Mäntylä, T. & Nilsson, L. (1997). Remembering to remember in
adulthood: A population-based study on aging and prospective memory, Memory,
4, 81-92.
The authors report that results of previous research are
mixed with regards to age-related differences in prospective memory, and they
suggest that some of the mixed evidence may be due to individual-difference
variables which are related to aging. Thus, one particular goal of the study
was to examine the ability of various individual-difference variables to
mediate any age differences in prospective memory. To this end they measured
demographic, psychometric, and blood variables. Additionally, the researchers
wanted to have a more “fine-grained description” of the description of the
relationship between age and prospective memory by including a large,
representative sample of subjects. They also investigated the suggestion that
age effects in prospective memory are related to the requirement for
self-initiated retrieval operations.
The authors chose to use a more natural prospective
memory task, which was to remind the experimenter to sign a document at the end
of the session.
Method
One hundred individuals in each of 10, 5-year cohorts
between 30 and 80 years were participants. Comparisons using demographic data
indicated that the sampled participants were likely representative of the
population.
The experiment took place in two, 2-hour sessions a week
apart. During the first session health and cognitive functioning variables were
measured; the second session included the memory and additional cognitive
functions. At the start of the second session subjects were asked to remind the
experimenter to sign a paper when the session was over. At the end of the
session the experimenter said: “Now we have finished the session.” If the
subject made no response after 15 s, the experimenter asked if there was “something
else to be done.” If the subject did not remember what was to be done, he or
she was explicitly asked to recall the task. These successive reminders were uncued retrieval, cued retrieval, and content
retrieval.
Results and Discussion
Forty and 45 year old subjects showed the highest uncued recall, with a gradual, non-significant decline with
increasing age until the 80-year olds, where the difference was significant.
The findings in cued recall were similar, except that the decline extended back
to the 75-year-old cohort. A derived measure, conditionalized
recall, was also examined. This measure represents the proportion of the
subjects in each age group who eventually recalled the task content who
remembered the task before any prompting. Results on this variable were similar
to other recall measures.
Do determine the effect of the potential moderating
variables the authors performed a logistic regression, using the variables
listed in Table 1. They split the sample into two parts, a validation sample
and a cross-validation sample. Results of the logistic regression on the
cross-validation sample indicated that only age and word-fluency were
significant predictors of differences in prospective memory.