Psychonomic Bulletin & Review
2010, 17 (2), 193-199
doi:10.3758/PBR.17.2.193

Brief Reports
Expanding the mind’s workspace:
Training and transfer effects with
a complex working memory span task
Jason M. Chein and Alexandra B. Morrison
Temple University, Philadelphia, Pennsylvania

In the present study, a novel working memory (WM) training paradigm was used to test the malleability of
WM capacity and to determine the extent to which the benefits of this training could be transferred to other
cognitive skills. Training involved verbal and spatial versions of a complex WM span task designed to emphasize
simultaneous storage and processing requirements. Participants who completed 4 weeks of WM training demonstrated significant improvements on measures of temporary memory. These WM training benefits generalized to
performance on the Stroop task and, in a novel finding, promoted significant increases in reading comprehension. The results are discussed in relation to the hypothesis that WM training affects domain-general attention
control mechanisms and can thereby elicit far-reaching cognitive benefits. Implications include the use of WM
training as a general tool for enhancing important cognitive skills.

Practice can yield remarkable levels of achievement
but little generalization. For example, Ericsson and Chase
(1982) reported the illustrative case of a college student
who, following many hours of practice on a digit-span task
(an often used measure of short-term memory [STM]),
could successfully recall over 80 randomly ordered digits.
However, the individual was limited to a typical STM span
of about seven items for other, even closely related, memoranda. The specificity of the improvements observed in
that and other classical studies of skill acquisition and
expertise have led many to conclude that the benefits of
practice on a given task do not generally extend into other
realms of performance (Chase & Simon, 1973; Engle &
Bukstel, 1978). By this account, although individuals may
exhibit innate differences in certain domain-general capacities, training strategies seeking to promote superior
performance through influence on these general capacities would be destined to fail.
Several recent studies, however, have invigorated an interest in the plausibility of using repetitive mental exercise
to enhance one domain-general ability—working memory
(WM)—and, in so doing, to concurrently improve performance in other cognitive skills (Jaeggi, Buschkuehl,
Jonides, & Perrig, 2008; Klingberg et al., 2005; Persson &
Reuter-Lorenz, 2008). Training regimens used in these and
other recent efforts appear to have targeted domain-general
processes that individuals utilize to broadly support complex cognition. For example, in a study by Verhaeghen,
Cerella, and Basak (2004), WM training was shown to in-

fluence recall from WM by expanding the capacity of attention. Likewise, others have demonstrated that WM training
can impact domain-­general cognitive control mechanisms
(Klingberg et al., 2005; Klingberg, Forssberg, & Westerberg, 2002), interference resolution processes (Persson &
Reuter-Lorenz, 2008), WM updating processes (Dahlin,
Neely, Larsson, Bäckman, & Nyberg, 2008), and even general fluid intelligence (Jaeggi et al., 2008).
Although it provides a very promising foundation, the
small corpus of existing WM training studies is limited in
two important ways. First, prior demonstrations of transfer
included measures closely related to those used to estimate
WM itself, rather than to more distant tasks. To address this
limitation, we sought to examine a broader battery of measures with the expectation that other tasks known to correlate with individual differences in WM capacity might
also benefit from WM training. Second, previous studies
utilized training tasks that are not commonly employed
in the basic behavioral or psychometric literatures (e.g.,
atypical variants of the n-back task in Verhaeghen et al.
[2004] and Jaeggi et al. [2008]; a battery of video-gamelike tasks, such as that used in Klingberg et al., 2005).
Thus, previous findings are somewhat disjointed from the
larger behavioral literature and offer only limited insights
into the specific WM mechanisms influenced by training
(e.g., encoding, strategic processing, updating, etc.).
Encouraged by findings regarding the malleability of
WM, in the present study, we test a novel approach to increasing WM capacity through repeated practice with an

J. M. Chein, jchein@temple.edu

193

© 2010 The Psychonomic Society, Inc.

