Sunteți pe pagina 1din 9

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/275355443

Measuring Working Memory With Digit Span and the Letter-Number


Sequencing Subtests From the WAIS-IV: Too Low Manipulation Load and Risk
for Underestimating Modality Effects

Article  in  Applied Neuropsychology: Adult · April 2015


DOI: 10.1080/23279095.2014.992069 · Source: PubMed

CITATIONS READS

11 2,225

1 author:

Jens Egeland
University of Oslo
86 PUBLICATIONS   1,726 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

ADHD, neuromotor dysregulation and chronic pain. View project

Effect of Physical Activity in Psychosis (Ephaps randomized controlled trial) View project

All content following this page was uploaded by Jens Egeland on 01 June 2016.

The user has requested enhancement of the downloaded file.


This article was downloaded by: [University of Oslo]
On: 27 April 2015, At: 00:44
Publisher: Routledge
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,
37-41 Mortimer Street, London W1T 3JH, UK

Applied Neuropsychology: Adult


Publication details, including instructions for authors and subscription information:
http://www.tandfonline.com/loi/hapn21
Click for updates
Measuring Working Memory With Digit Span and
the Letter-Number Sequencing Subtests From the
WAIS-IV: Too Low Manipulation Load and Risk for
Underestimating Modality Effects
ab
Jens Egeland
a
Division of Mental Health & Addiction, Vestfold Hospital Trust, Tønsberg
b
Institute of Psychology, University of Oslo, Oslo, Norway
Published online: 24 Apr 2015.

To cite this article: Jens Egeland (2015): Measuring Working Memory With Digit Span and the Letter-Number Sequencing
Subtests From the WAIS-IV: Too Low Manipulation Load and Risk for Underestimating Modality Effects, Applied
Neuropsychology: Adult, DOI: 10.1080/23279095.2014.992069

To link to this article: http://dx.doi.org/10.1080/23279095.2014.992069

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained
in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the
Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and
are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and
should be independently verified with primary sources of information. Taylor and Francis shall not be liable for
any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever
or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of
the Content.

This article may be used for research, teaching, and private study purposes. Any substantial or systematic
reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any
form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://
www.tandfonline.com/page/terms-and-conditions
APPLIED NEUROPSYCHOLOGY: ADULT, 0: 1–7, 2015
Copyright # Taylor & Francis Group, LLC
ISSN: 2327-9095 print/2327-9109 online
DOI: 10.1080/23279095.2014.992069

Measuring Working Memory With Digit Span and the


Letter-Number Sequencing Subtests From the WAIS-IV:
Too Low Manipulation Load and Risk for Underestimating
Modality Effects
Jens Egeland
Division of Mental Health & Addiction, Vestfold Hospital Trust, Tønsberg, and
Institute of Psychology, University of Oslo, Oslo, Norway
Downloaded by [University of Oslo] at 00:44 27 April 2015

The Wechsler Adult Intelligence Scale (WAIS) is one of the most frequently used tests
among psychologists. In the fourth edition of the test (WAIS-IV), the subtests Digit Span
and Letter-Number Sequencing are expanded for better measurement of working memory
(WM). However, it is not clear whether the new extended tasks contribute sufficient
complexity to be sensitive measures of manipulation WM, nor do we know to what
degree WM capacity differs between the visual and the auditory modality because the
WAIS-IV only tests the auditory modality. Performance by a mixed sample of 226
patients referred for neuropsychological examination on the Digit Span and Letter-Number
Sequencing subtests from the WAIS-IV and on Spatial Span from the Wechsler Memory
Scale-Third Edition was analyzed in two confirmatory factor analyses to investigate
whether a unitary WM model or divisions based on modality or level/complexity best fit
the data. The modality model showed the best fit when analyzing summed scores for each
task as well as scores for the longest span. The clinician is advised to apply tests with
higher manipulation load and to consider testing visual span as well before drawing
conclusions about impaired WM from the WAIS-IV.

Key words: attention/perception, cognitive/learning, Digit Span, Letter-Number Sequencing, tests,


