Archives of Clinical Neuropsychology

20 (2005) 291–299

Test–retest reliability of the Extended Complex Figure

Test—Motor Independent administration (ECFT-MI)
Stacey E. Woodrome a , Philip S. Fastenau a,b,∗
a
Department of Psychology, Purdue University School of Science, Indiana University Purdue University
Indianapolis (IUPUI), 402 N. Blackford Street, Indianapolis, IN 46202-3275, USA
b
Department of Psychiatry, Indiana University School of Medicine, Indiana University Purdue University
Indianapolis (IUPUI), Indianapolis, IN 46202-3275, USA
Accepted 5 August 2004

Abstract

The Extended Complex Figure Test (ECFT; Fastenau, 1996b, 2003a), adds recognition and matching
trials to the Rey–Osterrieth Complex Figure Test (ROCFT). An alternate administration of the ECFT,
the ECFT–Motor Independent version (ECFT-MI; Fastenau, 2003a) measures visuospatial memory
without the motoric demands of the ROCFT. ECFT-MI test–retest reliabilities were examined in 55
healthy adults tested two times separated by one week. The Recognition Trial showed strong reliability
for the Total Score (r = .80, p < .0001). Reliabilities for Recognition Detail Scale and Subscales (r
= .62–.74, p < .0001) and for the Global Scale (r = .51, p ≤ .005) were attenuated by fewer items
per subscale but were comparable to other memory tests. Matching Trial reliability coefficients were
attenuated by extreme ceiling effects and restriction of range, but scores were very consistent between
the two administrations. This study supports the temporal stability of the ECFT-MI.
© 2004 Published by Elsevier Ltd on behalf of National Academy of Neuropsychology.

Keywords: Rey; Memory; Cognitive; Cognition; Neuropsychological Tests; Learning

The Rey–Osterrieth Complex Figure Test (ROCFT) was developed as a measure of visual
perceptual abilities and memory (Osterrieth, 1944; Rey, 1941; Rey & Osterrieth, 1993). The
most common method of administration of the ROCFT consists of a copy trial and immediate
recall trial, followed 20–60 min later by a delayed recall trial. Comparisons among these trials
can help differentiate between different processes (e.g., a normal copy trial rules out motor

∗
Corresponding author. Tel.: +1 317 274 2961; fax: +1 317 274 6756.
E-mail address: pfastena@iupui.edu (P.S. Fastenau).

0887-6177/$ – see front matter © 2004 Published by Elsevier Ltd on behalf of National Academy of Neuropsychology.
doi:10.1016/j.acn.2004.07.006
292 S.E. Woodrome, P.S. Fastenau / Archives of Clinical Neuropsychology 20 (2005) 291–299

and constructional deficits in a poor immediate recall drawing). However, the ROCFT lacks
a recognition trial and perceptual matching trial. As Fastenau (1996b) discussed, an effective
memory test compares free recall and recognition abilities in order to separate retrieval deficits
from encoding and storage deficiencies. In addition, a complete spatial memory test should
ensure that poor performances are not due to perceptual deficits. The Extended Complex Figure
Test (ECFT; Fastenau, 1996b, 2003a; Fastenau & Manning, 1992) was developed to address
these issues.
In the design of the ECFT, Recognition and Matching Trials follow the delayed recall trial
of the ROCFT (Fastenau, 1996b, 2003a). Thus, the ECFT can answer two important questions
that the ROCFT cannot: (1) Are poor performances on the copy trial due to perceptual deficits
or motor/constructional deficits? (2) Are poor performances on free recall due to retrieval or
encoding/storage deficits? These additional components have been normed in a healthy adult
population (Fastenau, Denburg, & Hufford, 1999), and preliminary norms are available for
children (Sasher & Fastenau, 2001). The advantages of using the ECFT over other memory tests
have been presented elsewhere, and evidence for reliability and validity have been documented
elsewhere (see Fastenau, 2003b for a review).
The ROCFT is a common component of neuropsychological batteries, used by almost
two-thirds of neuropsychologists (Knight, Kapland, & Ireland, 2003). However, due to the
fine motor demands of this task, some individuals (e.g., those with hemiparesis, spasticity,
or amputation of the preferred hand) are unable to complete the measure. Consequently, the
Extended Complex Figure Test—Motor Independent administration (ECFT-MI; Fastenau,
2003a) was developed to assess spatial memory in these populations. The ECFT-MI eliminates
the motor demands associated with the ROCFT and with the standard administration of the
ECFT. A “study trial” (visually tracing the design) replaces the copy trial, and Recognition
and Matching Trials follow immediately thereafter (i.e., the two free recall trials are omitted).
Although the ECFT-MI, which excludes copy and recall trials, cannot differentiate between
recall and retrieval deficits, it does provide a measure of spatial memory for patients with
significant motor impairments who are unable to complete the full ECFT. In fact, Keller and
Fastenau (2000) demonstrated that the ECFT-MI yields recognition scores that are identical
to those generated following a copy trial for the ROCFT, lending support to the validity of this
measure.
Thus, the ECFT-MI has a potentially valuable place in neuropsychological assessment in that
it can now fill the spatial memory void in batteries for patients with serious motor impairments.
There is no research to date on the reliability of the ECFT-MI. In the present study, temporal
stability (i.e., test–retest reliability) was examined.