194     Chein and Morrison
adaptive complex WM (CWM) span task and examine the
generalizability of the resulting WM improvements into
the broader landscape of cognitive ability. By training participants using a CWM span task, we hoped to bridge the
WM training literature to the wider body of empirical work
that implicates WM in complex cognition. This variety of
WM task has served as the cornerstone of the psychometric (individual differences) literature on WM for over 30
years and is arguably the most reliable and widely studied
predictor of complex cognition (Conway et al., 2005). The
hallmark characteristic of CWM tasks is the interweaving
of storage (of the test items) and processing (the secondary decision-making task) components, which places a
premium on the participant’s ability to coordinate maintenance in the face of a concurrent distraction.
We anticipated that, because CWM tasks place a strong
demand on mechanisms linked to domain-general attention control (Engle & Kane, 2004), a training paradigm
built around this task would result in both increases of
WM span and more far-reaching benefits. Accordingly, the
aims of the present study were twofold. First, we sought
to examine the impact of a novel training paradigm, an
adaptive version of a CWM span task. Second, we sought
to expand the known boundaries of generalization by administering a battery of cognitive skills assessments.
Method
Participants
Forty-two Temple University undergraduates (25 female) completed the study and were compensated monetarily. Trained participants had a mean age of 20.1 years, and control participants had a

mean age of 20.6 years. All participants were native English speakers with no prior events or illnesses that we expected to have an
impact on WM performance.
Cognitive Assessments
A battery of computerized and handwritten tests assessed participants’ cognitive skill levels before and after WM training. The
battery of measures was chosen for its previously documented relationship to individual differences in WM ability (see, e.g., Kane
et al., 2004). Temporary memory measures included verbal and
spatial STM and verbal and spatial CWM span. The verbal and
spatial STM tasks assessed immediate short-term recall of serially
presented (­1-sec stimulus, 1-sec intertask interval) letters and locations, respectively. A schematic of each CWM measure is shown in
Figure 1. The verbal CWM task also involved recall of letters, but
each to-be-remembered letter was preceded by 4 sec of repeated,
participant-paced, lexical decisions. The spatial CWM task used the
same timing parameters but tested memory for locations presented
with interleaved symmetry decisions. Participants were given unlimited time to attempt recall, and, after each trial, participants received
feedback indicating both the number of correctly recalled items and
the overall accuracy of decision making on the intervening distractor
task. For each temporary memory measure, participants were first
asked to attempt recall of three test items. Following two successful
trials at a given length, the number of test items was increased by
one. The number of test items increased until participants failed to
correctly recall two successive trials at a given list length. A participant’s span was then defined as the maximum list length for which
all items were recalled in the correct serial order.
A logical reasoning assessment included two tests of verbal reasoning (ETS’s Inference and Nonsense Syllogisms tests) and two
tests of spatial reasoning (ETS’s Surface Development and Paper
Folding tests). Cognitive control was measured using the Stroop
Color–Word Interference Test (3 blocks of 60 trials each, 50% incongruent). General fluid intelligence was measured with Raven’s
Advanced Progressive Matrices (APM), using only the odd- or

PROCESSING (4 sec) STORAGE (1 sec)

Verbal
CWM

blick

Number of repetitions (#)
based on trial list length

Spatial
CWM

m

RECALL

P
X
D
K

S
Z
F
L

T V
B C
G J
M N

FEEDBACK
Recall:
6 out of 8
Decisions:
92%

repeat # times

Recall:
6 out of 8
Decisions:
92%

Variable number of
processing decisions
(participant paced)
Figure 1. Schematic depiction of the training tasks. Each of the training sessions included 16 trials of the
verbal complex working memory (CWM) task and 16 trials of the spatial CWM task. Participants began
each session with a span of four test items. One additional test item was presented following 2 consecutive
successful trials, and one fewer item was presented following 2 consecutive unsuccessful trials.

Working Memory Training and Transfer     195
25

Percentage CWM Span Increase

% improvement, spatial

20

% improvement, verbal

15
10
5
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
–5

Training Session

Figure 2. Percentage working memory span increases over 20 training
sessions for verbal and spatial complex working memory (CWM) tasks.
Results are expressed as a percentage increase relative to performance
on the first day of training.