WAIS, working memory

The concept of working memory (WM) has attained revisions of the tests. In Norway, the Wechsler Adult
widespread significance among psychologists working in Intelligence Scale (WAIS) is the only test used by a
different fields. WM is related to fluid intelligence (Shelton, majority of practicing psychologists (Vaskinn & Egeland,
Elliot, Matthews, Hill, & Gouvier, 2010) and might mediate 2012), and it is also rated as the most frequently used test
state-dependent decrements in learning (Egeland et al., among European psychologists (Evers et al., 2012). It thus
2003) and attention (Andersen, Hovik, Skogli, Egeland, & seems reasonable to assume that this test is also the most
Øie, 2013). It is also closely linked to executive function widespread measure of WM. Historically, the Wechsler
and memory (Unsworth & Spillers, 2010). Deficits in tests were divided into two factors, performance and verbal
WM are associated with most clinical states. Reflecting this IQ. A third factor called “freedom from distractibility” had
close relationship, the concept has been more closely inte- already been identified in the Wechsler-Bellevue Scale in
grated into the Wechsler intelligence scales with successive 1952 (Cohen, 1952), although it did not become clinically
relevant until the publication of Kaufman’s (1979) book,
Address correspondence to Jens Egeland, Ph.D., Division of Mental Intelligent Testing With the WISC-R. The third factor was
Health & Addiction, Vestfold Hospital Trust, P. O. Box 2267, Tønsberg, later further developed and divided into separate measures
3103 Norway. E-mail: jens.egeland@siv.no
2 EGELAND

of Processing Speed and Working Memory. In the fourth is often used in clinical research. One example is the
revision of the WAIS (Wechsler, 2008), both the Digit Span finding of Martinussen, Hayden, Hogg-Johnson, and
test and the Letter-Number Sequencing subtest were further Tannock (2005) that participants with attention-deficit
developed to provide a better assessment of the manipu- hyperactivity disorder (ADHD) are more impaired in
lation element in WM. WM is still only tested in the manipulation compared with simple storage.
auditory modality. The question asked here is whether the On the other hand, possible domain-specific impair-
changes made are sufficient for considering the Wechsler ments have continued to be of interest within research in
Adult Intelligence Scale-Fourth Edition (WAIS-IV) a dyslexia, specific language impairment, and nonverbal
relevant tool for measuring WM. learning disorders. Shah and Miyake (1996) demonstrated
In the Letter-Number Sequencing subtest in the WAIS- that reading span correlated well with reading comprehen-
IV, participants now receive 3 training tasks and the number sion measures but not with spatial ability measures,
of tasks are extended from 21 to 30 by introducing 6 tasks whereas the spatial span showed the opposite pattern.
involving only one letter and number and by extending the However, Miake (2001) emphasized that the correlations
number of three-letter tasks to 9 instead of 3. Sequential between reading comprehension and WM are higher for
Span was introduced as a new subtask in the WAIS-IV Digit complex reading span than for digit span, although these
Span, adding to Digit Forward and Backward to increase measures also correlate. Reduced phonological awareness,
manipulation WM load. The base-rate figures are offered which is highly associated with reduced digit span perfor-
based on the maximal span for each of the tasks as are base mance, is characteristic of dyslexia. In a meta-analysis of
Downloaded by [University of Oslo] at 00:44 27 April 2015

rates for difference scores between the different conditions. 53 studies, Melby-Lervåg, Lyster, and Hulme (2012) found
Originally, the concept of WM was introduced by that participants with dyslexia performed 0.71 standard
Baddeley and Hitch (1974) and was referred to as the multi- deviations below healthy controls on auditory short-term
modal model of WM. Two modality-specific slave systems memory measures, including Digit Span Forward. Of
were posited: the phonological loop and the visual-spatial studies comparing impairment within both the auditory
scratchpad. The slave systems were equivalent to attention and visual domains, Gooch, Snowling, and Hulme (2011)
span or short term-memory capacity and were involved in found that digit recall was impaired in dyslexia but not in
simple storage of information for a short period of time. ADHD, while visual span was impaired only in ADHD.
Whenever the information had to be manipulated in some Witruk, Ho, and Schuster (2002) found no support for
way, a more complex overarching non-modality-specific a general, modality-independent deficit in children with
system called the central executive had to be activated. dyslexia, as found in those with ADHD. The previously
The unique contribution of the model was the postulation cited study by Martinussen et al. (2005) also revealed
of the central executive because short-term memory was a modality effect in ADHD, with larger effect sizes in both
already an established concept (Miller, 1956). When visuospatial storage and manipulation tasks compared with
referring to WM today, it is unclear whether one refers to the corresponding auditory WM tasks. Preßler, Krajewski,
the central executive element solely or whether the slave and Hasselhorn (2013) identified groups of children charac-
systems are considered as well. Engle (2010) argue that it terized by either visual or phonological WM deficits and
is the complex span tests that are important for variability found that good phonological WM was favorable for the
in WM. Individual differences appear mainly when there later acquisition of reading, writing, and mathematics,
is demand for some distracting activity or retrieval from whereas good visual WM was a prerequisite for arithmetic
secondary memory (Unsworth & Engle, 2007). According skills. These studies indicate that modality differences in
to this view, only the complex tests have external validity WM capacity may be important when considering reading
in the sense of being able to account for differences in and writing impairments. Aboitiz, Aboitiz, and Garcia
learning ability specifically (Unsworth & Engle, 2007) or (2010) claimed that the phonological loop is vital both for
other cognitive and adaptive functions more broadly (Engle, the phylogenetic and ontogenetic development of language.
2010). This is in line with Miake’s (2001) view, calling for In their meta-analysis of brain mechanisms underlying
the use of WM span tasks that are different from simple WM, Nee et al. (2013) found that a content division (i.e.,
short-term memory tasks such as digit span by imposing modality) was more prominent than functional divisions
dual-task demand on the participant (i.e., some sort of (i.e., storage and manipulation).
simultaneous mental manipulation while processing infor- Summing up, it seems to be a convergence among
mation). In the Engle (2010) view of WM, the phonological researchers that complex span tasks are best for detecting
loop and visual-spatial scratchpad posited by Baddeley and test variability and to account for daily life variability
Hitch (1974) become less central, as they are considered in performance. The first question posed here, however,
merely two of what could be a dozen modality- or is whether the changes made in both the Digit Span
material-specific stores. and Letter-Number Sequencing subtests in the WAIS-IV
The distinction between simple storage and complex are sufficient for differentiating them from the simple
manipulation tasks seems now to be well established and forward condition measuring short-term memory capacity.
WAIS WORKING MEMORY 3

