Aging Neuropsychology and Cognition 1382-5585/03/1001-020$16.

00
2003, Vol. 10, No. 1, pp. 20–27 # Swets & Zeitlinger

Effects of Age, Domain, and Processing Demands

on Memory Span: Evidence for Differential Decline
Joel Myerson, Lisa Emery, Desirée A. White, and Sandra Hale
Washington University, MO, USA

ABSTRACT
Downloaded By: [Washington University in St Louis] At: 23:37 21 April 2010

Analysis of cross-sectional data from the normative sample of the Wechsler Memory Scale – Third Edition
(WMS-III) revealed different patterns of age-related differences in memory span measures depending on the
type of memory item, processing demands, and the age of the older adult group. Regression of memory span
on age revealed that the slope for Spatial Span raw scores was significantly more negative than the slope for
Digit Span raw scores. There was no significant difference, however, either between the slopes for forward
and backward Digit Span or between the slopes for forward and backward Spatial Span. Regression of
Letter-Number Sequencing raw scores on age showed a distinctive, curvilinear pattern. Taken together, the
present findings suggest that at least two mechanisms are involved in age-related differences in memory
span. One mechanism, associated with a relatively linear decrease in memory span as a function of age, may
differentially affect the storage of different types of information (e.g., sequences of digits vs. spatial
locations). The other mechanism, evidenced by the curvilinear trend in Letter-Number Sequencing scores,
may be tentatively attributed to a decline in executive aspects of working memory that becomes increasingly
pronounced with age.

It is well established that older adults perform than verbal storage (e.g., Jenkins, Myerson,
more poorly on measures of working mem- Joerding, & Hale, 2000; Myerson, Hale, Rhee,
ory than do younger adults (Salthouse, 1994; & Jenkins, 1999), or executive aspects of working
Verhaeghen, Marcoen, & Goossens, 1993). There memory may be more affected than storage
is no consensus, however, regarding which capacity (e.g., Dempster, 1992; Moscovitch &
aspects of the working memory system are most Winocur, 1992; West, 1996).
affected by aging. Most models of working It is commonly believed that older adults show
memory assume two types of components (e.g., a greater discrepancy between forward and back-
Baddeley, 1986; Engle, Laughlin, Tuholski, & ward digit span than young adults, and such a
Conway, 1999): storage components that may be differential deficit could be a consequence of a
specific to the type of information (e.g., verbal and specific age-related decline in executive or pro-
visuospatial) and processing or executive function cessing aspects of working memory. This is
components that are used for selecting, manip- because forward digit span may primarily mea-
ulating, and coordinating information in the sure storage, whereas backward digit span may
storage components. Aging could have an effect reflect both storage and processing because it
on any or all of these components. For example, requires that a person must maintain numbers in
spatial storage may be more affected by aging memory and also manipulate those numbers.

Address correspondence to: Joel Myerson, Department of Psychology, Campus Box 1125, Washington University,
Saint Louis, MO 63130, USA. E-mail: [email protected]
Accepted for publication: August 20, 2002.
 AGE DIFFERENCES IN MEMORY SPAN MEASURES 21

Alternatively, if the additional processing re- age differences in memory span for visuospatial
quired by backward span tasks is spatial in stimuli (e.g., locations and gestures) than for
nature (Costa, 1975; Hoshi et al., 2000; Rapport, verbal stimuli (e.g., digits and words).
Webster, & Dutra, 1994; Weinberg, Diller, In contrast, Salthouse, Kausler, and Saults
Gertsmann, & Schulman, 1972), then an age (1988) reported evidence against a differential
deficit in spatial processing could result in back- age-related decline in spatial memory span. Salt-
ward span being more age-sensitive than forward house et al. presented participants with a matrix of
span. letters, with certain letters highlighted, and
All of this, however, proceeds from the pre- instructed them to remember either the locations
mise that backward span actually is more affected of these highlighted letters or the letters them-
by aging than forward span. According to the selves. Regression analyses indicated that the rate
technical manual for the Wechsler Adult Intelli- of decrease in memory performance with age
gence Scale – Third Edition (WAIS-III; Psycho- was similar for both verbal (letters) and spatial
Downloaded By: [Washington University in St Louis] At: 23:37 21 April 2010

