)
Three measures of RM are also included in Table 1. Forward digit and minute (length 5 60 mm) hands on the clock. The second hand
span was assessed by presenting lists of digits verbally at a rate of rotated 360º per minute whereas the minute hand rotated 90º per
one digit per second. Participants were required to recall aloud each minute. Pressing the space bar on the computer keyboard reset both
list in the correct order. After three trials with two digits, the num- hands of the clock to zero. The time when the space bar was pressed
ber of digits increased to three, and so on up to a possible maximum was recorded by the computer program. The computer screen and
of nine digits. Testing was terminated when there were errors on two keyboard were placed so that the left-hand edge of the screen was at
of the three trials with a particular list length. Span was calculated
by averaging the lengths of the longest three lists accurately recalled.
90º to the right of participants. This ensured that the clock itself was
outside of participants’ peripheral vision as they viewed the film, so
For sentence span, the experimenter read aloud sets of sentences of they would have to turn around to look at the clock. A stopwatch
4–6 words and then asked participants to recall the last word of each was used by the experimenter to record when participants monitored
sentence in the correct order. After three trials with one sentence,
the number of sentences increased to two, and so on up to a possible
the clock.
Design and Procedure. Participants were tested individually ei-
maximum of six sentences. Testing was terminated when there were ther in the Department of Psychology or in the Mental Health De-
errors on two of the three trials with a particular number of sen- partment, University of Aberdeen. Half of the participants in each of
tences. Span was calculated by averaging the lengths of the longest
the three groups were assigned to the event-based PM condition, and
three sets accurately recalled. Immediate free recall was based on a half were assigned to the time-based PM condition. In all three
procedure by Capitani, Della Sala, Logie, and Spinnler (1992). The groups, the participants assigned to the event-based condition were
experimenter read out five lists of 12 unrelated words at a rate of extremely well matched with the participants assigned to the time-
one word per second. Immediately after each list presentation, par- based condition in terms of the background variables shown in
ticipants were required to recall aloud the words in any order. The Table 1, with just one exception: The AD patients assigned to the
total number of words recalled out of 60 was recorded.
ANOVAs were conducted to compare the RM measures across those assigned to the time-based condition. AD patients assigned to
the three groups of participants. For digit span, there was a signifi- the event- and time-based conditions had similar MMSE scores and
cant effect of group [F(2,81) 5 13.15, MSe 5 1.71, p , .001]. There had been attending the Memory Clinic for similar lengths of time.
was also a significant effect of group for sentence span [F(2,81) 5 1. In the event-based condition, AD patients were first shown six
34.32, MSe 5 0.61, p , .001]. For both digit span and sentence span, pictures. Three of these showed animals that did not appear in the
post hoc Newman–Keuls comparisons revealed significant differ- film, and three showed items of furniture. The patients were asked
event-based condition achieved lower scores for free recall than did
ences between the young adults and the AD patients, and between to say whether each picture showed an animal or not; this prelimi-
the older adults and the AD patients, but not between the young and nary test was to ensure that any errors in responding to animals in
older adults. Finally, there was a significant effect of group for free the PM task were not due to problems other than memory deficits,
recall [F(2,81) 5 46.57, MSe 5 35.87, p , .001]. In this case, post such as severe visual agnosia. All the patients successfully distin-
hoc Newman–Keuls comparisons revealed significant differences
between all three groups.
guished between the pictures of animals and furniture.
Participants were told that they were going to view a film. They
were encouraged to watch it carefully because they would be asked
questions later about its content. They were also instructed to look
out for animals in the film and respond by saying “animal” as
quickly as possible after seeing one. It was stressed that they should
not respond to people or birds. When participants had demonstrated
full understanding of the instructions, the experimenter started the
film. The delay between the appearance of each animal and the par-
ticipant’s verbal response was recorded by the experimenter, who
was seated out of the participant’s view.
2. In the time-based condition, the experimenter began by starting
the computerized clock and explaining how it displayed the passage
of time. Participants were instructed that they would be required to
stop and reset the clock by pressing the space bar on the computer
keyboard whenever 3 min had elapsed. The experimenter himself
pressed the space bar after 3 min to demonstrate how the clock was
Table 1
Background Data for Participants in Experiment 1
Young*
Older†
SD
AD Patients‡
Measure
M
SD
M
M
SD
Age (years)
Education (years) 13.83
Digit span
Sentence span
Free recall
25.40
4.96
3.11
1.22
0.79
7.53
67.27
12.28
7.08
4.39
25.93
4.24 68.54
3.56 10.13
8.01
2.05
1.34
0.97
5.86
6.82
4.54
35.07
1.37
0.56
5.35
2.92
4.03 19.46
Note—F, females; M, males; AD, Alzheimer’s disease. *n 5 30, 9 F,
21 M. †n 5 30, 18 F, 12 M. ‡n 5 24, 14 F, 10 M.

PROSPECTIVE AND RETROSPECTIVE MEMORY
875
were no responses longer than 4 sec.) The delay between
the onset of the animal’s appearance and the participant’s
response was recorded to the nearest second.2 In the time-
based condition, a successful PM response was recorded
if the participant stopped the clock between 2 sec before
and 2 sec after the clock had reached 3 min.3 The absolute
time difference between the participant’s response and the
target time was recorded to the nearest second.
PM performance. Figure 1 displays the mean num-
bers of PM successes out of five in the event- and time-
based conditions for young adults, older adults, and AD
patients.4 It can be seen that the young adults performed
at ceilingin bothconditions.Performance for the two older
groups was higher for the event-based task than for the
time-based task. The young group performed better than
the older group and the older group performed better than
the AD patients, with the difference between the two older
groups being similar for the event-based and time-based
tasks.
Because the young participants performed perfectly in
both conditions,they were not included in a two-way fac-
torial ANOVA on these data with group (2 levels: older
and AD patients)and task (2 levels:event-basedand time-
based) as factors. There were significant effects of both
Figure 1. Mean numbers of PM successes out of five in the event-
based and time-based conditions for young adults, older adults,
and AD patients in Experiment 1. Error bars represent 1 SE.
F
5
MS 5
p ,
group [ (1,50) 18.72,
MS 5
1.33,
p ,
1.33,
.001] and task
.001], but there was
F ,
e
then automatically reset to zero. Participants were instructed to be as
accurate as possible and that an early response was as erroneous as
a late response.1 They were told that they could check the clock as
often as they wished. The number and timings of the participants’
clock checks were noted by the experimenter.
F
5
[ (1,50) 13.26,
no significant interaction between them (
e
1).
The upper portion of Table 2 shows the mean latencies
for successful PM responses in the event-based condition
and the mean absolute distances away from the target
times for successful PM responses in the time-based con-
As in the event-based condition, participants were told that they
were going to view a film that they should watch carefully because
they would be asked questions later about its content. Participants
were then reminded of all the task instructions. When they had
demonstrated that they understood all the instructions, the experi-
menter started the film and the clock simultaneously by pressing a
single button connected to the video player and the computer.
At the end of the film in both the event- and time-based condi-
tions, participants were asked to recall what they had been instructed
to do while watching the film. If they did not spontaneously recall
the PM task instructions, they were prompted to think whether there
was anything else they had to do. If they failed to recall the PM task
instructions with this prompt, they were reminded of what they were
supposed to do and were asked if they recognized the instructions.
Participants in both conditions were then shown 24 photographs
one at a time. They were told that some of the scenes appeared in the
film whereas others did not. Their task was to respond “yes” or “no”
to each photograph depending on whether or not they recognized the
scene as having appeared in the film.