1. Methods

1.1. Sample

A total of 55 adults participated in a study of the test–retest reliability of the ECFT-MI as

well as several new versions of established neuropsychological measures. Participants were
between 21 and 60 years of age at the time of assessment. The level of education ranged from
 S.E. Woodrome, P.S. Fastenau / Archives of Clinical Neuropsychology 20 (2005) 291–299 293

Table 1
Participant demographics
Variable M S.D.
n 55
Age (years) 36.9 12.5
Education (years completed) 15.8 2.8
WRAT-3 Reading Standard Score 107.2 7.7
Beck Depression Inventory 9.0 6.4
Gender (% female) 81.8
Race (%)
Caucasian 83.6
African American 7.3
Asian/Pacific Islander 1.8
Biracial 7.3
Handedness (% right handed) 85.5

10 years to more than 4 years of graduate school; most of the participants were well educated
(44% had completed at least two years of college). Participants were recruited at a large, urban,
midwestern university, with attempts to recruit a diverse sample that would be relatively rep-
resentative of the general population. They were recruited through announcements in courses
with high enrollments of diverse and part-time students; through an electronic list serve for
university students, faculty, office staff, and service workers; and through flyers posted by time
clocks in service areas of the university. Table 1 presents the demographic information of the
participants.

1.2. Procedure

After giving informed consent, participants were tested twice (T1, T2), one week apart.
Participants received financial compensation for completing both phases of testing. The par-
ticipants completed the ECFT-MI at both time points at the end of a 2-h neuropsychological
test battery (Table 2). Although several memory tests were administered within the same test-
ing session in the present study, a prior study using the ECFT and three other memory tests
showed no order effects on test scores (Fastenau, Denburg, & Domitrovic, 1997). Trained
neuropsychological technicians administered the measures.

1.3. Measures

The ECFT-MI consists of Recognition and Matching Trials. Both have three scales (Global,
Detail, and Total), and the Detail Scale has two subscales (Left Detail and Right Detail).
Administration of the ECFT-MI begins with the following verbal instructions: “I have a picture
that I want you to study very carefully. After you have studied it, I will test your memory to see
how much of the picture you remember. Trace the picture in your mind, as if you were drawing
294 S.E. Woodrome, P.S. Fastenau / Archives of Clinical Neuropsychology 20 (2005) 291–299