The importance of modality differences for variability in relevant only in special cases but not necessarily in standard
WM is less clear. Modality-specific impairments may play assessments because it does not posit a fundamental threat
a role in dyslexia, particularly when measured with simple to generalizing the auditory tests to WM in general. A best
span tasks, but they may perhaps account for less variability fit for the unitary model would be taken as an indication
in more complex manipulation tasks. The next research that the Backward, Sequencing, or Letter-Number Sequen-
question is therefore related to the amount of modality-spe- cing tasks are not sufficiently complex to be differentiated
cific variability in the span tests in the WAIS-IV. In the from the simple Forward task. If a level model gives best
clinic, we were interested in whether it is sufficient to test fit to the data, the WAIS has accomplished what was
WM only in the auditory modality as is done in the intended with the revision of the WM tasks from the third
WAIS-IV. to the fourth revision. If the modality model gives the best
In the technical and interpretative manual of the WAIS- fit, one must be careful in inferring both a general WM defi-
IV (Wechsler, 2008, p. 15), Digit Span Forward is described cit and an impairment in the executive element as well. The
as a test involving “attention, encoding and auditory WAIS-IV offers both measures of length of best span or
processing,” while the term WM is reserved for the Back- total score earned for each condition. It is an empirical
ward and Sequencing tasks as well as the Letter-Number question as to which set of measures best reflects the under-
Sequencing subtest. There is no reference to modality lying structure of WM or whether they show the same struc-
specificity in describing these tasks. The question for the ture. In the CFA presented here, the three hypothesized
clinician is whether (a) to limit interpretations of impaired models are tested both with the total points for each subtask
Downloaded by [University of Oslo] at 00:44 27 April 2015

performance to a possible modality-specific impairment, and with the maximal length of each task. For the Spatial
or (b) to interpret impairment in the latter three WM tests Span in the WMS-III, no such division between span length
as a level effect in manipulation specifically or (c) as and total scores is offered.
a general effect. This could be crucial when examining
participants for either reading or language deficits or a more
general deficit expected, for example, in ADHD or schizo- METHODS
phrenia. How often this leads to problems of interpretation
of the test findings when test participants exhibit a range Participants
of different cognitive problems is not clear. This could
be rephrased as a question of what is the most potent Two hundred and twenty-six sequentially collected test
dimension of WM when using the span tests offered in the protocols were taken from the author’s part-time private
Wechsler tests and whether it could be tested in practice. Permission was granted by the Norwegian Data
a confirmatory factor analysis (CFA). Inspectorate. The participants had been referred for specia-
The last two revisions of the WAIS have been standar- lized neurocognitive assessment. Table 1 shows the compo-
dized together with new versions of the Wechsler Memory sition of the sample divided according to diagnosis or cause
Scale (WMS). This scale offers a visual analogue to the of referral/tentative diagnosis before assessment to give the
Digit Span subtest in the WAIS, the Spatial Span subtest, reader a picture of the heterogeneity of the cognitive and
which is equivalent to the Corsi Block Test. A CFA of psychological conditions involved. The age of the parti-
the WMS-Third Edition (WMS-III) showed a separate cipants ranged from 16 to 82 years old (Mage ¼ 38.1 years,
WM factor composed of Spatial Span and Letter-Number
Sequencing (Millis, Malina, Bowers, & Ricker, 1999). A TABLE 1
joint WMS-III and WAIS-Third Edition (WAIS-III) CFA Cause of Referral/Tentative Diagnosis
also showed a WM factor composed of Spatial Span N %
together with the three WAIS-III tests comprising the
WM index in that test. A CFA of the WAIS-IV and the Attention-deficit hyperactivity disorder 55 24.3
Learning disorders 45 20.0
WMS-Fourth Edition (Holdnack, Zhou, Larrabee, Millis,
Mild cognitive impairment 12 5.3
& Salthouse, 2011) showed two models that fit the data Cerebrovascular accidents 12 5.3
equally well: a seven-factor model with separate auditory Traumatic brain injury 12 5.3
and visual WM factors and a five-factor solution with a Acquired cognitive dysfunction of unclear origin 33 14.6
unitary WM factor. Addiction 14 6.2
Affective disorders 11 4.9
In the current study, we analyzed the raw scores of the
Toxic solvent exposure 6 2.7
WMS-III Spatial Span Forward and Backward tasks Anxiety disorders/posttraumatic stress disorder 6 2.6
together with the WAIS-IV Digit Span and Letter-Number Fatigue 10 4.4
Sequencing raw scores to determine whether a unitary WM Schizophrenia 5 2.2
factor, level, or modality factors provide the best fit to the Mental retardation 1 0.4
Other 4 1.8
data. If a unitary factor or a level model offers the best
Total 226 100.0
fit, one could claim that the testing of visual WM may be
4 EGELAND