logical Corporation, 1997a): (locations) materials. A second study (Salthouse,

1995) examined parallel verbal and visuospatial
The ability to repeat digits in the order of
versions of three different working memory tasks,
verbal (forward) presentation . . . tends to
and this study also failed to find evidence that the
remain relatively stable with aging in normally
visuospatial memory abilities of older adults are
functioning men and women
differentially affected relative to their verbal
. . . Digit Span Backward, however, is more
memory abilities.
affected by aging and by impairment . . .
Given the mixed evidence with respect to
[N]ormally functioning adults over 70 years
both forward versus backward span and verbal
old show a greater discrepancy [than younger
versus visuospatial memory span, a large scale
adults], with a significantly shortened back-
study is needed to resolve these issues. The data
ward span. (pp. 187–188)
from the standardization tables of the Wechsler
The experimental literature, however, presents Memory Scale – Third Edition (WMS-III) and
conflicting evidence on this point. On the one the WAIS-III (Psychological Corporation, 1997b)
hand, Babcock and Salthouse (1990) reported a afford the opportunity for such a study. The
meta-analysis that indicated a greater effect of age standardization sample included more than a
on backward than forward digit span. On the other thousand adults of various ages, and each age
hand, a meta-analysis by Verhaeghen et al. (1993) group was selected to be representative of the
showed that the effect size for backward digit United States population with respect to sex,
span did not differ from that for forward digit level of education, race=ethnicity, and geo-
span. Moreover, Gregoire and Van der Linden graphic region. The WMS-III includes measures
(1997) analyzed data from the standardization of forward and backward digit span as well
sample for the French adaptation of the Wechsler as forward and backward spatial span (Psycho-
Adult Intelligence Scale – Revised (WAIS-R) and logical Corporation, 1997a). The WMS-III also
found that the slope of the age-related decrease in includes a new measure of working memory, the
forward digit span was the same as that for back- Letter-Number Sequencing subtest, in which a
ward digit span. person is presented with a series of numbers and
The evidence regarding a differential age- letters, and then must recall the numbers in
related decline in verbal and visuospatial memory numerical order followed by the letters in alpha-
span is also conflicting. For example, a recent betical order. Because Letter-Number Sequencing
study by Myerson et al. (1999) showed that presumably places an even higher load on the
although young adults and older adults had simi- ability to manipulate information online than
lar digit spans, the older adults had much lower either forward or backward digit span, it may
spatial spans than young adults. Other studies provide a more sensitive test of the effects of age
(e.g., Dolman, Roy, Dimeck, & Hall, 2000; on the processing or executive aspects of working
Jenkins et al., 2000) have also reported greater memory.
 22 JOEL MYERSON ET AL.

METHOD Forward and Digits Backward, the test begins with

series of two numbers. For Digits Forward, the test
continues to a maximum of eight numbers, and for
Sample
Digits Backward the test continues to a maximum of
All data were obtained from the standardization tables
seven numbers. Examinees are given two trials at each
in the WMS-III Administration and Scoring Manual
series length, and the test continues until both trials of a
(Psychological Corporation, 1997b). As described in
series length are failed. One point is awarded for each
the WAIS-III – WMS-III Technical Manual (1997a), the
trial that the examinee answers correctly. The total raw
WMS-III standardization sample consists of 1,250
score for Digit Span is the sum of the trials answered
adults partitioned into 13 groups (ages 16–17, 18–19,
correctly for both Digits Forward and Digits Backward.
20–24, 25–29, 30–34, 35–44, 45–54, 55–64, 65–69, 70–
The maximum possible score for the Digit Span subtest
74, 80–84, 85–89). Each of the first 11 groups (through
is 30 (16 points for Digits Forward and 14 points for
age 74 years) contains data from 100 participants,
Digits Backward).
whereas the last two age groups contain data from 75
The Spatial Span subtest also consists of two parts:
participants each. For the following analyses, the first
Spatial Span Forward and Spatial Span Backward. For
Downloaded By: [Washington University in St Louis] At: 23:37 21 April 2010