6
dition (i.e., those that fell within the 2-sec time win-
dow). Because these response times are not strictly com-
parable across the two conditions (e.g., possible response
times for the event- and time-based conditions were
0–4 sec and 0–2 sec, respectively), separate ANOVAs
were conducted on these data. For the event-based condi-
F
5
tion, there was a significant effect of group [ (2,39)
MS 5
p ,
47.80,
0.23,
.001]. Newman–Keuls compar-
isons reveealed that the young adults were significantly
p ,
faster than the older adults(
were marginally faster than the AD patients(
the time-based condition,the effect of group was also sig-
.001), and the olderadults
p 5
.08). For
F
5
MS 5
p ,
0.22, .001]. On
nificant [ (2,36) 24.14,
e
Finally, the digit span, sentence span, and free recall tests were
administered to all participants (Table 1).
Table 2
Response Time Data and Recognition
Memory Performance in Experiment 1
Results
Young
Older
AD Patients
Recall of PM instructions. When asked at the end of
the experiment to recall what they had been instructed to
Measure
M
SD
M
SD
M
SD
do while watchingthe film, all participantsspontaneously Response times (sec)
Event based
0.2
0.3
0.3
0.4
1.5
0.5
0.3
0.5
1.8
1.6
0.8
0.6
recalled the PM instructions correctly with the exception
of 3 AD patients who required only a minimal prompt.
PM scoring. In the event-based condition, a success-
ful PM response was recorded if the participantcalled out
“animal” within 4 sec of the appearance of an animal in
Time based*
Recognition memory (% correct)
Event based
Time based
93.04 8.71 90.28 8.72 68.06 5.98
92.21 9.42 91.96 6.21 71.54 9.37
Note—AD, Alzheimer’s disease. *One older adult and 2 AD patients
the film (which was on screen for 3 sec). (In fact, there missing because of zero PM successes.

876
MAYLOR, SMITH, DELLA SALA, AND LOGIE
Newman–Keuls comparisons, the young adults were not sponses and clock-checking behavior leading up to PM
significantlycloser to the target times than the olderadults failures. In theformer case, 1 older adultand 2 AD patients
p .
(
.10), but the older adults were significantly closer had to be omitted from the analysis because they never
p ,
succeeded; in the latter case, all 15 young adults, 5 older
than the AD patients (
Analysis of the distances away from the target times adults, and 1 AD patient had to be omitted because they
all
.001).
(i.e., accuracy) for
PM responses in the time-based never failed. We initially analyzed the data separately for
condition (i.e., regardless of whether or not they fell suc- each of the five 3-min intervals. There were no obvious
6
cessfully within the 2-sec time window) showed that trends over the course of the experiment, so the data for
M 5 SD 5
AD patients (
curate than older adults (
who were, in turn, less accurate than young adults (
SD 5
10.0 sec,
16.5 sec) were less ac- each participant were averaged across the available trials
M 5
SD 5
6.6 sec,
16.6 sec), (maximum of five).
M 5
0.4 sec). Note also that AD patients made the six 30-sec periods leading up to PM success for each
Figure 2 shows the mean numbers of clock checks in
0.3 sec,
significantly fewer responses (whether accurate or inac- group of participants. Clearly, clock checking was most
curate) in the time-based condition than did the older frequent (and very similar across the three groups) in the
M
5
M
5
adults ( for AD patients 4.0/5; for older adults
final period of the 3-min time interval culminating in a
event-and PM response. Clockcheckingin the precedingfive periods
t
5 2 p , both
4.8/5) [ (25)
3.20,
.005]. Thus, in
time-based conditions, the AD patients missed more PM was much less frequent, particularly for young adults. An
targets entirely than the older group and also tended to be ANOVA was conducted on these data with group (3 lev-
slower (event-based)and less accurate (time-based) when els) and period (6 levels) as between- and within-subjects
they did respond. In comparison with the young adults, factors, respectively. There were significant effects of
F
5
MS 5
p ,
the older adultsmissed more targets and were slower when group [ (2,36) 8.27,
MS 5
0.87,
p ,
0.27,
F
.001] and period
.001], but there was
e
F
5
they did respond in the event-based task, but they pro- [ (5,180) 217.05,
duced almost as many responses in the time-based task, no interaction between them [ (10,180) 1.19,
e
5
MS 5
e
p .
albeit with much less accuracy.
0.27,
.10]. Post hoc Newman–Keuls comparisons in-
Clock-checking behavior. For the time-based condi- dicated that young adults checked the clock significantly
p ,
tion, we calculated the number of times participants less often than both older adults (
.05) and AD pa-
.001), and that older adults checked the clock
p ,
p ,
checked the clock in each of the six 30-sec periods over tients (
each 3-min intervalleadingup to a PM target time. In con- marginally less often than AD patients (
trast to previous studies (Einstein et al., 1995; Hohaus In contrast to the J-shaped functions of Figure 2, clock
.08).
et al., 2000; Park et al., 1997), we looked separately at checkingin the periods leading up to PM failure remained
clock-checking behavior leading up to successful PM re- flat for both older adults and AD patients, with the mean
Figure 2. Mean numbers of clock checks leading up to a successful PM re-
sponse in each 30-sec period of the 3-min intervals in the time-based condition
of Experiment 1. Error bars represent 61 SE.

PROSPECTIVE AND RETROSPECTIVE MEMORY
877
number of clock checks in each period never exceeding1. sign of an interaction between age and type of PM task.
These data were analyzed with group (2 levels) and period Thus it appears that our particular event- and time-based
(6 levels) as between- and within-subjectsfactors, respec- tasks do not produce differential age deficits.
F
5
tively. There was a significant effect of group [ (1,19)
In contrast to decline in PM performance with normal
.05], indicating that older adults aging, there was no age-related deficit in performance on
0.91) checked the clock more often than AD pa- the ongoing task of remembering the film for a later
M 5
MS 5 p ,
5.63,
M 5
0.79,
e
(
tients (
0.54). Note that this is the opposite result to recognition task. Note that age deficits in PM tasks have
the trend observed for periods leading up to PM success, now been observed in the context of similar, superior, and
where AD patients checked the clock more often than inferior performance on the ongoing task by older adults
older adults (Figure 2). The effect of period was not sig- in comparisonwithyoungadults(e.g., Mayloret al., 2002).
F
5
MS 5
p .
.10], and there
nificant [ (5,95) 1.05,
0.30,
Third, event-and time-basedPM performancesin terms
e
F ,
was no interactionbetween group and period (
1). Al- of numbers of successful responses were equallyimpaired
though older adults actually checked the clock slightly in AD patients in comparison with age-matched controls
more often in Periods 1–5 prior to a PM failure than prior (Figure 1). AD patients also tended to be slower (event
to a PM success, it should be remembered that different based) and were less accurate (time based) than older
subsets of older adults were included in the two analyses. adults. Our prediction was that the time-based PM task
Recognition memory performance. The lower por- would be more impaired by AD than the event-based PM
tion of Table 2 summarizes performance in the recogni- task because (1) time-based tasks are more adversely af-
tion memory task after film viewing. It can be seen that fected by normal agingthan event-based tasks due to their
the mean percentages of correct recognition responses greater reliance on self-initiated retrieval processes, and
(i.e., “yes” to old scenes from the film and “no” to new (2) memory tasks that are particularlyimpaired by normal
scenes) were higher for the young and older adults than aging are often also particularly impaired by dementia.