even-numbered problems for a given test session (counterbalanced
across participants) and allotting 45 min for test administration.1
Finally, reading comprehension was measured with the Nelson–
Denny Reading Test (Forms G and H used in counterbalanced fashion across sessions), using standard test procedures and allowing
20 min for test completion.
Training Paradigm
WM training utilized the verbal and spatial CWM measures used
for pre- and posttraining cognitive assessment, modified only in that
the training tasks were adaptive to the participant’s performance
level (i.e., increased or decreased the number of test items on the
basis of participant performance). Participants began each session
of training with a list length of four test items (letters or locations); if
all four items were correctly recalled for 2 trials in a row and at least
75% of the lexicality or symmetry decisions were made correctly on
these trials, the list length was increased to five items. This process
of increasing the stimulus count by one following 2 correct trials
continued throughout the training session. Likewise, 2 successive
incorrect trials caused the list length to be reduced by one item. Each
training session lasted between 30 and 45 min and included 16 trials of the verbal CWM task and 16 trials of the spatial CWM task.
The overall length of the session was variable, because trial onsets
were participant paced and average trial length varied according to
participant performance (trials with a larger number of test items
took longer to complete).
Procedure
At the onset of the study, all participants completed a battery of
standard cognitive tests in a laboratory located on the Temple University campus. Following the initial assessment, half of the participants were randomly selected to participate in the subsequent 4-week
WM training regimen; the other half were assigned to an untrained
control condition. Training required participants to complete the
WM training exercises on 5 days of each week over the duration of
4 weeks. Training sessions were carried out using software that participants downloaded onto their personal computers (1 participant
had difficulty downloading the training software and was switched
to the untrained group), and trained participants were required to log
and submit their daily training results electronically. The activities of
untrained participants were not regulated during the 4-week interval.
Participants in both groups returned 4–5 weeks after their initial assessment for a second assessment, which included alternate versions
of the tests used in the first assessment.

Results
A first step in analysis was to track participants’ CWM
span improvements over the 20 training sessions. Mean
­session-to-session performance over the 20 sessions indicated a steady rate of improvement across the 4 weeks
of training for both verbal and spatial CWM, as shown in
Figure 2.
Analyses next focused on comparisons of participants’ cognitive skill levels before and after the training
interval for the following measures: temporary memory
(verbal CWM, spatial CWM, verbal STM, spatial STM),
ETS reasoning tasks (inference, nonsense syllogisms,
paper folding, surface development), cognitive control
(Stroop), general fluid intelligence (Raven’s APM), and
reading comprehension (Nelson–Denny). Through these
comparisons, we sought to answer two central questions
about the impact of WM training: (1) Does WM training
lead to an increase in the capacity of temporary memory?
and (2) Do the benefits of WM training transfer to specific
untrained tasks?
Since the temporary memory and ETS reasoning assessments each included performance on multiple tests,
we created composite scores for each participant. The
composite was calculated by normalizing the first and second assessment scores relative to first assessment performance and then averaging the normalized scores for each
of the tasks in the composite. For all assessed measures,
paired-samples t tests were used to compare first and second assessment scores within trained and control groups,
and an independent-samples t test was used to evaluate the
difference in the improvements attained by trained versus
untrained participants (Table 1). We entered the study with
the directional hypothesis that training would improve test
scores; thus, all t tests were one-tailed.
For the temporary-memory composite, both control and
trained participants showed significant test–retest improvements (i.e., significantly increased performance on the sec-

196     Chein and Morrison
Table 1
Task Performance Scores in Cognitive Measures Before and After the Training Period

Measure
Temporary memory
Cognitive control
Reading comprehension
ETS reasoning battery
Fluid intelligence

Control Group
Test–Retest
t
df
p Value
1.96
21
.032
1.81
21
.043
0.87
20
.200
2.09
21
.024
0.31
21
.380

Trained Group
Test–Retest
t
df
p Value
7.06
19
,.005
2.77
19
.006
2.36
18
.015
4.15
19
,.005
0
19
.500

ond assessment). It is important to note that the magnitude
of the improvement was significantly larger among trained
participants, as confirmed by an independent-­samples t test
contrasting the size of the training effect (i.e., difference in
performance from first to second assessment) for trained
versus control participants (Table 1). A repeated measures
ANOVA was used to further establish the strong influence
of training on temporary memory capacity and revealed a
highly significant interaction between assessment day and
training group [F(1,40) 5 7.483, p , .0005, η p2 5 .428].
Since training was composed of verbal and spatial CWM
tasks, we further examined trained participants’ performance on these specific measures. Groupwise gains were
statistically significant for each measure [t(19) . 4.64,
p , .005], and, of the 20 trained participants, 18 showed
increased verbal CWM performance, and 15 showed increased spatial CWM performance.
Having found a significant impact of training on WM,
we next examined transfer from training to measures that
assessed more disparate cognitive abilities. Since generalization presupposes improvement on the trained tasks,
we considered generalization effects both in the entire
group of 20 trained participants and in the subgroup of
15 “successfully” trained participants, who demonstrated
improved performance on spatial CWM, the more stringent of the two WM measures.
On our measure of fluid intelligence (Raven’s APM), no
significant improvement was recorded in either the trained