two age groups (ages 16–19 years) were excluded,

each part, the examiner taps a series of cubes at the rate
leaving a total sample size of 1,050 (ages 20–89 years)
of about one cube per second. Following presentation,
partitioned into 11 groups. Potential participants were
the examinee either taps the cubes in the same order as
screened with a medical and psychiatric self-report
the examiner (Spatial Span Forward) or in reverse order
questionnaire (see Table 2.1 in the WAIS-III technical
(Spatial Span Backward). For both Spatial Span
manual; Psychological Corporation, 1997a).
Forward and Spatial Span Backward, the test begins
Each age group was selected to be representative of
with series of two cubes and continues to a maximum of
the United States population with respect to sex, level of
eight cubes. Examinees are given two trials at each
education, race=ethnicity (White, African-American,
series length, and the test continues until both trials of a
Hispanic, or Other), and geographic region (Northeast,
series length are failed. One point is awarded for each
North Central, South, and West). A stratified sampling
trial the examinee answers correctly. The maximum
plan, based on 1995 data from the United States Bureau
possible score for the Spatial Span subtest is 32 (16
of Census, was used to ensure representativeness
points each for Spatial Span Forward and Backward).
(Psychological Corporation, 1997a). For example, in
In Letter-Number Sequencing, the test administrator
the population at large, 50.1% of young adults aged 25–
says a series of alternating numbers and letters at the
29 years and 58.0% of older adults aged 75–79 years
rate of about one per second. Following presentation,
were female, whereas, in the WMS-III sample the
the examinee first reports the numbers in ascending
percentages for the corresponding age groups were 50.0
numerical order, then reports the letters in alphabetical
and 58.0%. More specifically, all of the age groups in
order. The test begins with series of two items (one
the WMS-III sample were 50% female except for those
number and one letter) and continues to a maximum of
with participants 65 years or older. For those older age
eight items (four numbers and four letters). Examinees
groups, the percentage of female participants was based
are given three trials at each series length, and continue
on corresponding census data. Similarly, for each age
until all three trials of a series length are failed. Because
group the percentages of participants with different
of differences in administration, the maximum possible
levels of education (i.e., 8 or less, 9–11, 12, 13–15, and
raw score for Letter-Number Sequencing is lower than
16 or more years of education) in the WMS-III sample
that for either Digit or Spatial Span. The maximum
closely approximated the corresponding percentages in
possible score for Letter-Number Sequencing is 21.
the United States population (compare Tables 2.5 and
2.7 in the WAIS-III technical manual; Psychological
Analysis
Corporation, 1997a). Notably, age differences in the
For the total Digit Span, total Spatial Span, Spatial Span
distribution of levels of education reflect historical
Forward, Spatial Span Backward, and Letter-Number
trends in educational access (Snyder, 1993).
Sequencing subtests, weighted regression analyses
were based on reconstructed distributions of raw scores
Subtests for each age group calculated from the WMS-III
The Digit Span subtest of the WMS-III consists of two standardization tables (Psychological Corporation,
parts: Digits Forward and Digits Backward. For each 1997b). For Digits Forward and Backward, separate
part, the test administrator says a series of numbers at scaled scores are not included in the standardization
the rate of about one per second. Following presenta- tables. The tables, however, do provide the cumulative
tion, the examinee either repeats the numbers in the distributions of the longest sequence of digits that
order they were presented (Digits Forward) or in participants could recall both forwards and backwards,
reverse order (Digits Backward). For both Digits as well as the cumulative distributions for each age
 AGE DIFFERENCES IN MEMORY SPAN MEASURES 23

group of the within-subject difference between the

longest span forward and longest span backward. These
data were used to compare forward and backward digit
spans. Because the present analyses were based on
standardization tables rather than the raw data, regres-
sion analyses treated the data from each measure as
independent.
The WMS-III standardization tables (Psychological
Corporation, 1997b) for the total Digit Span, total
Spatial Span, Spatial Span Forward, Spatial Span
Backward, and Letter-Number Sequencing subtests
present the scaled score equivalents of subtest raw
scores normalized for each age group. The scaled
scores are normalized so that, for each age group, the
mean equals 10 and the standard deviation equals 3. In
Downloaded By: [Washington University in St Louis] At: 23:37 21 April 2010