for the AD patients. (Both omissions and false positives Logie et al.’s (2001) results would suggest that at least the
were numerically greater for the AD patients than for the first of these two assumptions does not always hold true
other groups.) The percentages of correct responses were (see also d’Ydewalle et al., 1999, for evidence of larger
analyzed with group (3 levels) and PM task (2 levels) as age deficits in event- than in time-based tasks). It seems
between-subjects factors. There was a significant effect more plausible that both event- and time-based PM tasks
F
5
MS 5
p ,
of group [ (2,78) 62.10,
67.53,
.001], but can vary (e.g., in terms of the degree of self-initiated ac-
e
F ,
s
no effect of PM task and no interaction (both
1). tivityrequired or environmentalsupport available)in such
Post hoc Newman–Keuls comparisons revealed that the a way that some event-based tasks could be more de-
AD patientsdiffered from both the young and olderadults manding than some time-based tasks. In other words, a
(who did not differ).
simple distinction between event- and time-based PM
tasks in terms of different cues for action (a particular
event or time, respectively) may be less useful than origi-
Discussion
The first point to emphasize from the results of Exper- nally supposed in predicting both aging and dementia
iment 1 is that deficits due to both normal aging and de- deficits (see also Rabbitt, 1996). Nevertheless, it is of in-
mentia were observed in PM performance, despite having terest that the present event- and time-based PM tasks are
both
ensured that all participants had successfully retained the similarly impaired by
PM task instructions. In other words, everyone knew and 2001) and dementia (present data).
normal aging (see Logie et al.,
understoodwhat they were required to do, but olderadults,
Finally, there was greater clock checking in the final
and particularlyAD patients,simply failed to carry out the 30-sec period leading up to a target time than in the pre-
PM task at the appropriate time. We are therefore dealing vious five periods in the time-based PM task (Figure 2),
here purely with failures of the PM component (see May- as others have observed (e.g., Einstein et al., 1995; Park
lor, 1998).
et al., 1997). However, if clock-checkingbehaviorprior to
Second, performance in bothevent-and time-based PM PM success and failure is considered separately,the final-
tasks was impaired by normal aging in such a way that period increase was evidentonlyprior to PM success. This
older adults made fewer successful responses than young finding is important because it helps to explain why pre-
adults, and were slower (event based) or less accurate vious studies of aging (which have not distinguished be-
(time based) in executing responses generally. The age- tween clock checking prior to PM success and failure)
related deficit shown in Figure 1 appeared greater for the have found that older adults check the clock less often
time-based task than for the event-based task, consistent than young adults, particularly in the final period before
with the majority of previous studies, but here, such a the target time (Einstein et al., 1995; Hohaus et al., 2000;
comparison is obviously difficult to interpret because of Park et al., 1997). Since older adults are more likely than
the presence of ceiling effects for young adults. However, young adults to forget the PM task, exactly this pattern
in another study using identical tasks (Logie, Maylor, would be expected.In contrast, the present results showed
Della Sala, & Smith, 2001), PM performance was taken no effects of aging or dementia on clock-checkingbehav-
down from ceiling by requiring participantsto perform an ior in the final 30-sec period leading up to PM success, but
additional arithmetic verification task. PM success was there were significantdifferencesbetween groupsin clock-
greater for young than for older adults, but there was no checking in the earlier periods such that AD patients and

878
MAYLOR, SMITH, DELLA SALA, AND LOGIE
Table 3
were aged between 55 and 80 years. Both of these groups were paid
for their participation in the study. AD patients who were able to un-
derstand the task instructions (n 5 18) fulfilled the same selection
criteria as given for Experiment 1. They were aged between 57 and
81 years and had been attending the Memory Clinic for between 8
and 65 months (M 5 33.2, SD 5 18.1). Their MMSE scores ranged
from 17 to 30 (M 5 20.9, SD 5 3.8).
Table 3 summarizes background information on the three groups
of participants. The older adults and AD patients did not differ sig-
nificantly in age [t(36) 5 20.29, p . .10]. There was a significant
effect of group in an ANOVA on years of education [F(2,55) 5 14.58,
MSe 5 6.99, p , .001], with significant differences on Newman–
Keuls comparisons between the young and older adults ( p , .001)
but not between the older adults and the AD patients ( p . .10).
For the RM measures shown in Table 3, there were significant
overall effects of group in ANOVAs for digit span [F(2,55) 5 16.76,
MSe 5 1.01, p , .001], for sentence span [F(2, 55) 5 15.44, MSe 5
1.07, p , .001], and for free recall [F(2,55) 5 24.60, MSe 5 28.76,
p , .001]. As in Experiment 1, Newman–Keuls comparisons re-
vealed significant differences between all three groups with the ex-
ception of young versus older adults for digit span and sentence
span.
Apparatus and Stimuli. Two 17-min silent color films were
used in this experiment. One was the film used in Experiment 1, in
which five different animals appeared at approximately 3-min inter-
vals, each for 3 sec (the “animal film”). The other was very similar
to the original film in style and content, the main difference being
that five different clocks appeared after approximately 3, 6, 9, 12,
and 15 min. Each clock was on the screen for exactly 3 sec. As be-
fore, the clock scenes blended smoothly with the preceding and sub-
sequent scenes. Because this second film differed slightly from the
first, a second set of 24 color photographs (12 scenes from the film
and 12 new scenes) was prepared for the two-choice visual recogni-
tion test following the viewing of the “clock film.”
Background Data for Participants in Experiment 2
Young*
SD
Older†
SD
AD Patients‡
SD
Measure
M
M
M
Age (years)
Education (years)
Digit span
Sentence span
Free recall
21.15 2.91
15.05 1.79
7.08 1.19
4.62 0.99
31.30 6.51
68.05 7.98
11.95 3.12
6.75 0.96
4.23 0.95
26.80 5.37
68.72 5.85
10.56 2.85
5.30 0.79
2.83 1.16
19.17 3.67
Note—F, females; M, males; AD, Alzheimer’s disease. *n 5 20, 11 F,
9 M. †n 5 20, 14 F, 6 M. ‡n 5 18, 13 F, 5 M.
more
older adults actually checked
adults (Figure 2).
often than young
This finding resembles two previous results in the lit-
erature. First, Park et al. (1997) observed that youngadults
checked the clock more frequently under dual-task condi-
tions than undersingle-taskconditions.The more frequent
clock checkingfound in our older participantsis therefore
consistent with the view that one impact of normal aging
is a reduction in processing resources (see Salthouse,
1991, for a summary, but see Baddeley, Logie, Bressi,
Della Sala, & Spinnler, 1986, and Duff & Logie, 2001, for
caveats). Second, Maylor (1998) asked participants at the
end of an event-based PM experiment to rate how often
they had thoughtaboutthe PM componentduringthe task.
Regression analysis relating these self-rated thoughts to
PM performance revealed that for the same level of PM
performance, older participants reported thinking about
the PM task more than young participants.In otherwords,
The analogue clock display used in Experiment 1 was also used
for both event-based (Maylor, 1998) and time-based in Experiment 2. The film and the computer clock were synchro-
nized so that PM response times were recorded automatically by the
(present study) tasks, successful PM performance was
computer. As before, the computer screen and keyboard were placed
preceded by different behaviorsin youngand olderadults.
to the right of the participant so that the clock could not be seen dur-
ing film viewing.