SD ¼ 14.3). Fifty-three percent of participants were men Spatial Span Backward is differentially impaired.
and 47% were women. The mean General Ability Index Thus, in this model, Digit Span Forward and both
was 92 (SD ¼ 15) and ranged from 57 to 141. Spatial Span conditions are considered to measure
simple storage, while Digit Span Backward, the
Sequential task, and the Letter-Number Sequencing
Measures
Span measure manipulation.
From the WAIS-IV Digit Span subtest, the following mea- 3. The modality model. In this model all auditory tests
sures were collected: Digit Span Forward, Backward, and are grouped together (all three subtasks of Digit
Sequencing total scores, as well as longest span for each Span þLetter-Number Sequencing) in one factor
task. Total score and longest score were also collected from hypothesized to differ from a visual factor composed
the Letter-Number Sequencing subtest. To give background of Spatial Span Forward and Backward.
information as to the overall intellectual functioning of
the participants, the General Ability Index was computed CFA was conducted on raw scores for each of the four
based on the three subtests comprising the Verbal Compre- models, applying the LISREL 8.3 program (Jöreskog &
hension Index (Similarities, Vocabulary, and Information) Sörbom, 1993). The goodness-of-fit measures should be
and the three subtests comprising the Perceptual Reasoning interpreted as follows (Keith, 2005).
Index (Block Design, Matrix Reasoning, and Visual
Puzzles). Chi-square (v2). When applying chi-square in
Downloaded by [University of Oslo] at 00:44 27 April 2015

From the WMS-III, the total scores on the subtest Spatial comparing the hypothesized and observed models, a low
Span Forward and Backward were analyzed. value means a good fit.

v2/df. If the chi-squares for two models are the same,


Data Analyses
the more constrained or more parsimonious model is pre-
The goodness-of-fit for five different factor models was ferred. Parsimony is reflected in CFA models by df. Thus,
tested in two separate sets of analyses, where total scores the smallest v2 in comparison to df represents the best fit
were entered in the first set. In the second set, the relevant of the data and should be no more than 2 for a good model
span length was entered for the WAIS-IV data, because the fit (Jöreskog & Sörbom, 1993).
WAIS provides an assessment of these measures as well.
For Spatial Span, total scores for the forward and backward Goodness-of-fit index and adjusted goodness-of-fit
conditions were entered in both sets of analyses. The index. These measures show how much better the model
models read as follows: fits as compared with no model at all. The adjusted
goodness-of-fit index is adjusted for degrees of freedom.
1. One general WM factor underlying all six measures. Values can vary from 0 to 1. Values greater .90 indicate a
2. a, 2b, and 2c. Level of processing distinguishing good model fit.
between a simple storage factor and a manipulation
factor. To be comprehensive, we tested three variants Comparative fit index. Values of .95 or greater sug-
of the division in levels: In 2a, Digit Span Forward gest a good fit of the model to the data, and values greater
and Spatial Span Forward were considered simple than .90 suggest an adequate fit.
storage tasks, while Digit Span Backward, Sequential
Span, and Spatial Span Backward and the Letter- Root mean square error of approximation. The
Number Sequencing subtest were considered manipu- root mean square error of approximation (RMSEA) is a mea-
lation tasks tapping the central executive. Model 2b sure of approximate fit. Smaller values suggest a better fit,
tests the hypothesis that the two backward conditions with values of .06 or less suggesting a good fit and those
are not complex enough to sufficiently tax executive of approximately .08 suggesting an adequate fit.
functions and will therefore group together with the
forward conditions of the same task. The Sequential
task and the Letter-Number Sequencing Span will RESULTS
then comprise the more complex manipulation factor.
Model 2c is based on the findings by Wilde, Strauss, Table 2 shows the results of the CFA based on total scores
and Tulsky (2004) and Kessels, van den Berg, Ruis, from each of the WM subtests. Only the modality factor
and Brands (2008) that Spatial Span Backward differs model depicted in Figure 1 shows adequate or good fit
from Digit Span Backward by not being more between the model and the observed results for all fit
demanding than the forward condition. Replicating indexes. The high v2/df and RMSEA values clearly show
this finding in healthy controls, Bacon, Parmentier, that the other factor models are not valid. Table 3 shows
and Barr (2013) nevertheless found that in dyslexia, the equivalent fit statistics based on the best or longest span.
WAIS WORKING MEMORY 5