group (pretraining, M 5 12.2, SE 5 0.66; posttraining, M 5
12.2, SE 5 0.65) or the control group (pretraining, M 5 12.4,
SE 5 0.72; posttraining, M 5 12.2, SE 5 0.57) (see Table 1).
Examining only the subgroup of successfully trained participants, we similarly found no improvement in Raven’s APM
performance (pretraining, M 5 13.4, SE 5 0.51; posttraining, M 5 13.2, SE 5 0.55). Thus, training does not appear
to have affected performance on this measure.
Results for the ETS reasoning composite are also included in Table 1. Trained participants exhibited small,
but statistically significant, improvements (pretraining,
M 5 20.19, SE 5 0.12; posttraining, M 5 0.20, SE 5
0.13). However, control participants’ test–retest improvements (pretraining, M 5 0.17, SE 5 0.14; posttraining,
M 5 0.37, SE 5 0.15) were also statistically significant.
Despite slightly larger gains among trained participants,
an independent-samples t test directly contrasting improvements for the trained versus control groups was nonsignificant. Gains observed on these reasoning measures
may, thus, reflect only test–retest improvements and not a
true effect of transfer from training. This interpretation is
further supported by the finding that ETS reasoning improvements were no larger for the subset of successfully
trained participants (pretraining, M 5 20.18, SE 5 0.15;
posttraining, M 5 0.16, SE 5 0.16).
By contrast, for both Stroop and reading comprehension
performance, trained participants exhibited improvements
following training that exceeded the improvements found

Stroop Congruency Effect

Reading Comprehension
245

Trained
Untrained

Nelson–Denny Scaled Score

Inverted Stroop Effect (msec)

0
–20

Differences Between Groups
t
df
p Value
Cohen’s d
4.49
40
,.005
1.420
1.81
40
.039
0.570
1.80
38
.040
0.580
1.39
40
.090
0.439
0.24
40
.410
0.076

–40
–60
–80
–100
–120
–140
–160

Trained
Untrained

240

235

230

225
Pretest

Posttest

Pretest

Posttest

Figure 3. Pre- and posttest performance in the Stroop and Nelson–Denny reading comprehension measures for matched trained and untrained participants. The Stroop congruency
effect is inverted to better illustrate training gains.

in control participants (Figure 3). Both control and trained
participants showed significant test–retest improvements
in Stroop performance (i.e., reduction of the Stroop congruency effect). An independent-samples t test of the
change scores indicated significantly greater improvement
among trained participants than among controls (Table 1),
although a repeated measures ANOVA did not yield a significant training group (trained, control) 3 assessment day
(pretraining, posttraining) interaction [F(1,40) 5 2.39, p 5
.13, η 2p 5 .056]. Still, confidence in the impact of training on Stroop performance is encouraged by the finding
that reductions in the Stroop congruency effect were more
substantial among the subgroup of successfully trained
participants, and a repeated measures ANOVA based on
this training subgroup produced a marginally significant
interaction between training group (successfully trained,
control) and assessment day (pretraining, posttraining)
[F(1,35) 5 4.03, p 5 .053, η p2 5 .103].
Effects of training on reading comprehension2 are summarized in Table 1 and Figure 3. Examination of pre- and
posttraining reading comprehension performance showed
significant test–retest improvements for trained participants, but not for control participants. A direct contrast of
reading comprehension improvements in trained versus
untrained groups was also significant. Repeated measures
ANOVAs showed a marginally significant interaction between training group and assessment day when data from
the entire training group were considered [F(1,38) 5 3.30,
p 5 .077, η 2p 5 .08]; and, again, transfer from training
was more strongly indicated in the subset of successfully trained participants [F(1,34) 5 6.94, p 5 .013,
η 2p 5 .174].
Our findings, which suggest an effect of WM training on Stroop and reading comprehension performance,
can be informed further by a closer examination of inter­
individual differences in the influence of training on WM
capacity itself. As we have already shown above, participants for whom training promoted an increase in WM span
(i.e., successfully trained participants) exhibited stronger
improvements for each transfer measure. Taking a more
nuanced view, one might further expect to find a correlation between WM improvements and gains on the transfer
measures, with greater improvements in those for whom
the WM training was most effective (i.e., produced larger
increases in WM span) and weaker improvements in those
for whom it was relatively less effective. Despite our relatively small sample size, we accordingly tested correlations between WM span increases and both Stroop and
reading comprehension improvements. For the Stroop test,
the correlations were weak [r(18) 5 2.01 for verbal WM;
r(18) 5 .12 for spatial WM] and not significant. However,
the results from the reading comprehension measure were
more encouraging. The correlation between verbal WM
increases and Nelson–Denny improvements was modest
[r(18) 5 .24, p 5 .13], and there was a strong and statistically significant relationship between trained participants’
spatial CWM span increases and reading comprehension
improvements [r(18) 5 .49, p , .005], as is shown in Figure 4. Among control participants, there was no observed