Design and Procedure. Participants were tested individually ei-
Thus PM success in old age seems to be achieved by
greater effort during the task, either in terms of more
thoughts about the task (event based) or more checks on
ther in the Department of Psychology or in the Mental Health De-
partment, University of Aberdeen. Half of the participants in each of
the three groups were assigned to the related cue–action condition,
and half were assigned to the unrelated cue–action condition. In all
three groups, the participants assigned to the related condition were
extremely well matched with the participants assigned to the un-
related condition in terms of the background variables shown in
Table 3, with just one exception: Participants assigned to the related
condition achieved slightly lower scores on the digit span task than
those assigned to the unrelated condition, but this was equally the
case for all three groups. AD patients assigned to the two conditions
had similar MMSE scores and had been attending the Memory
Clinic for similar lengths of time.
the clock (time based).5
It is interesting to note that prior to a PM failure, older
adults and AD patientscheckedthe clock once on average
during the final 30-sec period. This observation is consis-
tent with the view that age-related deficits in PM are not
due to the completeforgettingof the PM task demands but
instead reflect temporary or momentary lapses of inten-
tion (see Craik & Kerr, 1996; Maylor, 1993a,1996a; West
& Craik, 1999).
Comparisons between the effects of aging and demen-
tia on RM versus PM, and factor analysis, are postponed
until after the presentation of Experiment 2, in which we
examined performance on two event-based PM tasks dif-
fering in terms of the relationship between the PM cue
event and the action required in response to the cue.
As in Experiment 1, the experimenter began by checking that all
the AD patients in the unrelated cue–action condition were able to
recognize animals (see Experiment 1 for procedure). All could dis-
tinguish perfectly between pictures of animals and pictures of furni-
ture.
Participants were instructed that they were going to view a film
and that they should watch it carefully because they would be asked
questions later about its content. They were also instructed to look
out for clocks (related cue–action condition) or animals (unrelated
cue–action condition). Participants in the related condition were
shown the clock film, whereas participants in the unrelated condi-
tion were shown the animal film. In response to the appearance of a
EXPERIMENT 2
Method
Participants. Three new groups of participants (young adults,
older adults, and AD patients) were recruited from the same sources
as described for Experiment 1. The young participants (n 5 20) were specified target event, all participants were instructed to stop the
aged between 18 and 29 years and the older participants (n 5 20)
clock by pressing the space bar of the computer keyboard as soon as

PROSPECTIVE AND RETROSPECTIVE MEMORY
879
possible to reset the hands of the clock to zero. (Note that target
Table 4
Response Time Data and Recognition
Memory Performance in Experiment 2
events did not occur at exactly 3-min intervals, so there was no sug-
gestion that the time on the clock was informative as to when the
next target would appear.) Participants in the unrelated cue–action
condition were additionally warned not to respond to people or birds.
The delay between the appearance of each target event and the press-
ing of the space bar was recorded by the computer. When partici-
pants had demonstrated that they understood all the task instruc-
tions, the experimenter started the film and clock simultaneously
by pressing a single button connected to the video player and the
computer.
In both conditions, participants were asked at the end of the film
to recall what they had been instructed to do while watching the film
(see Experiment 1 for details). The two-choice recognition test was
then completed, followed by the digit span, sentence span, and free
recall tests.
Young
Older
AD Patients
Measure
M
SD
M
SD
M
SD
Response times (sec)
Related
Unrelated
0.2
1.2
0.4
1.0
1.3
1.7
1.0
1.3
1.2
1.9
0.6
0.7
Recognition memory (% correct)
Related
Unrelated
84.17 12.39 79.17 14.03 66.20 11.31
87.92 8.66 84.58 8.57 61.57 11.56
Note—AD, Alzheimer’s disease.
iment 1 for a PM response to qualifyas successful because
of the extra time required to turn and locate the space bar
of the computer keyboard. (No participant took longer
than 6 sec from target onset to make a response.) Response
timeswere again recorded to the nearest second(see note 2).
PM performance. The mean numbers of PM successes
out of five in the related and unrelated conditions are
shown in Figure 3 for young adults, older adults, and AD
patients. Performance was at or close to ceiling for young
adults. The young adults outperformed the older adults,
who, in turn, outperformed the AD patients. The related
and unrelated conditions resulted in similar performance
in all three groups.
Results
Recall of PM instructions. When asked at the end of
experiment to recall what they had been instructed to do
while watching the film, all participantsspontaneouslyre-
called the PM task instructions correctly with the follow-
ing exceptions.In the related condition,4 AD patients re-
quired a minimal prompt. In the unrelated condition, 2
older adultsand 4 AD patientsrequired a minimal prompt,
and 1 AD patient required more specific prompting. It
should be noted that all of these participantswho failed to
spontaneously recall the PM task instructions carried out
the PM task successfully on at least one occasion.
PM scoring. In both conditions, a successful PM re-
sponse was recorded if the participant pressed the space
bar of the computer (which reset the clock)within 6 sec of
the appearance of a clock (related condition)or an animal
(unrelated condition) in the film (each appearing on
screen for 3 sec). We allowed 2 sec longer than in Exper-
As in Experiment 1, the young group was excluded
from a two-way factorial ANOVA on these data with
group (2 levels: older and AD patients) and task (2 levels:
related and unrelated) as factors. This revealed a signifi-
F
5
MS 5
p ,
1.33,
cant effect of group [ (1,34) 7.22,
.02]. Neither the effect of task nor the interactionbetween
F ,
e
group and task reached significance (both
s
1).
The mean latencies for successful PM responses in
each conditionare shown in Table 4. It can be seen that the
young adults were faster than the older adults and AD pa-
tients, and that responses in the related condition were
faster than those in the unrelated condition. A two-way
factorial ANOVA revealed significant effects of group
F
5
5
MS 5
p ,
p ,
[ (2,52) 5.45,
0.81,
0.81,
.01] and condition
.01], with no inter-
F
MSe 5
[ (1,52) 9.11,
eF ,
action between them (
1). Post hoc Newman–Keuls
comparisons revealed that the young adults were signifi-
cantly faster than boththe older adultsand the AD patients
p ,
( s .01), who did not differ.
Recognition memory performance. The mean per-
centages of correct responses on the recognition memory
test completed after film viewing are shown in Table 4.
These data were analyzed with group (3 levels) and PM
task (2 levels: related and unrelated) as between-subjects
F
5
factors. There was a significant effect of group [ (2,52)
MS 5
p ,
20.44,
126.56,
.001], with Newman–Keuls
comparisoens revealing that the AD patients performed
significantly less well than either the young adults or the
older adults, who themselves did not differ significantly.
There was no significantdifference between the two tasks
Figure 3. Mean numbers of PM successes out of five in the
related cue–action (clock–clock) and unrelated cue–action
(animal–clock) conditionsfor young adults, older adults, and AD
patients in Experiment 2. Error bars represent 1 SE.
F ,
(
1) and no interaction between group and task
F
5
MS 5 p .
[ (2,52) 1.07,
126.56,
.10].
e

880
MAYLOR, SMITH, DELLA SALA, AND LOGIE
Table 5
by normal aging. The very close similarity between the
numbers of PM successes in the event-based condition of
Experiment 1 and the related and unrelated event-based
conditions of Experiment 2 (see Figures 1 and 3) is fortu-
itous because it allows us to combine the data across ex-
periments to conduct more quantitative comparisons.
Table 5 shows the correlations between age and perfor-
mance for older adults only (since young adults were at or
close to ceiling in the event-based tasks).7 It can be seen
that for these participants, with an age range of 55 to 80
years, there were significant negative correlations with
age for all measures. The strongest correlation was for
recognition, which is surprising in view of the absence of
significantdifferencesbetween the youngand olderadults
in both experiments. However, a scatterplot of age against
recognition scores for the older adults in Table 5 revealed
that those in their 50s and early 60s performed above the
young mean whereas the over-75s performed well below
the youngmean. Two factors may have contributedto this:
(1) The older adultsmay have been more familiar with the
Aberdeen area than the young adults, and (2) the recogni-
tion task may have been a particularlydifficult one for the
very old participantsbecause the scene as depicted on the
still photograph may have matched the film exactly for
only a brief period of time. The recognition correlation
was followed by free recall, PM, sentence span, and digit
span. Note again that there is no strong evidence that PM
is unusual in its sensitivity to the effects of normal aging.
