West, R., & Craik, F. I. M. (2001). Influences on
the efficiency of prospective memory in younger and older adults. Psychology
and Aging, 16, 682-696.
The authors had several goals for the present studies
with respect to the processes involved in prospective memory (PM), and
especially processes which might underlie a perceived age-related decline in PM
performance. These goals included 1) evaluation of the
transfer-appropriate-processing (TAP) explanation of PM. (The TAP hypothesis
refers to the change in the relationship between the type of processing
required by the ongoing activity and that required by the PM task.) 2)
Evaluation the noticing + search model of PM. Based on previous aging research,
they expected the age-related decline in PM to be found in the search or
effortful component of the model, rather than the noticing, or automatic,
component. Also, they wished to evaluate an indication from other research that
some age-related declines result from disruption of the prospective component.
3) They investigated the possibility that the accessibility of the intention in
an activated state would afford a partial explanation of the age-related differences
in PM. They expected that high ongoing-task PM response congruity would mediate
(moderate) age differences in cue accessibility. 4) The final goal was to
evaluate the contribution of cognitive resources on PM, particularly working
memory, processing speed, and inhibitory control.
Experiment 1
This was a 24 design with two between-subject
variables (Age, Type of Ongoing Activity) and 2 within-subject variables (type
of PM cue, immediacy of response). For the perceptual ongoing activity, Ss
classified words into one of four color categories, while for the semantic task
they classified them into one of four superordinate categories. The perceptual
PM signal was a change in the type face from lower to upper case, while the
semantic PM signal was presentation of a word from a particular category.
There were 40 young adults (M = 22.55 yrs) and 40 older
adults (M = 69.48 yrs). Subjects were assigned to the type of ongoing activity
conditions randomly. The older adults had more years of education than the younger,
but did not differ in vocabulary scores.
In the PM task Ss learned the summary
statement-response connections to a 3-perfect criterion. In the test phase
there were 16 blocks of 20 trial which consisted of 19 ongoing activity trials
and one PM-cue trial. In the ongoing activity, the words from various
categories were presented in the center of the computer screen in lower case
letters and one of 4 colors. Ss were to press a key indicating the color
or the superordinate category. On PM-cue trials the S had to press a different
key. For example, red upper case - d, blue upper case - f, tuber - j, and
building part - k. There were 8 perceptual and 8 semantic cues with half of
each type being immediate and the remaining being delayed.
To assess other processes which might be factors in PM,
the Ss were administered additional tasks. These included the WAIS Digit
substitution task (speed), the SOPT (working memory), and a Stroop task
(inhibitory control).
Results
Analysis of the ongoing activity indicated that younger
and older adults were similar when dealing with color cues, but older adults
were better when dealing with categories. Analysis of the proportion of the
PM-cues eliciting a correct prospective-memory response were analyzed, (Table
1, p. 687). All the main effects were significant. Younger adults (M = .63)
were better than older adults (M = .56); perceptual cues (M = .61) were better
than semantic cues (M = .37); immediate responses were better (M = .57) than
postponed responses (M = .41). The Cue Type X Delay interaction was significant
indicating that semantic cues were retained better than perceptual cues.
The compatibility of the cue and type of processing
required in the ongoing task was only marginally significant, although in
the direction predicted by the TAP
hypothesis. The same was true for the effect of age on the interaction.
Components of PM and age. They looked at the
number of confusion and omission errors (Figure 1). Young adults and older
adults had similar confusion errors, but older adults had more omission errors
than younger adults. The Ongoing Activity X Type of Error interaction was also
significant. The authors concluded that age and ongoing activity affect the
prospective component of PM.
Age, PM, and resources. The effect of various
variables were evaluated using regression models (Tables 2 and 3), and results
indicated that age had an independent influence on the ability to recall the PM
cues and intentions. The authors also concluded that processing speed was the only
cognitive resource which had a unique influence on PM.
Experiment 2
There were three goals for this experiment. 1) To examine
the role of PM-cue accessibility to the age-related decline in the efficiency
of PM; 2) to further explore predictions from the TAP hypothesis; and 3) to
provide more control over the possible confound between the number of PM cues
and their defining features and differences in the distinctiveness of the cue
types.
The authors reduced the number of semantic cues to 2,
with each semantic PM cue appearing four times. To reduce the uniqueness of
those repetitions 2 words from the other 2 categories were also repeated 4
times. Also, to reduce the distinctiveness of the perceptual PM cues, they
introduced some to-be-ignored perceptual lures. In addition, they included some
semantic PM lures. As in the first experiment Ss performed additional
tasks to assess other processes. These included the Digit Symbol task (speed),
a Stroop task (inhibitory control), and an Alpha Span task (working memory).
There were 24 Ss in each age group. The older
adults had more education and higher vocabulary scores than the younger adults.
The older adults were also slower, and had more interference on the Stroop
task, but did not differ significantly on the Alpha Span task. The procedure
was similar to the procedure from Experiment 1 with the exception of the
changes in the PM cues, and some changes in the Stroop and working memory
tasks. One major addition was the measurement of RT for all responses.
Results
Performance on the ongoing activity was very high, and
there were no significant main effects.
Results of the PM accuracy revealed that the younger
adults (M = .85) were more accurate than the older adults (M = .63), Ss
doing the semantic activity (M = .82) did better than those doing perceptual
activity (M = .66), and color cues (M = .79) produced better performance than
semantic cues (M = .69). There was some support for the TAP hypothesis as can
be seen in the table below, which shows the proportion of correct PM responses
as a function of type of ongoing activity and type of cue.
Type of Cue Ongoing
Activity
Category Color
Semantic .82 .55
Perceptual .81 .77
There was only partial support because there was not the
predicted crossover interaction, and the triple interaction with age was not
significant.
PM cue accessibility. The authors performed two, 2
x 2 x 3 ANOVAs on the RTs. The variables included were age (younger, older), type of ongoing processing (semantic,
perceptual), and trial type (ongoing, perceptual cue, semantic cue). Analyses
were performed separately on data from correct cues and lures. It was assumed
that the latency scores were a reflection of the accessibility of the PM cues.
Figure 2 shows that the increased latency from ongoing activity to the two PM
cue trials was greater in the older adults than in the younger adults.
Resources. The authors concluded that the effects
of age were mediated by speed and inhibitory control.
General Discussion
The authors provided some explanation for the failure to
find the full crossover interaction predicted by the TAP hypothesis, with some
discussion about the failure to find little support for the attenuation of
age-related differences in PM because of a processing match.
They also concluded that there were greater effects of age on the prospective component than on the retrospective component, even using moderate numbers of cues and intentions. However, they had concerns about the purity of the measures of the efficiency of PM. Also, there were no age differences, and very few errors overall, in the second experiment, which the authors attributed to the high degree of learning of the cue-intention associations.