Table 2
Neuropsychological tests in order of administration
Rey Auditory Verbal Learning Test (RAVLT; Immediate Recall Trials)
Brief Visual Memory Test—Revised (BVMT; Immediate Recall Trials)
WAIS-III Digit Span
WAIS-III Digit Symbol
WAIS-III Letter—Number Sequencing
Trail Making Test
Paced Auditory Serial Addition Test (PASAT)
Rey Auditory Verbal Learning Test (Delayed Trials)
Brief Visual Memory Test—Revised (Delayed Trials)
Extended Complex Figure Test—Motor Independent Version (ECFT-MI)

it. Make sure to remember what all the different parts look like and where each part belongs.”
After revealing the figure the examiner says, “Trace the picture in your mind, as if you were
drawing it. Make sure to remember what all the parts look like and where each part belongs.
Try to remember everything about the picture. You have 3 min to study the picture” (Fastenau,
2003a). After the time elapses, the examinee is shown 30 different vertical arrangements of
five choices. The individual must select which of the items was part of the picture just studied.

Table 3
ECFT-MI means (S.D.), by Time, by Trial and by ECFT-MI scale or subscale
Max Time 1 Time 2

M (S.D.) M (S.D.)
Recognition Time
Raw Score
Global 7 5.3 (1.4) 6.0 (1.2)
Left Detail 9 4.8 (2.2) 6.2 (2.3)
Right Detail 11 6.9 (2.2) 7.9 (1.9)
Detail 23 13.7 (4.3) 16.7 (4.1)
Total 30 19.0 (5.4) 22.7 (4.9)

Scaled Score
Global 18 12.3 (3.6) 14.1 (4.0)
Left Detail 18 12.0 (3.1) 14.2 (3.6)
Right Detail 18 10.8 (3.2) 12.1 (3.1)
Detail 18 11.4 (3.3) 13.6 (3.4)
Total 18 11.2 (3.1) 13.5 (3.2)

Matching
Raw Score
Global 2 2.0 (0.2) 2.0 (0.0)
Left Detail 4 3.8 (0.5) 3.8 (0.5)
Right Detail 4 3.7 (0.6) 3.8 (0.4)
Detail 8 7.4 (0.9) 7.6 (0.8)
Total 10 9.4 (0.9) 9.6 (0.8)

Note. Max = maximum score possible. Matching Trial raw scores do not convert to scaled scores.
 S.E. Woodrome, P.S. Fastenau / Archives of Clinical Neuropsychology 20 (2005) 291–299 295

One of these choices is from the ROCFT and the remaining four are distracters containing
frequent and/or clinically significant errors.
The Matching Trial follows the Recognition Trial. It consists of 10 items where five choices
are vertically arranged to the left of a reproduced component from the ROCFT. Only one of the
choices is a duplicate of the reproduced component. More detailed descriptions and samples
of the Recognition and Matching tasks can be found in Fastenau (1996b).

1.4. Data analysis

The means and standard deviations were calculated for the Recognition and Matching scales
and subscales at T1 and T2 (Table 3). Pearson correlations were computed between T1 and
T2 for each of the scale and subscale scores.

2. Results

The results of the analyses of the Recognition scales and subscales at Time 1 and Time 2
are summarized in Table 4. All of the Recognition scales and subscales (Global, Left Detail,
Right Detail, Detail, and Total) at Time 1 were significantly correlated with the same scales
and subscales at Time 2 (p < .0001). The correlations ranged from moderate to strong (r
= 0.51–0.80). As depicted in Table 4, the magnitude of the reliability coefficients grossly
corresponded with the number of items on the scale. The Recognition Global Scale, which
consists of only seven items, had the lowest reliability. In addition, there was a prominent ceiling
effect, especially at Time 2. The Left and Right Detail Recognition subscales had moderate
reliability, with 9 and 11 items respectively. The Recognition Total Scale, which consists of the
most items, had the strongest reliability. The Recognition Detail Scale reliability was slightly

Table 4
Pearson correlations between Time 1 and Time 2, by ECFT-MI scale or subscale
Scale Pearson r No. of items
Recognition Trial
Global Scale 0.51*** 7
Left Detail Subscale 0.72**** 9
Right Detail Subscale 0.62**** 11
Detail Scale 0.74**** 23
Total Scale 0.80**** 30
Matching Trial
Global Scale – 2
Left Detail Subscale 0.25 4
Right Detail Subscale 0.09 4
Detail Scale 0.23 8
Total Scale 0.22 10