For dementia, clearly the AD patients performed sig-
nificantly worse than the age-matched older adults across
all memory tasks. To compare the effects of AD on PM
and RM directly, we pooled the data from the older adults
and AD patients from the event-based condition of Ex-
periment 1 and the related and unrelated event-based con-
Correlations Between Age and Scores for Older Adults
in the Event-Based Condition of Experiment 1 and Related
and Unrelated (Event-Based) Conditions of Experiment 2
Measure
r (33)
PM successes
Recognition
Digit span
Sentence span
Free recall
2.45**
2.60**
2.39*
2.33*
2.56**
*p # .05. **p , .01.
Discussion
As in Experiment 1, there was good retention of the PM
task instructions and yet there were PM deficits associ-
ated with both normal aging and dementia. Thus older
adults made numerically fewer successful PM responses
and were significantlyslower in producingthem than were
young adults. AD patients were significantly less suc-
cessful than older adults but not significantlyslower.
In terms of the numbers of successful PM responses,
there was no difference between the related and unrelated
cue–action conditionsfor any of the groups. However, PM
responses were significantly faster when the cue and ac-
tion were related than when they were unrelated. Our ex-
pectation was that the effects of both aging and dementia
would be reduced in the related (clock–clock) condition
relative to the unrelated (animal–clock) condition be-
cause, in the former case, retrieval of the appropriate ac-
tion (i.e., stop the clock) in response to the cue (clock)
would be primed by the cue itself. This manipulationwas
apparently not strong enough to influence whether or not
participants successfully responded; however, it was at
least successful in producing faster response times in the
related than in the unrelated condition.6 Contrary to pre-
diction, there was no evidence of differential effects of
condition on the PM response times across the three
groups. However, with very few trials contributing to the
mean response times for each participant, it is probably
n 5
ditions of Experiment 2 (
65) and calculated stan-
dardized scores. The means are presented in Table 6,
t
together with the results of tests comparing the older
adults and AD patients.8 The strongest effect of AD was
for recognition, followed by free recall, sentence span,
digit span, and finally PM. Again, the ordering of RM
tasks did not exactly correspond to predictions, perhaps
because of the use of an unusualrecognitiontest with non-
verbal materials. Nevertheless, there were striking par-
allels between the patterns for normal aging and demen-
3
unrealisticto expectsignificant group task interactions
for this particularmeasure of performance. The results do
at least suggest that even the performance of AD patients
can be sensitive to subtle manipulations of the PM task.
COMBINED ANALYSES
Effects of Normal Aging and Dementia
on PM Versus RM
Table 6
Standardized Scores for Older Adults and AD Patients in the
Event-Based Condition of Experiment 1 and Related and
Unrelated (Event-Based) Conditions of Experiment 2
In this section, we address the crucial issue raised in the
introductionof the relative effects on PM and RM of both
normal aging and dementia. For normal aging, the com-
parison between young and older adults in both experi-
ments revealed modest PM deficits in the context of sub-
stantial decline in free recall, but no decline in digit span,
sentencespan, or recognition.Thus, we have evidencethat
performance on these particular laboratory-based PM
tasks is neither unusually spared (see Einstein & Mc-
Older
(n 5 35)
AD Patients
(n 5 30)
Difference
t(63)
Measure
M
SD
M
SD
PM successes
Recognition
Digit span
Sentence span
Free recall
0.48
0.64
0.59
0.57
0.63
0.64
0.77
0.80
0.67
0.77
20.56
20.74
20.68
20.66
20.74
1.06
0.67
0.75
0.92
0.69
4.91*
7.62*
6.56*
6.24*
7.52*
Daniel, 1990) nor unusually impaired (see Craik, 1986) Note—AD, Alzheimer’s disease. *p , .001.

PROSPECTIVE AND RETROSPECTIVE MEMORY
881
n 5
time-based conditions (
92), except that in both these
Table 7
Correlations Between PM and RM Scores For the Older
Adults in the Event-Based Conditions of Experiments 1 and 2
(n 5 35) and Factor Analysis Loadings (Oblique Solution)
cases the separation between the two factors was even
clearer, with recognitionloadingsresembling those of the
other RM measures.
Correlations
DS SS
2.02 .09
.53** .39* .48**
.42* .43*
Loadings
Recog
FR
.22
Factor 1 Factor 2
CONCLUSIONS
PM successes
Recognition
Digit span
Sentence span
Free recall
.37*
–
.00
.66
.81
.79
.77
.97
.48
2.11
.02
The present study is the first to demonstratesignificant
deficits in the PM component of laboratory-controlled
PM tasks in patients suffering from AD in comparison
with age-matched controls. Importantly, the AD patients’
PM failures were not due to misunderstanding or forget-
ting the task instructions; instead, AD patients simply
failed more often to carry out the PM task in response to
–
–
.59**
–
.24
Note—Recog, recognition; DS, digit span; SS, sentence span; FR, free
recall. *p , .05. **p , .01.
tia with respect to RM (see Tables 5 and 6). Contrary to the appropriate cue (as, indeed, did older adults in com-
Huppert and Beardsall’s (1993) proposal,there was no ev- parison with young adults; see also Maylor, 1998). The
idence that PM was more impaired by AD than was RM. findings demonstrate that PM can be investigatedexperi-
A rather different (but neverthelessimportant)question
mentally in AD, although in achieving above-floor per-
is the extent to which PM deficits associated with normal formance from AD patients, there were ceiling effects for
aging and dementia can be predicted by RM deficits. This young adults on the same tasks.
was addressed by again focusing on the event-based con-
Contrary to predictions, there was no evidence of dif-
ditions from the two experiments. For normal aging, we ferential effects of AD on time- versus event-based PM
conducted a multiple regression analysis on the older tasks (Experiment 1) or on unrelated versus related
n 5
adults’ data (
35) in which the predicted variable was
cue–action conditions (Experiment 2). In the case of Ex-
the number of PM successes, and the predictor variables periment 1, the dementia findings mirror our normal
were the four measures of RM and age. (For correlations aging data using the same tasks (Logie et al., 2001). We
between PM and RM, see Table 7.) There were no signif- would therefore argue that time-based tasks may not nec-
icant independentcontributionsto the variance, but the ef- essarily be more demanding than event-based tasks in
t 5 2
1.78,
p 5
fect of age approachedsignificance [
.08]. terms of the degree of self-initiated activity required or
environmental support available. For example, in our
R2
With age includedin the regression equation, increased
from .218 to .296; age therefore accounted for an addi- event-based task, participantshad to remember to scan the
tional 7.8% of the variance over and above the measures whole of each scene for the appearance of an animal,
of RM.
For an assessment of the effects of dementia,the data for get” about the PM task for a couple of minutes after mak-
n 5
whereas in our time-based task, participants could “for-
older adults and AD patients (
multipleregression analysiswith PM successes as the pre-
dictedvariableand the RM measures, age, and group (0
65) were entered into a ing a successful response. Note also that it is very difficult
only
to design time- and event-based tasks that differ
in
5
9
terms of whether the cue for action is a particular time or
5
older adults; 1
AD patients) as predictor variables.
event. In both Experiment 1 and previous studies (e.g.,
Again, no variable made an independent contribution to Einstein et al., 1995), the task instructions were slightly
t 5
the variance, althoughgroup approachedsignificance[
lengthier for the time-based task in order to explain the
increased from .300 to .343 when clockface, the locationof the space bar, and so on. In sum-
group was added to the regression equation, AD thereby
accountingfor an independent4.3% of the variancein PM not always expect a greater deficit in time- than in event-
2
p 5
R2
1.96,
.055].
mary, there are good reasons for supposingthat we should
when RM measures and age were taken into account.
based PM tasks as a result of either normal aging or de-
mentia. Experiment 2 was an attempt to focus on a ma-
nipulation of retrieval demands within event-based tasks.