TABLE 2
CFA of Unitary, Level, and Modality Models of WM-Span Subtest Scores From WAIS-IV and WMS-III (N ¼ 226)

Model χ2 (df) χ2/df GFI AGFI CFI RMSEA

1. Unitary WM 32.43 (9) 3.60 .95 .90 .94 .109


2a. Level model: DF þ SF vs. DB, DS, SB, LN 27.13 (8) 3.39 .96 .90 .94 .104
2b. Alternative level model: DF, DB, SF, SB vs. DS, LN 30.83 (8) 3.85 .96 .88 .94 .114
2c. Alternative level model: DF, SF, SB vs. DB, DS, LN 32.52 (8) 4.06 .95 .88 .93 .118
3. Modality model 6.55 (8) 0.82 .99 .97 1.0 .0

Note. Numbers in bold satisfy criteria for adequate fit between model and data. CFA ¼ confirmatory factor analysis; WM ¼ working memory; WAIS-IV ¼
Wechsler Adult Intelligence Scale-Fourth Edition; WMS-III ¼ Wechsler Memory Scale-Third Edition; GFI ¼ Goodness-of-Fit Index; AGFI ¼ Adjusted
Goodness-of-Fit Index; CFI ¼ Comparative Fit Index; RMSEA ¼ root mean square error of approximation; DF ¼ Digit Span Forward; DB ¼ Digit Span
Backward; DS ¼ Digit Span Sequence; LN ¼ Letter-Number Span; SF ¼ Spatial Span Forward; SB ¼ Spatial Span Backward.

DISCUSSION

The study is based on clinical data. The research literature


convincingly shows that WM can be subdivided according
to both modality and level of processing (Martinussen et al.,
Downloaded by [University of Oslo] at 00:44 27 April 2015

2005) and that complex WM tasks better detect ecologically


valid variability. The first issue addressed here is whether
the difference in complexity or manipulation load between
the simple storage test (i.e., the forward span condition on
the one side) and the manipulation tasks (i.e., the backward,
sequencing conditions, and the Letter-Number Sequencing
subtest) are large enough to stand out as separate factors
in a CFA. The second issue addressed is whether modality
differences can confound interpretations of WM impair-
ment when only the auditory modality is assessed as is
the case in the WAIS-IV. Both issues are highly relevant
for clinical practice because the WAIS-IV is so frequently
in use in the clinic. The results show that the differences
in level of processing were not large enough to separate
FIGURE 1 Two-factor modality-based model for Digit Span, Spatial
Span, and Letter-Number Sequencing. DF ¼ Digit Span Forward; DB ¼
a manipulation or complex WM element from the simple
Digit Span Backward; DS ¼ Digit Span Sequence; LN ¼ Letter-Number storage element. However, the CFA showed that a unitary
Span; SF ¼ Spatial Span Forward; SB ¼ Spatial Span Backward. WM model did not provide a good fit for the data either.
The modality model was the only model that had satis-
factory fit to the data when analyzing total scores, and
The fit measures were best for the modality model as well, it was superior to the other models when analyzing span
although the v2/df was not less than 2, as is considered the length as well.
cutoff for a good model fit. Thus, the modality model based The findings are consistent with the findings of Kessels
on total scores was superior to all other models. et al. (2008) of a verbal and spatial factor in a principal

TABLE 3
CFA of Unitary, Level, and Modality Models of WM-Span Best Performance Scores From the WAIS-IV and WMS-III (N ¼ 226)