Note. **** p < .0001, *** p < .005. All other correlations, p > .05. –: the Pearson r could not be correlated due to a
lack of variation of scores.
296 S.E. Woodrome, P.S. Fastenau / Archives of Clinical Neuropsychology 20 (2005) 291–299

Table 5
Change in Matching scale and subscale scores between Time 1 and Time 2
Difference % of sample
Global Scale 0.0 96.4
1.0 3.6
Left Detail Subscale 0.0 78.2
1.0 18.2
2.0 3.6
Right Detail Subscale 0.0 69.1
1.0 29.1
2.0 0.0
3.0 1.8
Detail Scale 0.0 65.5
1.0 21.8
2.0 11.0
3.0 0.0
4.0 0.0
5.0 1.5
Total Scale 0.0 63.6
1.0 23.7
2.0 11.0
3.0 0.0
4.0 0.0
5.0 1.8

lower than expected, despite consisting of 23 items and despite being free of ceiling and floor
effects; however, it still had moderate reliability.
Matching scales and subscales yielded much smaller retest reliabilities. It was suspected
that the low retest correlations on ECFT-MI Matching were due in large part to ceiling effects
and restriction of range. Therefore, post hoc analyses were conducted to assess the degree to
which participants’ Matching scores changed from T1 to T2. Table 3 shows that the mean
scores for the Matching subscales are nearly perfect (i.e., substantial ceiling effects). Table 5
also demonstrates that Matching scores were very consistent, with more than 87% of the
sample showing a change of 1 point or less on every scale. This consistency reflects test–retest
stability; it also illustrates the further restriction of range on the changes in scores, which
appears to have attenuated the reliability coefficient.

3. Discussion

The Recognition Total Scale of the ECFT-MI appears to have good test–retest reliability.
It is most likely the case that the reliability coefficients for the remaining Recognition scales
and subscales were affected by having a limited number of items and, for one subscale, ceiling
effects. Likewise, Matching reliability coefficients appear to have been greatly attenuated by
 S.E. Woodrome, P.S. Fastenau / Archives of Clinical Neuropsychology 20 (2005) 291–299 297