Factor Analysis
The data from the 35 older adults in the event-based This was clearly insufficient to influence PM success, al-
conditions of Experiments 1 and 2 were further analyzed
though response times were more sensitive to the manip-
by conducting a factor analysis on PM and RM scores. ulation, even in AD patients. Like West and Craik (1999),
Table 7 shows both the correlation matrix and the load- we conclude that useful information may be obtained by
ings for the two factors produced by the oblique solution examiningnot onlywhether or not participantsrespond to
reference structure. It can be seen that there is clear evi- PM targets, but how quickly they respond.
dence for separate (although correlated) factors of RM
(factor 1) and PM (factor 2), and these accounted for
The clock-checkingbehaviorin the time-based PM task
of Experiment 1 revealed that older adults and AD pa-
49.9% and 21.3% of the variance, respectively. The same tients checked the clock more often than did young adults
pattern emerged from factor analyses of (1) older adults prior to making a successful PM response. Previous stud-
n 5
and AD patients in all event-based conditions (
65), ies (e.g., Einstein et al., 1995; Park et al., 1997) have at-
and (2) older adults and AD patients in both event- and tributed age deficits in time-based PM tasks to reduced

882
MAYLOR, SMITH, DELLA SALA, AND LOGIE
clock-checking (self-initiated) behavior, but they did not 1996;onits relationshipto PM, see Bisiacchi,1996;Glisky,
distinguishbetween PM successes and failures. The pres- 1996; McDaniel, Glisky, Rubin, Guynn, & Routhieaux,
ent breakdown of results is more consistent with the view 1999). Certainly, patients with lesions of the frontal lobes
that momentary lapses of intention are primarily respon- have been found to be particularly impaired on PM tasks
sible for time-based PM deficits in normal aging and de- (Cockburn,1995;Shallice& Burgess, 1991). It remains to
mentia (as they are in event-based tasks; see Maylor, be determined whether the unique variance associated
1993a, 1996a; West & Craik, 1999), although deficits in with normal aging and dementia can be attributable to
other factors such as time estimation (e.g., Craik & Hay, frontal decline.
1999) and distractibility(e.g., Hasher & Zacks, 1988;Mc-
Dowd, Oseas-Kreger, & Filion, 1995) may also play a ential effects of either normal aging or dementia on PM
role. and RM, the results from factor analysis of the older
Finally, although the present study did not show differ-
Importantly, the present data allowed us to compare the adults’ data revealed separate PM and RM factors. Thus,
effects of both normal agingand dementiaon PM and RM the present findings from objective measures of memory
tasks. Note, first, that the rank ordering of the four RM converge with data from a self-rated questionnaire of
tasks with respect to correlationswith age for older adults everyday PM and RM failures, which revealed distinct
(Table 5) was identical to that of the differences between PM and RM factors in its latent structure (Crawford,
the standardized scores of older adults and AD patients Smith, Maylor, Della Sala, & Logie, in press). Together,
(Table 6). This provides further evidence of similarity be- these results are at least consistent with arguments that
tween the effects of normal aging and dementiaon RM. In PM andRM tasksmay dependondifferentprocessesand/or
contrast, on this analysis, PM was actually more impaired different brain areas, with PM tasks more reliant on exec-
by normal aging than at least some of the RM tasks utive processes such as strategic planning, self-initiated
(namely, span tasks), whereas PM was less impaired by retrieval, and task interruption or switching (see Brandi-
dementia than all four of the RM tasks. This result would monte et al., 1996; McDaniel et al., 1999, for discussion).
seem to suggest that PM is not disproportionately im-
paired by dementia relative to RM (or, at least, that our
particularPM tasks are not more impaired than any of our
sample of RM tasks). This is consistent with our ques-
tionnaire data (Smith et al., 2000), in which carers re-
ported that AD patients’ PM failures were no more fre-
quent (albeit more frustrating) than their RM failures in
everyday life. Together, these findings would seem to
argue against Huppert and Beardsall’s (1993) proposal
that PM tasks are particularly sensitive to the early stages
of dementia.However, it should be acknowledgedthat the
present AD patients were not necessarily in the early
stages of dementia since they had been attending the
Memory Clinic for an average of 31.5 months. Further
empirical work (requiring a longitudinaldesign) is there-
REFERENCES
Baddeley,A., Logie, R., Bressi, S., Della Sala, S., & Spinnler, H.
(1986). Senile dementia and working memory. Quarterly Journal of
Experimental Psychology, 38A, 603-618.
Bisiacchi, P. S. (1996). The neuropsychological approach in the study
of prospective memory. In M. Brandimonte, G. O. Einstein, & M. A.
McDaniel (Eds.), Prospective memory: Theory and applications
(pp. 297-317). Mahwah, NJ: Erlbaum.
Brandimonte,M., Einstein, G. O., & McDaniel,M. A. (Eds.)(1996).
Prospective memory: Theory and applications. Mahwah, NJ: Erl-
baum.
Camp, C. J., Foss, J. W., Stevens, A. B., & O’Hanlon, A. M. (1996).
Improving prospective memory task performance in persons with
Alzheimer’s disease. In M. Brandimonte,G. O. Einstein, & M. A. Mc-
Daniel (Eds.), Prospective memory: Theory and applications(pp. 351-
367). Mahwah, NJ: Erlbaum.
fore necessary to determine whether PM failures are ear- Capitani, E., Della Sala, S., Logie, R. H., & Spinnler, H. (1992).
Recency, primacy, and memory: Reappraising and standardising the
serial position curve. Cortex, 28, 315-342.
Ceci, S. J., Baker, J. G., & Bronfenbrenner, U. (1988). Prospective
remembering, temporal calibration, and context. In M. M. Gruneberg,
lier indicators of dementia than are RM failures.
The regression analysesreveal some similaritybetween
the effects of normal aging and dementia on PM such that
there were trends in both cases to suggest that neither ef-
fect can be entirely accounted for by RM deficits. Previ-
ous studies have also revealed significanteffects of age in
multiple regression analyses on PM performance after
measures of current cognitive ability (including fluid and
P. E. Morris, & R. N. Sykes (Eds.), Practical aspects of memory: Cur-
rent research and issues. Vol. 1: Memory in everyday life (pp. 360-
365). Chichester, U.K.: Wiley.
Ceci, S. J., & Bronfenbrenner,U. (1985).“Don’t forget to take thecup-
cakes outof theoven”: Prospectivememory, strategic time-monitoring,
and context. Child Development, 56, 152-164.
crystallized intelligenceand processing speed) have been Cockburn, J. (1995). Task interruption in prospective memory: A
frontal lobe function? Cortex, 31, 87-97.
Cockburn, J., & Smith, P. T. (1988). Effects of age and intelligence on
everydaymemory tasks. In M.M.Gruneberg,P. E. Morris,& R. N. Sykes
(Eds.), Practical aspects of memory: Current research and issues. Vol.
taken into account(e.g., Cockburn& Smith, 1991;Mäntylä
& Nilsson, 1997; Maylor, 1998). Further work is required
to determine what factors are responsible for the remain-
ing age-related, and now dementia-related, variance. One
possibility is that the relationship between normal aging
and PM may be particularly associated with decline in
frontal (dorso-lateral) functioning (on the frontal hypoth-
esis of aging, see Moscovitch & Winocur, 1992; Parkin,
1997; Perfect, 1997; Phillips & Della Sala, 1999; West,
2: Clinical and educational implications (pp. 132-136). Chichester,
U.K.: Wiley.