Model χ2 (df) χ2/df GFI AGFI CFI RMSEA

1. Unitary WM 54.48 (9) 6.05 .92 .82 .88 .152


2a. Level model: DF þ SF vs. DB, DS, SB, LN 52.07 (8) 6.51 .93 .81 .88 .158
2b. Alternative level model: DF, DB, SF, SB vs. DS, LN 41.38 (8) 5.17 .94 .85 .91 .138
2c. Alternative level model: DF, SF, SB vs. DB, DS, LN 55.20 (8) 6.90 .92 .80 .87 .164
3. Modality model 18.09 (8) 2.26 .97 .93 .97 .076

Note. Numbers in bold satisfy criteria for adequate fit between model and data. CFA ¼ confirmatory factor analysis; WM ¼ working memory; WAIS-
IV ¼ Wechsler Adult Intelligence Scale-Fourth Edition; WMS-III ¼ Wechsler Memory Scale-Third Edition; GFI ¼ Goodness-of-Fit Index; AGFI ¼ Adjusted
Goodness-of-Fit Index; CFI ¼ Comparative Fit Index; RMSEA ¼ root mean square error of approximation; DF ¼ Digit Span Forward; DB ¼ Digit Span
Backward; DS ¼ Digit Span Sequence; LN ¼ Letter-Number Span; SF ¼ Spatial Span Forward; SB ¼ Spatial Span Backward.
6 EGELAND

component analysis of Digit Span and Corsi Block Span in difference in simple span is also corroborated in memory
elderly individuals. Bowden, Petrauskas, Bardenhagen, tests or by a difference between the Perceptual Reasoning
Meade, and Simpson (2013) performed a CFA on all single and Verbal Comprehension Indexes in the WAIS-IV.
Digit Span Forward and Backward items and found no fit Lasonen, Leppämäki, and Hokkanen (2009) found that
for the forward/backward dichotomy. They concluded that only participants with dyslexia and not those with ADHD
forward and backward span reflect the same cognitive performed poorer than controls on the Digit Span subtest
ability. Neither Perry et al. (2001) nor Twamley, Palmer, of the WAIS-III. Goldstein, Beers, Siegel, and Minshew
Jeste, Taylor, and Heaton (2006) found differences between (2001) found that only those with a reading disorder or glo-
forward and backward span in patients with schizophrenia, bal deficits were impaired in the Digit Span in a comparison
otherwise expected to have a specific deficit in the manipu- of children with deficits in arithmetic and reading, global
lation element of WM. This was taken to mean that the learning disorders, and autism. These results are to be
backward span was not sensitive enough to capture differ- expected given the present findings, but to the author’s
ential impairments. Instead one has to use more complex knowledge, they have not been replicated with the WAIS-
tests, as suggested by Engle (2010) and Miyake (2001). IV extended span testing.
The fact that the differences in manipulation load were The Digit Span and Letter-Number Sequencing of the
lower than the size of the modality effect even when includ- WAIS are only two of the three tests comprising the
ing the Sequence task and the extended Letter-Number WAIS-IV WM Index score. However, the problem of gen-
Sequencing subtest indicates that the addition and extension eralizing an auditory test result to WM in general might be
Downloaded by [University of Oslo] at 00:44 27 April 2015

of the presumptive WM tasks in the WAIS-IV are not suf- accentuated when including the third subtest, Arithmetic.
ficient to test complex WM considered to be most clinically For the third version of the WAIS, the Arithmetic subtest
relevant. was found to measure verbal learning more than WM
The fit between the modality model and the observed (Gregoire, 2004), and it fit better into a verbal comprehen-
scores was not only better than the different complexity sion factor than a WM factor (Egeland, Bosnes, &
models, but it represented a good fit. The supremacy of Johansen, 2008). The CFA based on the normative sample
the modality model was therefore not due to the lack of sen- of the WAIS-IV showed the best fit between model and
sitivity to complexity differences. There could be several observed scores when the Arithmetic subtest was allowed
reasons for this. One is related to the sample: Could it be to load on both Verbal Comprehension and Working
an overrepresentation of participants with specific learning Memory (Wechsler, 2008). This finding had nevertheless
deficits either related to language or visuospatial proces- no consequences for the computation of index scores in
sing? Twenty percent were referred with a suspected the WAIS-IV.
learning deficit. Thirteen of these participants (5% of the There are some limitations to the study. The sample is
sample) were suspected to have a language-related learning not very well described in terms of diagnoses, only by
disorder, and five (2% of sample) were suspected to have initial cause of referral, which in some cases included
a possible nonverbal learning deficit, which is as expected a good diagnostic process. The purpose of the neuro-
in a mixed clinical sample. Another explanation seems psychological testing involving the WAIS-IV was to give
more plausible: As claimed by Engle (2010), simple tasks a functional description, rather than a diagnosis. In the cases
are more automatic and do not capture deficits in controlled where a definite diagnosis was set, the test results were part
attention, shifting, or inhibition that constitute the varia- of the diagnostic decision, thus confounding independent
bility seen in more complex tests. The closer one gets to and dependent variables. Nevertheless, the list of causes
testing sensory capacity with little demand for executive of referral shows that a wide array of clinical problems
monitoring, the more dominant the interindividual differ- was examined and contributed to variance in the examined
ences in preferred or dominant processing mode will be. measures. If only participants with specific learning deficits
If we then consider the span tests of the WAIS as more were examined, the modality effects would probably be
simple storage capacity tests and not tests of complex WM, even more prominent. The study is based on clinical file
it is important not to conclude that participants have a nor- data. If the study had been designed as a WM study from
mal WM based on such tests alone. Conversely, impaired the start, we would have included WM tests that involved
span test performance could not safely be interpreted as a higher cognitive load, and thus, we would have expected
anything more than a deficit in auditory processing capacity that complexity level would emerge as a factor. However,
unless it is supplemented by a corresponding test of the purpose of the study was not to test the dimensions of
visuospatial span, such as the one applied here from the WM as such, but to check the validity of the presumed
WMS. Detecting a modality supremacy or deficit in WM tests in the WAIS-IV. The finding of the prominence
individual cases, however, may be important both for the of modality advises the clinician on the one hand to also
diagnosis of learning deficits and reading disorders and apply tests of complex span and to include a spatial or
for guidance as to how patients will learn most effectively. visual measure before drawing firm conclusions about
The latter is particularly important if the modality a WM deficit.
WAIS WORKING MEMORY 7