the limited number of items on the scales and subscales, by ceiling effects, and by restriction
of range. Nonetheless, the consistency of Matching scores (>87% of the sample showing ≤1
point change) provides support for the temporal stability of the Matching Trial.
The ECFT-MI Recognition test–retest correlations are comparable to other, established mea-
sures of memory. The Figural Memory subtest from the Wechsler Memory Scale—Revised
(WMS-R; Wechsler, 1987) most closely parallels the administration of the ECFT-MI. Exami-
nees study the stimulus (they do not draw it) and then immediately pick out the correct design
from an array of choices in a recognition paradigm (no intervening free recall trials). Using
the standardization sample (N = 316) and a retest interval ranging from four to six weeks, the
average retest coefficient was 0.44. Other visual measures of memory on the WMS-R used
free recall and cued recall formats (Visual Reproductions I & II, Visual Paired Associates I
& II); average retest reliabilities for those non-recognition formats ranged from 0.46 to 0.59
(Wechsler, 1987).
The test–retest reliability on the revision of that scale, the Wechsler Memory Scale-III
(WMS-III; Wechsler, 1997), was assessed in a sample of 297 individuals. The average amount
of time between administrations was 35.6 days. On the WMS-III, the closest analog to ECFT-
MI Recognition is Visual Reproductions Recognition (although, it is preceded by immediate
free recall and delayed free recall trials). The retest reliability for that subtest was 0.56 in the
age range of the present study. Other visual measures of recognition memory on the WMS-III
(Faces I & II) ranged from 0.58 to 0.64 in the age range of the present study; cued recall
measures (Family Pictures I & II) range from 0.61 to 0.73 (Tulsky, Zhu, & Ledbetter, 1997).
Similar to the ECFT-MI, the more visually orientated measure, the Brief Visuospatial Mem-
ory Test—Revised (BVMT-R), was assessed for test–retest reliability in 71 individuals who
were tested an average of 55.6 days after the first exam. For the BVMT-R Recognition Trial,
no reliability coefficient was reported because of ceiling effects at both testing sessions, with
85% identifying all six targets accurately at Time 1 and with 100% achieving perfect scores at
Time 2. The retest correlations on free recall trials ranged from 0.60 to 0.84 (Benedict, 1997).
As seen here, memory tests tend to have relatively modest retest coefficients, and recognition
measures tend to have lower retest reliabilities than do free recall and cued recall trials. For
comparable visual recognition tasks reviewed here, retest reliabilities ranged 0.44 to 0.64 and
in some cases were too attenuated by ceiling effects to be computed.
Similarly, visual perceptual measures tend to be encumbered by ceiling effects and restric-
tion of range, which attenuate the reliability coefficients. Fastenau (1996a) observed dramatic
attenuation of the alpha reliability coefficient in a matching trial for WMS-R Visual Reproduc-
tions, which resembles the ECFT Matching Trial but has fewer items than ECFT Matching.
Although the WMS-R study measured internal consistency rather than temporal stability, the
attenuation of reliability in scales consisting of just a few, easy items is relevant to the present
study. More elaborate perceptual matching measures with more items show considerably bet-
ter retest coefficients. One-year test–retest reliability for the Hooper Visual Organization Test
(HVOT; Hooper, 1983) was 0.68 in a sample of older adults (Levin et al., 1991). For the
Judgment of Line Orientation, the manual reports a retest reliability of 0.90 in retest intervals
ranging from 1 to 21 days (Benton, Hamsher, Varney, & Spreen, 1983).
Therefore, the ECFT-MI Recognition retest reliabilities are commensurate with or exceed
those reported for other published visual recognition memory tests. It should be acknowledged
298 S.E. Woodrome, P.S. Fastenau / Archives of Clinical Neuropsychology 20 (2005) 291–299

that the retest interval in our study was briefer than those in the above studies; thus, it could
be beneficial to lengthen this interval in later research.
For the Matching scale and subscales, ceiling effects and restriction of range attenuated
test–retest reliabilities, but the consistency of Matching scores across the one-week interval
provides some support for the temporal stability of that trial. This study would have benefited
from including a clinical sample in addition to the healthy group in order to observe whether or
not the ceiling effects would remain. Furthermore, the fact that the sample was not representa-
tive of the population as a whole, being predominately female and more highly educated, might
have also influenced these findings. Ultimately, the Matching Trial is best used as a screening
instrument, consistent with the recommendations in the ECFT test manual (Fastenau, 2003a).
Future studies could examine retest reliability for the Matching Trial in a clinical sample of
individuals who show perceptual deficits.
In conclusion, present research on the ECFT-MI indicates that it will be a useful resource to
neuropsychologists assessing spatial memory in patients whose motor impairments are previ-
ously established. Thus, a neuropsychologist working with a patient who does not have serious
motor deficits may chose to use the standard ECFT to examine motor, perceptual, encoding,
and retrieval processes of spatial memory. However, in the case of serious motor impairments,
in which the ROCFT or the ECFT would normally be avoided, the ECFT-MI can now be ad-
ministered to acquire an index of spatial memory processing. Reliability data for the ECFT-MI
from clinical samples would be a useful supplement to the healthy data presented here.

Acknowledgements

This research was supported by a grant from Eli Lilly and Company to the second author.
The authors thank W. Thomas Hankins and Carrie A. McGinnis for their assistance with data
collection.