Change in Reading Comprehension

Working Memory Training and Transfer     197
40
30
20
10
0
–10
–20
–2

0

2

4

6

8

Change in Spatial CWM Span

Figure 4. Individual differences in Nelson–Denny reading comprehension improvements are correlated with increases in spatial
complex working memory (CWM) span for the group of trained
participants. Change scores were computed as the differences
between the first and second assessments.

relationship between changes in WM span and Stroop
performance [r(22) 5 .19 for verbal WM; r(22) 5 2.10
for spatial WM] or between changes in CWM span and
reading span [r(20) 5 2.028, p 5 .90 for verbal CWM;
r(20) 5 .097, p 5 .67 for spatial CWM].
Discussion
Our objectives in the study were twofold. First, we assessed the efficacy of a novel WM training paradigm, an
adaptive version of a CWM span task that could serve
as a bridge between the WM training and psychometric
literatures. Consistent with this aim and with earlier studies (Jaeggi et al., 2008; Klingberg et al., 2005; Klingberg
et al., 2002; Verhaeghen et al., 2004), we found that trained
participants improved significantly more than did controls
on temporary memory measures.
Second, we aimed to examine the scope of transfer
from training. If training influences specific processes
shared only between a particular cognitive measure and
the trained tasks, then narrow generalization would be expected. On the other hand, transfer to multiple (disparate)
measures could be taken as evidence that training has an
impact on a domain-general mechanism.
Others have shown that WM training benefits are not
task specific, but rather, that they extend beyond the
trained task by affecting WM processes, including interference buffering (Persson & Reuter-Lorenz, 2008) and
WM updating (Dahlin et al., 2008). A few tantalizing
studies have further demonstrated transfer of WM training
to other assessments of cognition, including measures of
cognitive control (Klingberg et al., 2005) and fluid intelligence (Jaeggi et al., 2008).
We administered a battery of cognitive skill assessments in order to determine whether WM training would
afford even broader generalization than has been previously demonstrated. Generalization was not universally
indicated: Some of the skills we assessed did not improve