REFERENCES Lasonen, M., Leppämäki, S., & Hokkanen, L. (2009). Adult dyslexia and
attention deficit disorder in Finland—Project DyADD: WAIS-III
cognitive profiles. Journal of Learning Disabilities, 42, 511–527.
Aboitiz, F., Aboitiz, S., & Garcia, R. R. (2010). The phonological Martinussen, R., Hayden, J., Hogg-Johnson, S., & Tannock, R. (2005).
loop: A key innovation to human evolution. Current Anthropology, A meta-analysis of working memory impairments in children with
51, 55–63. attention deficit/hyperactivity disorder. Journal of the American
Andersen, P. M., Hovik, K. T., Skogli, E. W., Egeland, J., & Øie, M. Academy of Child & Adolescent Psychiatry, 44, 377–384.
(2013). Symptoms of ADHD in children with high-functioning autism Melby-Lervåg, M., Lyster, S.-A. H., & Hulme, C. (2012). Phonological
are related to impaired verbal working memory and verbal delayed skills and their role in learning to read: A meta-analytic review.
recall. PLoS One, 8, e64842. Psychological Bulletin, 138, 322–352.
Bacon, A. M., Parmentier, F. B., & Barr, P. (2013). Visuospatial memory Miller, G. A. (1956). The magical number seven, plus or minus two: Some
in dyslexia: Evidence for strategic deficits. Memory, 21, 189–209. limits on our capacity for processing information. Psychological
doi:10.1080/09658211.2012.718789 Review, 63, 81–97.
Baddeley, A. D., & Hitch, G. J. (1974). Working memory. In G. Bower Millis, S. R., Malina, A. C., Bowers, D. A., & Ricker, J. H. (1999).
(Ed.), The psychology of learning and motivation (Vol. 8, pp. 47–90). Confirmatory factor analysis of the Wechsler Memory Scale-III. Journal
San Diego, CA: Academic Press. of Clinical and Experimental Neuropsychology, 21, 87–93.
Bowden, S. C., Petrauskas, V. M., Bardenhagen, F. J., Meade, C. E., & Miyake, A. (2001). Individual differences in working memory:
Simpson, L. C. (2013). Exploring the dimensionality of digit span. Introduction to the special section. Journal of Experimental Psychology:
Assessment, 20, 188–198. General, 130, 163–168.
Cohen, J. (1952). Factors underlying Wechsler-Bellevue performance of Nee, D. E., Brown, J. W., Askren, M. K., Berman, M. G., Demiralp, E.,
three neuropsychiatric groups. Journal of Abnormal Social Psychology, Krawitz, A., & Jonides, J. (2013). A meta-analysis of executive
47, 359–365. components of working memory. Cerebral Cortex, 23, 264–282.
Downloaded by [University of Oslo] at 00:44 27 April 2015