References

Benedict, R. H. B. (1997). Brief Visuospatial Memory Test—Revised: Professional manual. Odessa, FL: Psycho-
logical Assessment Resources Inc.
Benton, A. L., Hamsher, K. deS., Varney, N. R., & Spreen, O. (1983). Contributions to neuropsychological assess-
ment: A clinical manual. New York: Oxford University Press.
Fastenau, P. S. (1996a). An elaborated administration of the Wechsler Memory Scale—Revised. The Clinical
Neuropsychologist, 10, 425–434.
Fastenau, P. S. (1996b). Development and preliminary standardization of the “Extended Complex Figure Test”
(ECFT). Journal of Clinical and Experimental Neuropsychology, 18, 63–76.
Fastenau, P. S. (2003a). Extended Complex Figure Test (ECFT) manual. Los Angeles: Western Psychological
Services.
Fastenau, P. S. (2003b). Extended Complex Figure Test (ECFT): Rationale and empirical support for Recognition
and Matching. In J. A. Knight & E. F. Kaplan (Eds.), Handbook of Rey–Osterrieth Complex Figure Usage:
Clinical and research applications (pp. 255–282). Odessa, FL: Psychological Assessment Resources.
Fastenau, P. S., Denburg, N. L., & Domitrovic, L. A. (1997). Intentional and incidental memory: Order effects in
clinical testing. Professional Psychology: Research and Practice, 28, 32–35.
 S.E. Woodrome, P.S. Fastenau / Archives of Clinical Neuropsychology 20 (2005) 291–299 299

Fastenau, P. S., Denburg, N. L., & Hufford, B. J. (1999). Adult Norms for the Rey–Osterrieth Complex Figure Test
and for Supplemental Recognition and Matching Trials from the Extended Complex Figure Test (ECFT). The
Clinical Neuropsychologist, 13, 30–47.
Fastenau, P. S., & Manning, A. A. (1992). Development of a recognition task for the Complex Figure Test [Abstract].
Journal of Clinical and Experimental Neuropsychology, 14, 43.
Hooper, H. E. (1983). Hooper Visual Organization Test. Los Angeles: Western Psychological Services.
Keller, K. L., & Fastenau, P. S. (2000). Comparisons between visual encoding and visual-motor encoding on
the Rey: Support for using Extended Complex Figure Test (ECFT) norms for the Motor-Independent ECFT
(MI-ECFT) [Abstract]. Journal of the International Neuropsychological Society, 6, 114.
Knight, J. A., Kapland, E., & Ireland, L. D. (2003). Survey findings of Rey–Osterrieth Complex Figure usage. In
J. A. Knight & E. F. Kaplan (Eds.), Handbook of Rey–Osterrieth Complex Figure Usage: Clinical and research
applications (pp. 45–56). Odessa, FL: Psychological Assessment Resources.
Levin, B. E., Llabre, M. M., Reisman, S., Weiner, W. J., Sanchez-Ramos, J., Singer, C., et al. (1991). Visuospatial
impairment in Parkinson’s disease. Neurology, 41, 365–369.
Osterrieth, P. A. (1944). Le test du copie d’une figure complexe: Contribution à l’étude de la perception et de la
mémoire [The test of copying a complex figure: A contribution to the study of perception and memory]. Archives
of Psychology (Chicago), 30, 206–356.
Rey, A. (1941). L’examen psychologique dans les cas d’encephalopathie traumatique [Psychological examination
in cases of traumatic encephalopathy]. Archives of Psychology (Chicago), 28, 286–340.
Rey, A., & Osterrieth, P. A. (1993). Translations of excerpts from Andre Rey’s Psychological examination of
traumatic encephalopathy and P.A. Osterrieth’s The Complex Figure Copy Test. Clinical Neuropsychologist, 7,
4–21.
Sasher, T. M., & Fastenau, P. S. (2001). Preliminary child normative data for the Extended Complex Figure Test
(ECFT) [Abstract]. The Clinical Neuropsychologist, 15, 258.
Tulsky, D., Zhu, J., & Ledbetter, M. F. (1997). WAIS-III/WMS-III technical manual. San Antonio, TX: The Psy-
chological Corporation.
Wechsler, D. (1987). Wechsler Memory Scale—Revised (WMS-R). San Antonio, TX: The Psychological Corpora-
tion.
Wechsler, D. (1997). Wechsler Memory Scale—3rd Edition (WMS-III). San Antonio, TX: The Psychological Cor-
poration.