order to reconstruct the original distribution of raw

scores for each age group, we first matched each scaled
score to its corresponding raw score(s) and then
determined the probability of each range of raw scores
using a cumulative probability table for a normal
distribution. The midpoint of each range of raw scores
(or the corresponding range of z scores) was used as the
dependent variable in the regression equations, and the
product of the probability of each range and the number
of people in the age group was used as the regression
weight. The independent variable was the midpoint of
the age range for each of the 11 groups describe above
(e.g., the 20–24 group had ages ranging from 20 years
and 0 days to 24 years and 364 days with a midpoint of
22.5 years).
The maximum and mean scores for the Letter-
Number Sequencing subtest are markedly different
from those for the Digit Span and Spatial Span subtests.
Therefore, analyses comparing Letter-Number Sequen-
cing with Digit Span and Spatial Span were conducted
using z scores rather than raw scores as the independent
variable. For each subtest, the distributions of z scores
for the various age groups were computed using the Fig. 1. Forward and backward memory spans across all
mean and standard deviation of the 20–24-year-old age age groups. The top panel shows the mean
group (estimated from the reconstructed raw score longest series recalled for Digit Span Forward
distribution) as a reference. and Backward. The bottom panel shows the
mean raw scores for Spatial Span Forward and
Backward. In both panels, the error bars
indicate standard errors.
RESULTS

Means for the various measures at each age,

estimated from the reconstructed raw score no significant difference between the slopes for
distributions, are pictured in Figures 1–3, and Digit Span Forward and Digit Span Backward,
regression equations and correlation coefficients t(2096) ¼ 0.30, p > .05. Similarly, there was no
are reported in Table 1. As may be seen in significant difference in the slope of the regres-
Figure 1, there is no evidence suggestive of a sion lines for Spatial Span Forward and Spatial
differential decrease in forward versus backward Span Backward, t(2096) ¼ 1.21, p > .05. In addi-
digit span with age for either Digit Span (top tion, regressing the Forward-Backward difference
panel) or Spatial Span (bottom panel). There was for Digit Span on age indicated that none of the
 24 JOEL MYERSON ET AL.
Downloaded By: [Washington University in St Louis] At: 23:37 21 April 2010

Fig. 3. Mean z scores for Digit Span, Spatial Span, and

Letter-Number Sequencing across all age
groups. z Scores were computed using the mean
Fig. 2. Mean total raw scores for Digit Span and Spatial and standard deviation of the 20–24-year-old
Span across all age groups. The error bars age group as a reference.
indicate standard errors.

variance in the difference scores could be t(2096) ¼ 5.09, p < .001. Similar results were
accounted for by age (r2 ¼ .00). obtained when age groups that had more females
Some of the subtests, however, do show evi- than males were not included in the analysis (i.e.,
dence suggestive of a differential decrease. Nota- when the age range was restricted to 20–64 years).
bly, Spatial Span appears to decrease more as Finally, Letter-Number Sequencing scores
a function of age than does Digit Span, as may appear to show a pattern of decrease different
be seen in Figure 2. The slope of the regression from that of either Digit Span or Spatial Span,
for the total raw Spatial Span score (0.085 as may be seen in Figure 3. As noted previously,
points per decade) was significantly steeper than z scores based on the 20–24-year-old age
the slope of the regression for the total raw group were used to compare Letter-Number Se-
Digit Span score (0.054 points per decade), quencing to the other tests. Therefore, constrained

Table 1. Regression Equations and Correlation Coefficients for Memory Span Measures.

Equation r

Digit Span total raw score y ¼  0.054 x þ 19.18  .27

Spatial Span total raw score y ¼  0.085 x þ 19.18  .47
Digits Forward longest series y ¼  0.014 x þ 7.16  .23
Digits Backward longest series y ¼  0.015 x þ 5.44  .22
Spatial Forward raw score y ¼  0.040 x þ 10.04  .40
Spatial Backward raw score y ¼  0.044 x þ 8.94  .42
Digit Span z score y ¼  0.011 (x  22.5)  .26
Spatial Span z score y ¼  0.022 (x  22.5)  .45
Letter-Number z score y ¼  0.019 (x  22.5)  .45
Note. All reported correlations are significantly different from zero (p < .05).
 AGE DIFFERENCES IN MEMORY SPAN MEASURES 25