198     Chein and Morrison
differentially with WM training (e.g., reasoning tasks,
Raven’s APM). However, our results replicate the finding that WM training can lead to improvements in cognitive control, as measured in the Stroop task (Klingberg
et al., 2005). In a novel finding, we further demonstrate
that WM training can lead to improvements in reading
comprehension. As we discuss below, the observation of
generalization to two highly disparate cognitive measures
(Stroop and Nelson–Denny) leads us to favor the view
that WM training can indeed have an impact on a domaingeneral mechanism.
Of course, CWM span tasks place demands on many different processes, including encoding, covert maintenance,
attention, updating, interference resolution, and controlled
memory search (and perhaps many others). Training may
influence one or more of these processes. Accordingly,
transfer to different cognitive measures may depend on
the impact of training on different processes. For example,
training may improve Stroop performance by strengthening controlled attention, but it may improve reading comprehension by increasing the capacity of verbal storage.
The latter view may have particular intuitive appeal,
since increased verbal storage would presumably allow
participants to maintain an expanded verbatim representation of recent text and to thus integrate and assimilate
the meaning of the text more easily. However, if increased
verbal storage capacity explains Nelson–Denny reading
comprehension increases, then individual improvements
in verbal WM should strongly predict the magnitude of
improvements in reading comprehension. Instead, we
found that verbal WM improvements were only a weak
and nonsignificant predictor of reading comprehension
gains. Meanwhile, spatial WM increases were a much
stronger predictor.
These findings can be addressed from the more parsimonious view that CWM training affects reading comprehension in the same way that it benefits both WM (verbal and spatial) and Stroop performance: by enhancing
a domain-general attentional mechanism not specifically
tied to verbal or spatial storage per se, but to the coordination of information maintenance in the face of additional
processing demands. This interpretation is consistent
with the broader claim that CWM tasks are highly predictive of cognitive function specifically because they limit
rehearsal and domain-specific strategy use and, hence,
emphasize the role of domain-general attentional mechanisms (Cowan et al., 2005). Since verbal rehearsal strategies are so highly developed, the suppression of rehearsal
and the isolation of attention control mechanisms may be
more complete in spatial CWM span tasks (Conway et al.,
2005). Accordingly, the inclusion of a spatial CWM component in training may have been especially important in
producing the observed generalization effect. Indeed, our
correlational findings suggest that reading comprehension
improvements were particularly dependent on increased
spatial WM performance.
Admittedly, one difficulty for the claim that transfer arises from the impact of WM training on a domain­general attentional process is that both improved and unim-

proved measures included in our assessment battery have
been linked, on the basis of latent variable studies, to this
shared WM mechanism (Kane et al., 2004). The selectivity
of transfer invites the question of why training generalized
to some of the cognitive skills that we measured, but not
to others. Specifically, the training group did not show differential improvement on measures of reasoning or fluid
intelligence, which also putatively demand controlled attention. One way to address the apparent discrepancy is
to assume that tasks for which we observed improvements
simply place a higher demand on controlled attention than
do the unaffected tasks. Another possibility is that our
study lacked the statistical power to reveal modest training
effects in seemingly unaffected measures. Indeed, there
was a trend toward significance in the effect of training
on the ETS reasoning composite that would likely have
reached statistical significance in a larger sample.
However, such power limitations do not readily account
for our failure to replicate a transfer of WM training benefits to measures of fluid intelligence (as was observed
by Jaeggi et al., 2008), since we did not find even a trend
for improvement in trained participants on Raven’s APM.
Beyond statistical explanations, differences in the training
paradigms used for the two studies may explain the differences in transfer effects. The training program used by
Jaeggi et al. (2008) involved 400 trials per training session,
with a dual n-back training paradigm designed to emphasize binding processes and task management. Conversely,
our training paradigm included only 32 trials per session
and more heavily emphasized maintenance in the face of
distraction. Finally, the seemingly conflicting results may
be due to differences in intelligence test administration.
As was pointed out in a recent critique (Moody, 2009),
Jaeggi et al. (2008) used atypical speeded procedures in
administering their tests of fluid intelligence, and these
alterations may have confounded the apparent effect of
WM training on intelligence.
The selective pattern of transfer that we observed is informative in another regard. As was the case in almost
every previous WM training study (cf. Persson & ReuterLorenz, 2008), our experiment lacked an “active” control
group (i.e., a group engaged in alternate activities during
the training interval). We cannot, therefore, rule out that
observed improvements were the result of differences in
the motivation level or expectations of trained, relative
to untrained, participants: a placebo effect. However, an
explanation of this type does not explain the selectivity
of transfer. If trained participants were simply more motivated at the time of the second assessment, differences in
motivation level should have been as apparent on tests of
fluid intelligence and reasoning as on tests of WM, cognitive control, and reading comprehension. Additionally, the
observed correlation between spatial CWM and Nelson–
Denny improvements is not readily explained as a simple
placebo effect.
In sum, the present study demonstrates success in using
a CWM task to improve WM, cognitive control, and reading comprehension. Isolated demonstrations of generalization from WM training should be interpreted with