Cockburn, J., & Smith, P. T. (1991). The relative influence of intelli-
gence and age on everyday memory. Journal of Gerontology: Psy-
chological Sciences, 46, P31-36.
Craik, F. I. M. (1986). A functional account of age differences in mem-
ory. In F. Klix & H. Hagendorf (Eds.), Human memory and cognitive

PROSPECTIVE AND RETROSPECTIVE MEMORY
883
capabilities: Mechanisms and performances (pp. 409-422). Amster-
dam: Elsevier.
existing experimental paradigms. Applied Cognitive Psychology, 6,
507-524.
Craik, F. I. M., Anderson, N. D., Kerr, S. A., & Li, K. Z. H. (1995). Logie, R. H., Maylor,E. A., Della Sala, S., & Smith, G. (2001). Ef-
Memory changes in normal aging. In A. D. Baddeley, B. A. Wilson,
& F. N. Watts (Eds.), Handbook of memory disorders (pp. 211-241).
Chichester, U.K.: Wiley.
Craik, F. I. M., & Hay, J. F. (1999). Aging and judgments of duration:
Effects of task complexity and method of estimation. Perception &
Psychophysics, 61, 549-560.
fects of working memory demands on age differences in prospective
memory. Manuscript submitted for publication.
Mäntylä, T., & Nilsson, L.-G. (1997). Remembering to remember in
adulthood:A population-based study of aging and prospective mem-
ory. Aging, Neuropsychology, & Cognition, 4, 81-92.
Maylor,E. A. (1990).Age and prospectivememory. Quarterly Journal
of Experimental Psychology, 42A, 471-493.
Craik, F. I. M., & Kerr, S. A. (1996). Commentary: Prospective mem-
ory, aging, and lapses of intention.In M. Brandimonte, G. O. Einstein, Maylor, E. A. (1993a). Aging and forgetting in prospective and retro-
& M. A. McDaniel (Eds.), Prospective memory: Theory and applica-
tions (pp. 173-197). Hillsdale, NJ: Erlbaum.
Crawford, J. R., Smith, G., Maylor, E. A., Della Sala, S., &
Logie, R. H. (in press). The Prospective and Retrospective Memory
Questionnaire(PRMQ): Normative data and latent structure in a large
non-clinical sample. Memory.
spective memory tasks. Psychology & Aging, 3, 420-428.
Maylor,E. A. (1993b).Minimized prospectivememory loss in old age.
In J. Cerella, J. Rybash, W. Hoyer, & M. L. Commons (Eds.), Adult in-
formation processing: Limits on loss (pp. 529-551). San Diego: Aca-
demic.
Maylor, E. A. (1995). Prospective memory in normal aging and de-
Duff, S. C., & Logie, R. H. (2001). Processing and storage in working
memory span. Quarterly Journal of Experimental Psychology, 54A,
31-48.
d’Ydewalle, G., Luwel, K., & Brunfaut, E. (1999).The importance
of ongoing concurrent activities as a function of age in time- and
event-based prospectivememory. EuropeanJournal of CognitivePsy-
chology, 11, 219-237.
mentia. Neurocase, 1, 285-289.
Maylor, E. A. (1996a). Age-related impairment in an event-based
prospective memory task. Psychology & Aging, 11, 74-78.
Maylor, E. A. (1996b). Does prospective memory decline with age? In
M. Brandimonte, G. O. Einstein, & M. A. McDaniel (Eds.), Prospec-
tive memory: Theory and applications (pp. 173-197). Mahwah, NJ:
Erlbaum.
Einstein, G. O., Holland, L. J., McDaniel, M. A., & Guynn, M. J.
(1992). Age-related deficits in prospective memory: The influence of
task complexity. Psychology & Aging, 7, 471-478.
Einstein, G. O., & McDaniel,M. A. (1990).Normal agingand prospec-
tive memory. Journal of Experimental Psychology: Learning, Mem-
ory, & Cognition, 16, 717-726.
Maylor, E. A. (1998). Changes in event-based prospective memory
across adulthood. Aging, Neuropsychology, & Cognition, 5, 107-128.
Maylor,E. A., Darby, R. J., Logie,R. H., Della Sala, S., & Smith, G.
(2002). Prospective memory across the lifespan. In P. Graf & N. Ohta
(Eds.), Lifespan development of human memory (pp. 235-256). Cam-
bridge, MA: MIT Press.
Einstein, G. O., McDaniel, M. A., Manzi, M., Cochran, B., &
Baker, M. (2000). Prospective memory and aging: Forgetting inten-
tions over short delays. Psychology & Aging, 15, 671-683.
Einstein, G. O., McDaniel, M. A., Richardson, S. L., Guynn, M. J.,
& Cunfer, A. R. (1995).Aging and prospectivememory: Examining
the influences of self-initiated retrieval processes. Journal of Experi-
mental Psychology: Learning, Memory, & Cognition, 21, 996-1007.
McDaniel, M. A., Glisky, E. L., Rubin, S. R., Guynn, M. J., &
Routhieaux, B. C. (1999). Prospective memory: A neuropsycho-
logical study. Neuropsychology, 13, 103-110.
McDowd, J. M., Oseas-Kreger,D. M.,& Filion,D. L. (1995).Inhibitory
processes in cognition and aging. In F. N. Dempster & C. J. Brainerd
(Eds.), Interference and inhibition in cognition (pp. 363-400). San
Diego: Academic Press.
Einstein, G. O., Smith, R. E., McDaniel, M. A., & Shaw, P. (1997). McKhann, G., Drachmann, P., Folstein, M., Katzman, R., Price, P.,
Aging and prospective memory: The influence of increased task de-
mands at encoding and retrieval. Psychology & Aging, 12, 479-488.
Folstein, M. F., Folstein, S. E., & McHugh, P. R. (1975). Mini-
Mental State: A practical method for grading the cognitivestate of pa-
& Stadler, E. (1984). Clinical diagnosis of Alzheimer’s disease: Re-
port of the NINCDS-ADRDA work group under the auspices of the
Department of Health and Human Services task force on Alzheimer’s
disease. Neurology, 34, 939-944.
tients for the clinician. Journal of Psychiatric Research, 12, 189-198. McKitrick, L. A., Camp, C. J., & Black, F. W. (1992). Prospective
Glisky, E. L. (1996). Prospective memory and the frontal lobes. In
M. Brandimonte, G. O. Einstein, & M. A. McDaniel (Eds.), Prospec-
tive memory: Theory and applications (pp. 249-266). Mahwah, NJ:
Erlbaum.
memory interventionin Alzheimer’s disease. Journalof Gerontology:
Psychological Sciences, 47, 337-343.
Morris,R. G. (1996).The cognitive neuropsychologyof Alzheimer-type
dementia. Oxford: Oxford University Press.
Harris, J. E., & Wilkins, A. J. (1982). Remembering to do things: A
theoretical framework and an illustrative experiment. Human Learn-
ing, 1, 123-136.
Moscovitch,M.,& Winocur,G. (1992).Theneuropsychologyofmem-
ory and aging.In F. I. M. Craik & T. A. Salthouse(Eds.), The handbook
of aging and cognition (pp. 315-372).Hillsdale, NJ: Erlbaum.
Hasher, L., & Zacks, R. T. (1988). Working memory, comprehension, Nebes, R. D. (1992). Cognitive dysfunction in Alzheimer’s disease. In
and aging: A review and a new view. In G. H. Bower (Ed.), The psy-
chologyof learning and motivation(Vol. 22, pp. 193-225).New York:
Academic Press.