regression analyses (that excluded the 20–24- evidence of a specific age-related deficit in back-
year-old data) were conducted to determine if ward spatial span relative to forward spatial span.
the slope of the decrease in Letter-Number The results of a meta-analysis of experimental
Sequencing was different from that of either studies reported by Babcock and Salthouse (1990)
Digit Span or Spatial Span. That is, all regressions are sometimes taken as indicating greater age-
were forced through the point at which age ¼ 22.5, sensitivity of backward spans. However, the two
z ¼ 0.0. The slopes for Digit Span, Spatial Span, studies examining cross-sectional data from large,
and Letter-Number Sequencing were, respec- representative samples of people (i.e., the current
tively, 0.011, 0.022, and 0.019 z score units per study and that of Gregoire and Van der Linden)
decade. Results indicated that the slope of the provide no evidence that older adults show a
Letter-Number Sequencing regression line was specific deficit in backward spans compared to
significantly steeper than that for Digit Span, forward spans. These results are of particular
t(1896) ¼ 7.26, p < .001, and slightly less steep interest because they contradict the established
Downloaded By: [Washington University in St Louis] At: 23:37 21 April 2010

than that for Spatial Span, t(1896) ¼ 2.11, p < .05. belief – exemplified by the quote in the introduc-
The graph in Figure 3 reveals some curvature tion – that older adults are differentially impaired
in the relationship between Letter-Number in backward span.
Sequencing and age. This pattern was confirmed The present findings regarding Digit and
by polynomial regression, which indicated that Spatial Span are consistent with the results of
the regression of Letter-Number Sequencing on several previous studies which reported that older
age has a significant quadratic component, adults are differentially impaired in visuospatial
F(1,947) ¼ 32.08, p < .001. Digit Span also shows memory span compared with verbal memory
a significant quadratic component, F(1,947) ¼ span (Dolman et al., 1997; Jenkins et al., 2000;
6.69, p < .001; Spatial Span, however, does not, Myerson et al., 1999). They differ, however, from
F(1,947) ¼ 1.22, p > .05. the results of other studies (Salthouse, 1995;
Salthouse et al., 1988). Procedural and sampling
differences may be responsible for the conflicting
DISCUSSION nature of previous evidence. Regardless of the
correct interpretation of past discrepancies, how-
To summarize, analyses of cross-sectional data ever, the present findings based on a large norma-
from the WMS-III (Psychological Corporation, tive sample may help resolve the question of
1997b) provided no support for the common whether aging affects visuospatial memory span
belief that backward spans show greater age- more than verbal memory span. Comparisons
related declines than forward spans: There was no based on raw scores and those based on z scores
evidence of a specific age-related deficit in back- (in young adult standard deviation units) both
ward memory span for either verbal material (Digit revealed that slopes for the regression of spatial
Span) or visuospatial material (Spatial Span). span measures on age were significantly more
However, the slope of the regression of Spatial negative than slopes for corresponding digit span
Span on age was significantly more negative measures.
than the slope for Digit Span. Scores on the Park et al. (2002) recently reported the results
Letter-Number Sequencing subtest also decreased of a study of 345 adults aged 20–92 years that
more as a function of age than scores on the examined age-related differences in various
Digit Span subtest, but scores on the Spatial Span aspects of verbal and visuospatial memory includ-
subtest showed the largest decrease. ing short-term memory, working memory, and
The present findings regarding a lack of a long-term memory. Of particular relevance to
specific age-related deficit in backward digit the present study are Park et al.’s results for
span, relative to forward digit span, are consistent short-term memory, which was represented in
with those reported by Gregoire and Van der their study by Digit Span Forward and Backward
Linden (1997) for the French WAIS-R norma- and Spatial Span Forward and Backward (which
tive sample. In addition, our findings reveal no they referred to as Corsi Blocks although they
 26 JOEL MYERSON ET AL.