Egeland, J., Bosnes, O., & Johansen, H. (2008). Factor structure of the Perry, W., Heaton, R. K., Potterat, E., Roebuck, T., Minassian, A., &
Norwegian Version of WAIS-III in a clinical sample: The arithmetic Braff, D. L. (2001). Working memory in schizophrenia: ‘Online’ storage
problem. Assessment, 16, 292–300. versus executive functioning. Schizophrenia Bulletin, 27, 157–176.
Egeland, J., Sundet, K., Rund, B. R., Asbjørnsen, A., Hugdahl, K., Preßler, A.-L., Krajewski, K., & Hasselhorn, M. (2013). Working memory
Landrø, N. I.,… Stordal, K. I. (2003). Sensitivity and specificity capacity in preschool children contributes to the acquisition of school
for memory dysfunction in schizophrenia: A comparison with major relevant precursor skills. Learning and Individual Differences, 23,
depression. Journal of Clinical and Experimental Neuropsychology, 138–144.
25, 79–93. Shah, P., & Miyake, A. (1996). The separability of working memory
Engle, R. W. (2010). Role of working memory capacity in cognitive resources for spatial thinking and language processing: An individual
control. Current Anthropology, 51, 17–26. differences approach. Journal of Experimental Psychology: General,
Evers, A., Muniz, J., Bartram, D., Boben, D., Egeland, J., Fernandez- 125, 4–27.
Hermida, J. R., . . . Urbánek, T. (2012). Testing practices in the 21st Shelton, J. T., Elliott, E. M., Matthews, R. A., Hill, B. D., & Gouvier, Wm.
century: Developments and European psychologists’ opinions. D. (2010). The relationship of Working Memory, secondary memory,
European Psychologist, 17, 300–319. and fluid intelligence: Working Memory is special. Journal of Experi-
Goldstein, G., Beers, S. R., Siegel, D. J., & Minshew, N. J. (2001). mental Psychology: Learning, Memory, and Cognition, 36, 813–820.
A comparison of WAIS-R profiles in adults with high-functioning Twamley, E. W., Palmer, B. W., Jeste, D. V., Taylor, M. J., & Heaton,
autism or differing subtypes of learning disabilities. Applied Neuro- R. K. (2006). Transient and executive function working memory in
psychology, 8, 148–154. schizophrenia. Schizophrenia Research, 87, 185–190.
Gooch, D., Snowling, M., & Hulme, C. (2011). Time perception, Unsworth, N., & Engle, R. W. (2007). The nature of individual differences
phonological skills and executive function in children with dyslexia in Working Memory capacity: Active maintenance in primary memory
and/or ADHD symptoms. Journal of Child Psychology & Psychiatry, and controlled search from secondary memory. Psychological Review,
52, 195–203. 114, 104–132.
Gregoire, J. (2004). Factor structure of the French version of the Wechsler Unsworth, N., & Spillers, G. J. (2010). Working Memory capacity:
Adult Intelligence Scale-III. Educational and Psychological Attentional Control, secondary memory, or both? A direct test of the
Measurement, 64, 463–474. dual-component model. Journal of Memory and Language, 62,
Holdnack, J. A., Zhou, X., Larrabee, G. J., Millis, S. R., & Salthouse, T. A. 392–404.
(2011). Confirmatory factor analysis of the WAIS-IV/WMS-IV. Vaskinn, A., & Egeland, J. (2012). Testbruksundersøkelsen: En oversikt
Assessment, 18, 178–191. over tester brukt av norske psykologer [An overview of tests used
Jöreskog, K., & Sörbom, D. (1993). LISREL 8: Structural equation by Norwegian psychologists]. Tidsskrift for Norsk Psykologforening
modeling with the SIMPLIS command language. Lincolnwood, IL: (Journal of the Norwegian Psychological Association), 49, 658–665.
Scientific Software International. Wechsler, D. (2008). Wechsler Adult Intelligence Scale-Fourth Edition:
Kaufman, A. S. (1979). Intelligent testing with the WISC-R. New York, Technical and interpretative manual. San Antonio, TX: Pearson
NY: Wiley. Assessment.
Keith, T. Z. (2005). Using confirmatory factor analysis to aid in Wilde, N. J., Strauss, E., & Tulsky, D. S. (2004). Memory span on the
understanding the constructs measured by intelligence tests. In D. P. Wechsler scales. Journal of Clinical & Experimental Neuropsychology,
Flanagan & P. L. Harrison (Eds.), Contemporary intellectual assessment 26, 539–549.
(pp. 373–402). New York: Guilford. Witruk, E., Ho, C. S., & Schuster, U. (2002). Working Memory in
Kessels, R. P. C., van den Berg, E., Ruis, C., & Brands, A. M. A. (2008). dyslexia: How general is the deficit? In E. Witruk, A. D. Friederici,
The backward span of the Corsi Block-Tapping Task and its association & T. Lachmann (Eds.), Basic functions of language, reading and
with the WAIS-III Digit Span. Assessment, 15, 426–434. reading disorders (pp. 281–297). Dordrecht, Netherlands: Kluwer.

View publication stats

S-ar putea să vă placă și