Discussion on Psychology and Aging

Psychology and Aging
1998. Vol. 13. No. 1. 120-126
Copyright 1998 by the American Psychological Association, Inc.
0882-7974/98/$3.00
Aging and the Stroop Effect: A Meta-Analysis
Paul Verhaeghen and Lieve De Meersman
University of Leuven
In this meta-analysis, data from 20 studies comparing younger and older adults on the Stroop
interference effect, contained in 15 articles, were analyzed. No significant difference was found in
the Stroop interference effect, expressed as mean standardized difference, between the 2 age groups
(for younger adults: d = 2.04; for older adults: d = 2.17). Moderator variables were present, but
these did not produce age differences. Brinley analysis showed that a single regression line with a
slowing factor of 1.9 described the data well (R1 = .83) and confirmed that no Age X Condition
interaction was present in the data. Likewise, no Age X Condition interaction was found when the
data were fitted to the information loss model; the age ratio of decay rates was estimated to be 1.4.
Consequently, the apparent age-sensitivity of the Stroop interference effect appears to be merely an
artifact of general slowing.
The Stroop effect (Stroop, 1935) is one of the best-known
effects in cognitive psychology. Research participants typically
take much longer to name the ink color of a color word depicting
a color incongruent with the ink color (e.g., the word green printed
in red) than to name the color of a patch of color, of a string of
X s or symbols, or of noncolor words. In his extensive review,
MacLeod (1991) counted more than 700 articles dealing with this
effect, either examining it directly or using it as a tool to study
other cognitive processes, making the Stroop effect one of the
most well-replicated phenomena in experimental psychology.
Explanations advanced for the Stroop effect invariably center
around interference of two responses (viz., reading the word
and naming the color) associated with the same input. Recent
theories focus on the relative associate strength of the two response
tendencies of reading and color naming (Cohen, Dunbar,
& McClelland, 1990) or of the two task sets of reading and
color naming (Monsell, 1996). What presumably happens is
that when a color word is shown, the response tendency or
task set of reading the word, through life-long experience with
reading, is initially activated more strongly than the novel response
tendency or task set of naming the color. Another way
of formulating this mechanism is that in the Stroop task the
Paul Verhaeghen and Lieve De Meersman, Department of Psychology,
University of Leuven, Leuven, Belgium.
This research was conducted while Paul Verhaeghen was a postdoctoral
researcher at the National Fund for Scientific Research, Flanders,
Belgium. It was supported by a grant from the National Fund for Scientific
Research, Flanders, Belgium. Lieve De Meersman worked on this
study in partial fulfillment of her master’s degree in psychology from
the University of Leuven.
We wish to thank Tim Salthouse and Denise Park for providing us
with the exact means and standard deviations from their studies and
Marilyn Hartman for answering our enquiry about the Hartman and
Hasher (1991) study.
Correspondence concerning this article should be addressed to Paul
Verhaeghen, who is now at the Department of Psychology, 430 Huntington
Hall, Syracuse University, Syracuse, New York 13244-2340. Electronic
mail may be sent to [email protected].
participant has to inhibit the prepotent response of reading the
word (i.e., decrease its activation level) in favor of increasing
activation of the nondominant response of color naming. This
process of decrease-increase in activation unfolds over time
(Lindsay & Jacoby, 1994), and thus participants are typically
slower when naming the ink color of an incongruent color word
(the interference condition) than when naming the color of a
neutral stimulus (the baseline condition).
Recently, the Stroop effect has attracted the interest of cognitive
aging researchers because it seems a good measure of inhibitory
processes, a construct recently invoked to explain cognitive
deficits associated with aging (e.g., Hasher & Zacks, 1988).
The inhibition theory of cognitive aging claims that the working
memory problems exhibited by older adults find their origin
in detective inhibitory mechanisms. According to this theory,
defective inhibition either causes irrelevant information to enter
into working memory, consequently limiting its functional capacity,
or causes irrelevant material within working memory not
to be suppressed, resulting in distraction from the task.