F. I. M. Craik & T. A. Salthouse (Eds.), The handbook of aging and
cognition (pp. 373-445). Hillsdale, NJ: Erlbaum.
Park, D. C., Hertzog, C., Kidder, D. P., Morrell, R. W., & May-
horn, C. B. (1997). Effect of age on event-based and time-based
prospective memory. Psychology & Aging, 12, 314-327.
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.
Perfect, T. (1997). Memory aging as frontal lobe dysfunction.In M. A.
Conway (Ed.), Cognitive models of memory (pp. 315-339). Hove,
U.K.: Psychology Press.
Hohaus, L., Shum, D., White, J., & Caird, D. (2000, April). The ef-
fect of salience on prospective memory declines with age. Paper pre-
sented at the Eighth Cognitive Aging Conference, Atlanta.
Hultsch, D. F., Hertzog, C., Dixon, R. A., & Small, B. J. (1998).
Memory changein the aged. Cambridge: Cambridge University Press.
Huppert, F. A., & Beardsall, L. (1993). Prospective memory impair-
ment as an early indicator of dementia. Journal of Clinical and Ex-
perimental Neuropsychology, 15, 805-821.
Huppert, F. A., Johnson, T., & Nickson, J. (2000). High prevalence of
prospective memory impairment in the elderly and in early-stage de-
mentia: Findings from a population-based study. Applied Cognitive
Psychology, 14, S63-S81.
Kidder, D. P., Park, D. C., Hertzog, C., & Morrell, R. (1997).
Prospective memory and aging: The effects of working memory and
prospectivememory task load. Aging,Neuropsychology,& Cognition,
4, 93-112.
Phillips, L. H., & Della Sala, S. (1999). Aging, intelligence, and
anatomical segregation in the frontal lobes. Learning & Individual
Differences, 10, 217-243.
Prull, M. W., Gabrieli, J. D. E., & Bunge, S. A. (2000). Age-related
changes in memory: A cognitiveneuroscience perspective. In F. I. M.
Craik & T. A. Salthouse (Eds.), The handbookof aging and cognition
(2nd ed., pp. 91-153). Mahwah, NJ: Erlbaum.
Rabbitt, P. (1996). Why are studies of “prospectivememory” planless?
In M.Brandimonte,G. O. Einstein, & M. A. McDaniel (Eds.), Prospec-
Kvavilashvili,L. (1992). Remembering intentions:A critical review of

884
MAYLOR, SMITH, DELLA SALA, AND LOGIE
tive memory: Theory and applications (pp. 239-248). Hillsdale, NJ:
Erlbaum.
Salthouse, T. A. (1991). Theoretical perspectives on cognitive aging.
3. It may seem unduly strict to expect participants (especially older
people) to producea response within 2 sec after the target time. But note
that participants had the opportunity to watch the clock up to the target
time and therefore achieve accurate responses by anticipating the clock
Hillsdale, NJ: Erlbaum.
Shallice,T., & Burgess, P. W. (1991). Deficits in strategy application reaching 3 min. Results using more lenient time windowsfor correct re-
following frontal lobe damage in man. Brain, 114, 727-741.
Sinnott, J. D. (1989). Prospective/intentional memory and aging:
Memory as adaptive action. In L. W. Poon, D. C. Rubin, & B. A. Wil-
sponses are presented later.
4. Extending the time window for successful responses in the time-
based task to between 2 sec before and 4 sec after the target time resulted
son (Eds.), Everyday cognition in adulthood and late life (pp. 352- in means of 3.60 and 2.67 for the older group and AD patients, respec-
369). Cambridge: Cambridge University Press.
tively. Note that, on this criterion, the difference between the two older
Smith, G., Della Sala, S., Logie, R. H., & Maylor, E. A. (2000). groups was actually smaller for the time-based task than for the event-
Prospective and retrospective memory in normal aging and dementia:
A questionnaire study. Memory, 8, 311-321.
based task.
5. This latter suggestion may seem inconsistent with the finding that
performance on the ongoingtask (as measured by recognition memory)
did not differ between young and older participants. However, it is very
unlikely that the older group’s additional clock checks coincided with
any of the 12 scenes from the 17-min film used in the recognition test.
6. One possible explanation for this result is that the particular clocks
in the film may have been easier to detect than the animals. To test this
possibility,a simple vigilance task was conducted with a group of young
adults (n 5 12; mean age 5 21 years) recruited from the same sources
as those in Experiments 1 and 2. Participants were asked to view both the
Spinnler, H. (1999). Alzheimer’s disease: Neuropsychological defects
according to the topographicalspreading of neuronaldegeneration.In
F. Denes & L. Pizzamiglio (Eds.), Handbook of clinical and experi-
mental neuropsychology(pp. 699-746).Hove, U.K.: PsychologyPress.
Spinnler, H., & Della Sala, S. (1988). The role of clinical neuro-
psychology in the neurological diagnosis of Alzheimer’s disease.
Journal of Neurology, 235, 258-272.
We st, R. L. (1996). An application of prefrontal cortex function theory
to cognitive aging. Psychological Bulletin, 120, 272-292.
West, R. [L.], & Craik, F. I. M. (1999). Age-related decline in prospec- clock and the animal films (with order counterbalanced)and to press the
tive memory: The roles of cue accessibility and cue sensitivity. Psy-
chology & Aging, 14, 264-272.
Wilcock, J. K., Hope, R. A., Oppenheimer, C., Reynolds, J. P.,
space bar of the computer as quickly as possible whenever a target
(clocks or animals, depending on the film) appeared. Participants re-
sponded to all the clocks and animals within 6 sec of their appearance,
with the exception of 1 participant who missed an animal. The mean re-
Rossor, M. N., & Davies, M. B. (1989). Recommended minimum
data to be collected in research studies on Alzheimer’s Disease. Jour- sponse times were 1.6 sec (SD 5 0.7) for clocks and 1.5 sec (SD 5 0.6)
nal of Neurology, Neurosurgery, & Psychiatry, 52, 693-700.
Wilson, B. A., Cockburn,J., & Baddeley,A. D. (1985).TheRivermead
for animals, with no significant difference between them [t(11) , 1],
thereby discounting the possibility that clocks were easier to detect than
behavioural memory test. Titchfield, U.K.: Thames Valley Test Co. animals.
Zacks, R. T., Hasher, L., & Li, K. Z. H. (2000). Human memory. In
7. The correlations between age and RM scores were similar to those
F. I. M. Craik & T. A. Salthouse (Eds.), The handbook of aging and shown in Table 5 when the older adults assigned to the time-based con-
cognition (2nd ed., pp. 293-357). Mahwah, NJ: Erlbaum.
dition of Experiment 1 were also included.
8. The standardized scores for PM successes in the time-based condi-
tion of Experiment 1 were as follows: M 5 0.38, SD 5 0.95 for older
adults; M 5 20.47, SD 5 0.88 for AD patients [t(25) 5 2.37, p , .05].
9. There were significant correlations between PM and each of the
NOTES
1. The film length of 17 min ensured that if participants were late in
respondingafter 3 min had elapsed, there were always five opportunities RM measures (n 5 65): .46 (recognition),.39 (digit span), .45 (sentence
for PM responses.
span), and .50 (free recall).
2. Throughout,responses between 0 and 0.5 sec were entered as 0 sec,
between 0.5 and 1.5 sec as 1.0 sec, and so on. Hence, mean response
times of 0 sec were possible.
(Manuscript received August 18, 2000;
revision accepted for publication April 26, 2002.)