note that it was taken from the WMS-III). They processing to a verbal task (e.g., digit span) would
converted each of these measures to z scores result in an increased age deficit. Nevertheless,
based on the SD of the entire sample and then the present results failed to provide evidence
regressed these scores on age. suggestive of a differential decline in Digits Back-
The Park et al. (2002) regression results were ward relative to Digits Forward.
similar to those of the present study in that for The present results regarding age-related dif-
both digits and spatial locations, the slopes for the ferences in Letter-Number Sequencing may be
forward and backward scores were approximately contrasted with those for Digits Forward and
equal. In both studies, moreover, the slopes for the Digits Backward. Letter-Number Sequencing per-
spatial span measures were nearly twice as steep formance clearly decreased at a greater rate,
as those for the digit span measures. Specifically, particularly after age 60, than either forward or
Park et al. reported that the slopes for Spatial Span backward digit span. Some researchers (e.g.,
Forward and Backward were  0.025 and Dobbs & Rule, 1989; Engle et al., 1999) have
Downloaded By: [Washington University in St Louis] At: 23:37 21 April 2010

 0.021 (both standard errors ¼ 0.002) whereas suggested that backward span tasks require rela-
the slopes for Digit Span Forward and Backward tively little executive processing. Arguably, cate-
were  0.013 and  0.012 (both standard gorizing items as digits and letters and then
errors ¼ 0.003). Although they reported no sig- reordering them (i.e., numerically and alphabeti-
nificant differences in slope between any of their cally) does require substantial executive proces-
digit span and spatial span measures, this may be sing. If so, the present findings are consistent with
because of the family-wise correction for the the hypothesis of specific age-related deficits in
number of statistical tests performed, given the executive functions (e.g., Moscovitch & Winocur,
size of the difference and the relatively small 1992; West, 1996). Relatively little research,
standard errors. however, has been done on the Letter-Number
To facilitate comparison with the Park et al. Sequencing task since it was recently introduced
(2002) findings, we converted the WMS-III data (Psychological Corporation, 1997a, 1997b), and
to z scores based on the SD of the entire WMS-III tests of the hypothesis that this task makes special
sample. Regression analyses based on these trans- demands on the ability to manipulate information
formed data yielded values for the slopes that online are needed.
were very similar to those reported by Park et al. In conclusion, a single mechanism account
Specifically, the slopes for the transformed WMS- does not appear to be sufficient to explain the
III data were  0.025 for Spatial Span (Forward overall pattern of age-related differences in mem-
plus Backward) and  0.013 for Digit Span ory span performance on the WMS-III. Digit
(Forward plus Backward). The difference Span and Spatial Span both require recalling
between these Digit and Spatial Span slopes was items in the original and reverse order, yet Spatial
statistically significant: t(2096) ¼ 8.08, p > .001. Span decreases more with age than Digit Span.
Moreover, the similarity of these slopes to those Moreover, although both Digit Span and Letter-
reported by Park et al. clearly demonstrates the Number Sequencing use verbal material, the latter
replicability of the finding that spatial spans decreases more with age than the former. With
decrease more with age than digit spans. respect to the two more age-sensitive subtests,
It has been suggested that the extra processing Spatial Span showed a linear decrease with age
involved in backward digit span is spatial in whereas the decrease in Letter-Number Sequen-
nature (e.g., Costa, 1975; Hoshi et al., 2000; cing was significantly curvilinear. Taken together,
Rapport et al., 1994; Weinberg et al., 1972), but these findings suggest that at least two mechan-
the normative WMS-III data analyzed here pro- isms may be involved in determining the time
vide little support for the hypothesis. Aging course of age-related decline in memory span.
affects visuospatial processing more than verbal One of these mechanisms differentially affects the
processing (e.g., Hale & Myerson, 1996; Jenkins storage of spatial information and may begin
et al., 1999; Lawrence, Myerson, & Hale, 1998), affecting working memory performance relatively
and thus one might expect that adding spatial early in adulthood; the other mechanism, which
 AGE DIFFERENCES IN MEMORY SPAN MEASURES 27

has pronounced effects only in later adulthood, Jenkins, L., Myerson, J., Joerding, J.A., & Hale, S.
may be tentatively attributed to a decline in (2000). Converging evidence that visuospatial
cognition is more age-sensitive than verbal cogni-
executive aspects of working memory.
tion. Psychology and Aging, 15, 157–175.
Lawrence, B.M., Myerson, J., & Hale, S. (1998).
Differential decline of verbal and visuospatial
ACKNOWLEDGEMENTS processing speed across the adult life span.
Aging, Neuropsychology, and Cognition, 5,
The authors are at the Department of Psychology, 129–146.
Washington University, Saint Louis, MO, USA. Lisa Moscovitch, M., & Winocur, G. (1992). The neurop-
Emery was supported by National Institute on Aging sychology of memory and aging. In F.I.M. Craik
training grant AG00030. et al. (Eds.), The handbook of aging and cognition
(pp. 315–372). Hillsdale, NJ: Erlbaum.
Myerson, J., Hale, S., Rhee, S.H., & Jenkins, L. (1999).
REFERENCES Selective interference with verbal and spatial work-
Downloaded By: [Washington University in St Louis] At: 23:37 21 April 2010