Interestingly, only a limited number of studies have examined
the Stroop effect in a younger versus older adult contrast (see
Table 1). Nevertheless, scientists who study human aging typically
appear confident that these studies have yielded age effects
that are “almost universal” (West, 1996, p. 287) and that the
Stroop interference effect is ‘ ‘highly robust and age-sensitive in
that older adults have been found to show more Stroop interference”
(Kwong See & Ryan, 1995, p. 459). There is, however,
a problem with the interpretation of the results found in most
aging studies concerning the Stroop effect, and this has to do
with the way the Stroop effect is traditionally measured. Usually,
an interference score is calculated by simply subtracting reaction
time (RT) in the baseline condition from reaction time in the
interference condition. If this difference score is larger in older
than in younger adults (or if the Age X Condition interaction
in an analysis of variance [ ANOVA] is significant), researchers
conclude that the interference effect is larger in older adults.
However, general slowing theories (Cerella, 1990; Myerson,
Hale, Wagstaff, Poon, & Smith, 1990) predict a larger interference
effect in older adults when measured by the traditional
120
AGING AND THE STROOP EFFECT 121
Table 1
Sample of Studies, Along With Effect Sizes for Interference in Younger and Older Adults
Baseline latency
(ms)
Study Younger Older
Interference latency
(ms)
Younger Older
Note. Dashes indicate that the effect size could not be calculated. Exp. = experiment; BLE = biological life events.
Interference effect
size (d)
Younger Older
Conn, Dustman, & Bradford (1984)
Comalli, Wapner, & Werner (1962)
Dulaney & Rogers (1994), Exp. 1
Dulaney & Rogers (1994), Exp. 2
Dulaney & Rogers (1994), Exp. 3
Hartman & Hasher (1991)
Houx, Jolles, & Vreeling (1993), no BLE
Houx, Jolles, & Vreeling (1993), BLE
Kieley & Hartley (1997), Exp. 1
Kieley & Hartley (1997), Exp. 2
Kwong See & Ryan (1995)
Li & Bosman (1996)
Panek, Rush, & Slade (1984), men
Panek, Rush, & Slade (1984), women
Park et al. (1996)
Salthouse (1996)
Salthouse & Meinz (1995)
Spieler, Balota, & Faust (1996)
Weir, Bruun, & Barber (1997), Exp. 1
Weir, Bruun, & Barber (1997), Exp. 2
20
32
20
14
6
44
42
18
16
45
82
35
19
31
44
40
49
27
17
24
40
15
20
14
6
24
66
52
16
45
92
35
11
20
131
64
85
50
18
24
412
582
536
524
453
560
525
528
557
712
536
589
543
521
416
492
495
671
656
625
574
689
704
674
571
609
580
656
593
771
683
697
808
784
469
693
631
970
875
781
882
1,044
841
772
657
993
790
869
757
822
789
753
972
932
583
819
746
759
1,052
969
1,274
1,651
1,235
1,292
946
1,275
997
1,435
860
920
1,166
999
2,043
1,991
707
1,310
1,088
1,146
1,760
1,406
2.54
— 3.07
1.61
1.55
2.63
2.37
1.96
2.25
1.05
2.01
2.10
2.43
2.54
3.19
2.07
1.90
1.14
2.50
1.66
3.12
— 2.97
1.69
1.58
2.37
3.16
2.92
2.88
1.38
2.34
1.20
1.57
1.92
2.73
1.70
2.77
1.05
1.78
1.79
difference score, even if merely general slowing is present. The
reason for this is that the relation between younger and older
adults’ reaction times is typically not additive, but is much better
described by a linear function with a slope larger than 1. In
other words, the age difference in response time is not expected
to be constant across conditions, but typically grows larger with
increasing latency of the task, even if the data are governed by
genera] slowing. Consequently, general slowing theories would
predict the age difference in the Stroop task to be larger in
the interference condition than the baseline condition, simply
because participants need more time for processing in the interference
condition than in the baseline condition. Thus, before
it can be stated with confidence that there is true age sensitivity
in the Stroop interference effect, one needs to demonstrate that
the age difference exhibited in the interference condition is
larger than the age effect predicted from general slowing. It is
reassuring to see that some recent studies have attempted to
take general slowing into account by relying on some form of
ratio scores (Dulaney & Rogers, 1994; Kwong See & Ryan,
1995; Panek, Rush, & Slade, 1984; Spieler, Balota, & Faust,
1996) in at least secondary analyses. However, ratio scores assume
that the function relating reaction times of older participants
to young participants is merely proportional. This model
has been found to fit data less well than the more sophisticated
multilayered slowing model (Cerella, 1990) or the information
loss model (Myerson et al., 1990). Consequently, we use these
latter models in our attempt to model the meta-analytic data.