Working Memory Training and Transfer     199
some degree of caution, and further work in this area—
including replications, investigations of durability, tests
of relative efficacy, and considerations of the real-world
benefits of different WM training paradigms—is clearly
needed. Still, the present findings constitute an important step toward bridging the WM training literature and
a long history of empirical and psychometric work based
on CWM tasks. Moreover, the results are consistent with
the hypothesis that WM training can have an impact on
domain-general cognitive mechanisms; thus, these results benefit multiple areas of cognition. In particular, our
discovery that WM training can yield improvements in
reading skill—even among college-aged participants—
encourages enthusiasm for the increasingly popular belief
that WM training can be used as a general tool for promoting important cognitive skills.
Author Note
The authors thank Jenny Andrade for her role in data collection. Correspondence concerning this article should be addressed to J. M. Chein,
Department of Psychology, Temple University, 825 Weiss Hall, 1701
N. 13th Street, Philadelphia, PA 19122 (e-mail: jchein@temple.edu).
REFERENCES
Chase, W. G., & Simon, H. A. (1973). The mind’s eye in chess. In W. G.
Chase (Ed.), Visual information processing (pp. 215-281). New York:
Academic Press.
Conway, A. R. A., Kane, M. J., Bunting, M. F., Hambrick, D. Z.,
Wilhelm, O., & Engle, R. W. (2005). Working memory span tasks:
A methodological review and user’s guide. Psychonomic Bulletin &
Review, 12, 769-786.
Cowan, N., Elliott, E. M., Saults, J. S., Morey, C. C., Mattox, S.,
Hismjatullina, A., & Conway, A. R. A. (2005). On the capacity of attention: Its estimation and its role in working memory and
cognitive aptitudes. Cognitive Psychology, 51, 42-100. doi:10.1016/
j.cogpsych.2004.12.001
Dahlin, E., Neely, A. S., Larsson, A., Bäckman, L., & Nyberg, L.
(2008). Transfer of learning after updating training mediated by the
striatum. Science, 320, 1510-1512. doi:10.1126/science.1155466
Engle, R. W., & Bukstel, L. H. (1978). Memory processes among
bridge players of differing expertise. American Journal of Psychology,
91, 673-689. doi:10.2307/1421515

Engle, R. W., & Kane, M. J. (2004). Executive attention, working
memory capacity, and a two-factor theory of cognitive control. In
B. H. Ross (Ed.), The psychology of learning and motivation (Vol. 44,
pp. 145-199). New York: Elsevier.
Ericsson, K. A., & Chase, W. G. (1982). Exceptional memory. American Scientist, 70, 607-615.
Jaeggi, S. M., Buschkuehl, M., Jonides, J., & Perrig, W. J. (2008).
Improving fluid intelligence with training on working memory.
Proceedings of the National Academy of Sciences, 105, 6829-6833.
doi:10.1073/pnas.0801268105
Kane, M. J., Hambrick, D. Z., Tuholski, S. W., Wilhelm, O., Payne,
T. W., & Engle, R. W. (2004). The generality of working memory capacity: A latent-variable approach to verbal and visuospatial memory
span and reasoning. Journal of Experimental Psychology: General,
133, 189-217.
Klingberg, T., Fernell, E., Olesen, P. J., Johnson, M., Gustafsson, P., Dahlström, K., et al. (2005). Computerized training of
working memory in children with ADHD—A randomized, controlled
trial. Journal of the American Academy of Child & Adolescent Psychiatry, 44, 177-186.
Klingberg, T., Forssberg, H., & Westerberg, H. (2002). Training
of working memory in children with ADHD. Journal of Clinical &
Experimental Neuropsychology, 24, 781-791.
Moody, D. E. (2009). Can intelligence be increased by training on a
task of working memory? Intelligence, 37, 327-328. doi:10.1016/
j.intell.2009.04.005
Persson, J., & Reuter-Lorenz, P. A. (2008). Gaining control: Training
executive function and far transfer of the ability to resolve interference. Psychological Science, 19, 881-888.
Verhaeghen, P., Cerella, J., & Basak, C. (2004). A working memory
workout: How to expand the focus of serial attention from one to
four items in 10 hours or less. Journal of Experimental Psychology:
Learning, Memory, & Cognition, 30, 1322-1337. doi:10.1037/0278
-7393.30.6.1322
Notes
1. The Raven’s APM is not a speeded test. All participants completed
the test well within 45 min.
2. Second-assessment reading comprehension data were missing for
1 trained and 1 control participant; these participants were, therefore, necessarily excluded from analyses of the reading comprehension measure.

(Manuscript received July 17, 2009;
revision accepted for publication November 25, 2009.)