ing memory in young and older adults. Journals

Babcock, R.L., & Salthouse, T.A. (1990). Effects of of Gerontology: Psychological Sciences, 54(B),
increased processing demands on age differences 161–164.
in working memory. Psychology and Aging, 5, Park, D.C., Lautenschlager, G., Hedden, T., Davidson,
421–428. N.S., Smith, A.D., & Smith, P.K. (2002). Models of
Baddeley, A.D. (1986). Working memory. Oxford: visuospatial and verbal memory across the adult life
Clarendon Press=Oxford University Press. span. Psychology and Aging, 17, 299–320.
Costa, L.D. (1975). The relation of visuospatial Psychological Corporation. (1997a). WAIS-III – WMS-
dysfunction to digit span performance in patients III technical manual. San Antonio: Harcourt Brace
with cerebral lesions. Cortex, 11, 31–36. & Company.
Dempster, F.N. (1992). The rise and fall of the Psychological Corporation. (1997b). WMS-III ad-
inhibitory mechanism: Toward a unified theory of ministration and scoring manual. San Antonio:
cognitive development and aging. Developmental Harcourt Brace & Company.
Review, 12, 45–75. Rapport, L.J., Webster, J.S., & Dutra, R.L. (1994). Digit
Dobbs, A.R., & Rule, B.G. (1989). Adult age span performance and unilateral neglect. Neuropsy-
differences in working memory. Psychology and chologia, 32, 517–525.
Aging, 4, 500–503. Salthouse, T.A. (1994). The aging of working memory.
Dolman, R., Roy, E.A., Dimeck, P.T., & Hall, C.R. Neuropsychology, 8, 535–543.
(2000). Age, gesture span, and dissociation among Salthouse, T.A. (1995). Differential age-related influ-
component subsystems of working memory. Brain ences on memory for verbal-symbolic information
and Cognition, 43, 164–168. and visual-spatial information? Journal of Geron-
Engle, R.W., Laughlin, J.E., Tuholski, S.W., & tology: Psychological Sciences, 50B, 193–201.
Conway, A.R.A. (1999). Working memory, short- Salthouse, T.A., Kausler, D.H., & Saults, J.S. (1988).
term memory, and general fluid intelligence: A Investigation of student status, background vari-
latent variable approach. Journal of Experimental ables, and feasibility of standard tasks in cognitive
Psychology: General, 128, 309–311. aging research. Psychology and Aging, 3, 29–37.
Gregoire, J., & Van der Linden, M. (1997). Effects of Snyder, T.D. (1993). 120 years of American Education:
age on forward and backward digit spans. Aging, A statistical portrait. Washington, DC: U.S. Depart-
Neuropsychology, and Cognition, 4, 140–149. ment of Education.
Hale, S., & Myerson, J. (1996). Experimental evidence Verhaeghen, P., Marcoen, A., & Goossens, L. (1993).
for differential slowing in the lexical and nonlexical Facts and fiction about memory aging: A quantita-
domains. Aging, Neuropsychology, and Cognition, tive integration of research findings. Journal of
3, 154–165. Gerontology: Psychological Sciences, 48, 157–171.
Hoshi, Y., Oda, I., Wada, Y., Ito, Y., Yamashita, Y., Oda, Weinberg, J., Diller, L., Gertsmann, L., & Schulman, P.
M., Ohta, K., Yamada, Y., & Tamura, M. (2000). (1972). Digit span in right and left hemiplegics.
Visuospatial imagery is a fruitful strategy for the Journal of Clinical Psychology, 28, 361.
digit span backward task: A study with near-infrared West, R.L. (1996). An application of prefrontal cortex
optical tomography. Cognitive Brain Research, 9, function theory to cognitive aging. Psychological
339–342. Bulletin, 120, 272–292.