The goal of this meta-analysis was thus to bring together the
available data on the age effect in the Stroop task and to examine
whether age differences are truly larger in the interference condition
as compared with the baseline. This hypothesis was investigated
using two methods. First, traditional methods of data pooling
(Hedges & Olkin, 1985) were used to calculate the Stroop interference
effect (expressed as the mean standardized difference between
the baseline and the interference condition) for both younger
and older adults. We then examined whether this interference effect
was larger for older than younger adults, as predicted by the
inhibition theory. Second, data were subjected to Brinley analysis
(e.g., Cerella, 1990), that is, we examined whether a single (general
slowing) or two different (inhibition theory) curves were
needed to describe the relation between the performance of
younger and older adults in the baseline and the interference conditions.
In order to minimize variance among studies, only data
pertaining to the color-word effect were analyzed.
Method
Sample of Studies
Studies were collected by consulting the PsycLIT electronic database,
through personal contacts, and by checking references found in the
articles thus retrieved. One reference not thus retrieved was kindly
pointed out by one anonymous reviewer. The search was concluded in
April 1997. Criteria for inclusion: (a) The study included at least one
sample of younger (mean ages 30 years or younger) and older adults
(mean ages 60 years or older); (b) Stroop data, using the color-word
task, were reported, including at least a baseline condition consisting of
noncolor-word stimuli (color patches, symbols, or noncolor words) and
an interference condition, consisting of naming the ink color of an incongruent
color word; and (c) the data were reported in a format amenable
to meta-analysis.1 No study was excluded for reasons other than not
1 For the studies by Park et al. (1996), Salthouse (1996), and Salthouse
and Meinz (1995), data in the original articles were not reported
for each age group separately. These authors were so kind as to provide
us with the exact means and standard deviations for their younger (ages
30 and younger) and older (ages 60 and older) participants.
122 VERHAEGHEN AND DE MEERSMAN
satisfying these criteria. Thus, to the best of our knowledge, this sample
comprises the totality of the published literature on aging effects in the
Stroop task. Studies, along with some of their characteristics, are listed
in Table 1.
Data Pooling
The system for data pooling advocated by Hedges and Olkin (1985)
was used for the first type of analysis. We calculated the Stroop interference
effect for each age group (younger and older adults) for each
study separately by subtracting the baseline condition latency from the
interference condition latency and dividing this score by the pooled
standard deviation, yielding a mean standardized difference, or g{. This
method is advocated for within-subject designs by Dunlap, Cortina,
Vaslow, and Burke (1996); it yields smaller effect sizes than the method
using the standard deviation of the difference scores. Moreover, the
Dunlap et al. method allows for direct and precise calculation of effect
sizes from means and standard deviations without recurrence to an estimate
of the correlation between tasks. (One study, Comalli, Wapner, &
Werner, 1962, did not report standard deviations for the separate age
groups, and thus did not allow for a precise calculation of effect size.
This study was discarded from our analysis.) As is usual in this type of
analysis, we decided to keep independent groups within studies (viz.,
men vs. women in the Panek et al., 1984, study, and biological life
events vs. no biological life events in the Houx, Jolles, & Vreeling, 1993,
study) separate in the analyses. As a result, 19 effect sizes were calculated
for each age group. A small-sample correction factor was applied
to the individual effect sizes in accordance with the principles outlined
in Hedges and Olkin (1985, p. 81), converting the #, values to dt. These
19 effect sizes were then averaged using a weighting factor for sample
size (Hedges & Olkin, 1985, pp. 109-117), to yield estimates of the
average interference effect (d+) in younger and older adults, respectively.
One advantage of the Hedges and Olkin (1985) approach is that a
test statistic is available for between-groups comparisons. In this study,
this statistic (the between-groups homogeneity statistic, QB, which is
chi-square distributed with degrees of freedom equal to the number of
groups minus 1; Hedges & Olkin, 1985, pp. 154-155) was used to test
whether the interference effect was larger in older as compared to
younger adults. Moreover, Hedges and Olkin (1985, pp. 155-156) have
proposed a within-group estimate of homogeneity ((2w, chi-square distributed
with the number of studies minus 1 as number of degrees
of freedom), indicating whether the studies can be considered to be
homogeneous, that is, whether the effect size can be considered a good
point estimate of a single population value.
Brinley Analysis
For the Brinley analysis, the mean latency data of the older adults
were regressed on the mean latency data of the younger adults (20 data
points for each age group). Mean latency data were expressed in seconds
needed for responding to a single stimulus.2 Two popular general slowing
models were applied to the data.
The first model applied is the multilayered slowing model advanced
by Cerella (1990). In this model, it is stated that aging brings about
differential slowing in peripheral processes (i.e., input and output processes)
and central processes. Cerella demonstrated that such a model
yields young-old data that can be described by a linear function:
RTM = a + b RTyo, (1)
The b parameter, or the slope of the function, describes the ratio of older
over younger central processing tune and thus provides an index of agerelated
slowing in central processing.
In order to test whether different equations are needed for the two
conditions in the Stroop task, an interaction analysis approach was used
(Berry & Feldman, 1985; for a brief tutorial on using this technique in
the context of young-old Brinley plots, see Myerson, Wagstaff, & Hale,
1994). In mis method, the data are fitted to the following regression
equation:
bt RTy<>u 2 Cond + b2 Cond X RTyoi (2)
The Cond variable is a dummy variable, taking the value 0 in the baseline
condition and the value 1 in the interference condition. If the a2 parameter
is significant, different intercepts, one for each condition, are needed to
describe the data adequately; if the b2 parameter is significant, different
slopes, one for each condition, are needed for adequate description of
the data. All equations were fitted using the Statistical Package for
the Social Sciences (SPSS, 1996) weighted least squares algorithm,
weighting for sample size. The reader may note that this theoretically
guided use of Brinley analysis is quite different from the exploratory
use of the technique that has recently generated much controversy (e.g.,
Fisk & Fisher, 1994; Perfect, 1994).
The second general slowing model used to fit the data was the information-
loss model by Myerson et al. (1990). This model assumes that a
constant proportion of information is lost during each processing step
and that the amount of information lost is larger in older than in younger
adults. This model is described by the equation
— b RTvai (3)
The m term in this equation describes the ratio of the older over younger
decay rates, that is, it gives the proportional age difference in the speed
with which information is lost when propagated through the cognitive
system. This equation was fitted using the SPSS nonlinear regression
procedure, weighting for sample size. The interaction lest used here
was fitting Equation 3 separately for each condition and testing for the
difference in b and m values using the confidence intervals given by the
regression program.
Significance level of all statistical tests was set at .05.
Results
Data Pooling
Effect sizes for the individual studies are reported in Table
1. The average interference effect in the T9 studies was 2.04 for
the younger adults (limits of the 95% confidence interval were
1.90 and 2.18) and 2.17 for the older adults (limits of the 95%
confidence interval were 2.04 and 2.29). Thus, as expected, the
interference effect was very large and significantly different
from zero for both age groups. However, the effect size did
not differ reliably across groups, as indexed by nonsignificant
between-groups homogeneity, QB( 1) = 1.71, ns, indicating that
the interference effect was as large in younger adults as in
older adults. Both effect sizes were significantly heterogeneous,
Qw(18) = 60.55 and 118.28, respectively, indicating considerable
variance in effect size, even in this sample of very comparable
studies.
The heterogeneity in the data was further explored by examining
potential moderator variables. Three such variables were
investigated: the use of patches of color versus colored noncolor
words or colored symbols (usually Xs) in the baseline condition,
presenting items one at a time versus presenting more than
2 Marilyn Hartman (personal communication, April 8, 1997) pointed
out that the number of items used in the Hartman and Hasher (1991)
study was actually 100 and not 50, as mentioned in that article.
AGING AND THE STROOP EFFECT 123
Table 2
Effect Sizes far Younger and Older Adults as a Function of Selected Moderator Variables
Younger adults Older adults
Moderator
Color patches
Symbols or words
One item at a time
More items at a time
Computerized presentation
Printed materials
k
13
6
4
15
12
7
dt
1.90
2.44
1.45
2.24
2.25
1.59
LL/UL
1.73/2.06
2.16/2.72
1.16/1.73
2.07/2.40
2.07/2.42
1.34/1.84
a.
47.93″
1.95
12.15″
26.06″
19.73″
23.20″
GB
(moderator) d+
10.68′ 2.01
2.72
22.34′ 1.31
2.43
17.62″ 2.45
1.46
LL/UL
1.87/2.15
2.45/2.98
1.06/1.57
2.29/2.58
2.30/2.60
1.23/1.69
GB
Qu (moderator)
81.38′ 21.19′
15.71′
11.46′ 56.32′
50.50′
44.57′ 49.76′
23.95′
GB
(age)
1.01
2.03
0.47
3.10
3.16
0.55
Note, k – number of studies; d+ = mean weighted effect size; LL/UL = lower, respectively upper limits of the 95% confidence interval of d+’,
Qw = within-group homeogeneity statistic; QB = between-groups homogeneity statistic.
‘ Significant heterogeneity at p < .05.
one item at a time, and using printed stimuli for presentation
and a stopwatch for recording reaction times versus computerized
testing and reaction time measurement. Results of this analysis
are presented in Table 2. None of the moderator analyses
resulted in the desired combination of between-groups heterogeneity
and within-group homogeneity. However, two results are
noteworthy. First, all of these moderator variables had an influence
on the Stroop effect in both age groups, as indexed by
significant between-groups heterogeneity according to the appropriate
statistic. Thus, color patches produced a larger Stroop
effect than colored noncolor words or colored symbols, presenting
more items at a time produced a larger Stroop effect than
sequential presentation of items, and printed stimuli and stopwatch
measurements produced a larger Stroop effect than computerized
testing and recording. Second, none of the groupings
of effect sizes according to these moderator variables resulted
in a reliable age effect, as indexed by nonsignificant betweengroups
heterogeneity according to the appropriate statistic.
Thus, no third order (Age X Condition X Moderator Variables)
interactions appeared to be present in the data.
Brinley Analyses
A Brinley plot of the young-old data is provided in Figure
1. This figure illustrates how large the interference effect is:
There is almost no overlap between the data clouds for the
baseline and the interference condition. Mean latency (weighted
for sample size) in the baseline condition was 536 ms for the
younger adults and 669 ms for the older adults; weighted mean
latency for the interference condition was 807 ms for the
younger adults and 1180 ms for the older adults. The mean
interference effect was thus 271 pis for the younger adults and
511 ms for the older adults.
The multilayered slowing model fitted these data quite well
(R2 = .83), with a = -0.34, and b = 1.88. Fitting Equation 2
to the data resulted in the following estimates for the parameters:
a, = 0.03, a2 = -0.61, £, = 1.19, and i>2 = 0.99. However, the
a?, parameter was not significant, indicating that there was no
reliable intercept difference between the two conditions. Consequently,
Equation 2 was refitted, this time omitting the a2 Cond
term. This resulted in nonsignificance of the fc2 parameter
(QI = -0.27, ft, = 1.74, and &2 = 0.07), indicating that the
slope difference between conditions was not reliable once the
intercept term was fixed to be equal across conditions. Consequently,
a single regression was sufficient to describe the data.
Largest Cook’s D value for the individual data points was .007,
suggesting that no outliers were present.
One extra analysis was undertaken to examine whether the
slope estimated from the regression analysis might not be a
systematic distortion of the effect present in the individual experiments.
It is possible, for instance, that the data points for
the baseline condition systematically tend to lie below the regression
line, whereas the data points for the interference condition
tend to lie above the regression line (for a discussion of
this and related problems with Brinley analysis, see Perfect,
1994). In that case, the regression line is not a good estimator
0.5
•»»
*
• Baseline
o Interference
Regression
0 0.5 1 1.5 2 2.5
RT(youngef)
Figure 1. Brinley plot of baseline and interference conditions for
Stroop color-word tasks, along with best fitting linear function. RT =
response time.
124 VERHAEGHEN AND DE MEERSMAN
of the average within-study effect, and one would expect that
the slopes of the lines connecting the baseline data point with
the interference data point in individual studies would systematically
be larger than the slope estimated from the regression
analysis. When we calculated the slopes from all 20 individual
studies and averaged these, weighting for sample size, the mean
slope was equal to 1.83 (SD = 0.38), which is not larger than
the slope of the regression line estimated on all 40 data points
(viz., 1.88). Thus, the 1.88 slope of the regression line seems
a good approximation of the mean interference effect in the
individual studies.
From a purely statistical standpoint, this sequence of analyses
suggests that the data are adequately explained by a single function,
indicating that merely general slowing is present in the
data. However, the reader may have noticed that although the
relevant statistics are not significant, the intercepts and slopes
obtained for the baseline and interference condition are quite
different (for the baseline condition: a = 0.03, b = 1.19; for
the interference condition: a = —0.58, b = 2.18). This might
suggest that the general slowing effect in the data may be a
statistical artifact because of lack of power. There is, however,
a theoretical reason for mistrusting the coefficients of these
separate equations. This has to do with the point of intersection
of the two lines. In interaction analysis, one theoretically expects
this point of intersection to lie below and to the left of the two
data clouds. If this is not the case, the implication would be
that there exist viable younger adult latencies for which the
latency of older adults in the condition with the largest slope
(presumably the most complex condition) would actually be
smaller than that of the younger adults. In our data, the point
of intersection lies way to the right of the shortest latency of
the younger adults, and in fact even to the right of the mean of
the baseline distribution, namely at the point (616 ms, 764
ms). This implies that the regression line for the interference
condition lies below the regression line for the baseline condition
in a large part of the latency range of the baseline condition.
Theoretically, this makes no sense because this amounts to
claiming that a condition involving the color-naming process
plus a reading-suppression process would yield a smaller age
difference than a condition not involving the extra process of
reading suppression, but then only for a particular range of
reaction times. In fact, we think this situation might best be
understood as an indication that some curvilinearity is present
in the data.
Fitting the data to the information-loss model and weighting
for sample size resulted in a slightly better fit (R2 = .84; note
that the significance of the difference in R2 with the linear
regression analysis cannot be tested because this is not a nested
comparison), with b = 1.58, and m — 1.40. When separate
functions were estimated for baseline and interference conditions,
the 95% confidence intervals for the m and b parameters
overlapped (R2 = .65, b = 1.21, SE = 0.12, m = 0.95, SE =
0.17; and R2 = .68, b = 1.61, SE = 0.09, m = 1.50, SE =
0.26), indicating that the age difference in decay rate was not
different across conditions.
Discussion
The main result from this analysis is obvious: The interference
effect in the Stroop task appears not to be age sensitive. Rather,
the data seem governed by a general age effect, indicating that
the apparently larger age effect in the interference condition (a
271-ms interference effect for the younger adults vs. a 511-ms
effect for the older adults) is merely a side effect of general
slowing. Three sets of analyses led us to this conclusion.
First, when data were pooled, the interference effect appeared
to be as large in younger adults as in older adults, namely about
2.1 standard deviations. This is a very strong effect, indeed. The
age effect (.13 SD, viz., 2.17-2.04) was not significant and
quite small in comparison to the size of the interference effect
itself. The reader may also note that 9 out of 19 effect sizes (as
close to 50% as possible) were smaller for older than for
younger adults, further supporting the claim of no age difference.
Moreover, although moderators of the Stroop interference
effect could be identified in the data, grouping studies according
to these moderator variables did not lead to a significant age
difference in the Stroop effect in any of these groupings.
Second, when the multilayered slowing model was fit to the
data, we found no evidence for an Age X Condition interaction,
suggesting that the central slowing factor is equal across tasks.
Central processing in the Stroop task was found to take about
1.9 times as much time in older adults than in younger adults.
This analysis was corroborated when the slopes for individual
studies were computed and averaged: a mean slope of 1.8
emerged. The slowing factor found in the Stroop data is quite
large if one assumes that color naming is a lexical task; the
typical slowing factor in lexical processing is about 1.5 (Lima,
Hale, & Myerson, 1991). Rather, this central slowing factor of
1.8 or 1.9 falls within the 1.8-2.0 slowing range typically found
in nonlexical tasks (Cerella, 1990). It can be debated whether
this implies that color naming is not a lexical process or whether
this demonstrates that more than one lexical slowing factor must
be distinguished (Laver & Burke, 1993).
As is usual in young-old latency data, the intercept term in
the equation was significant and negative, indicating that there
is differential slowing in peripheral and central processes, and
that central processes are slowed to a larger extent than peripheral
processes. This result has implications for a good estimation
of the Stroop interference effect in an age-comparison approach.
First, the traditional difference score is not informative at all
because one would expect this difference score to be larger in
the slowest group even if merely general slowing is present.
Second, a ratio score is also less than optimal because a mere
proportional model is not completely adequate for describing
young-old differences in Stroop data. When one wishes to use
the Stroop interference score as a predictor variable in a correlational
analysis, one alternative to difference or ratio scores may
be to use hierarchical regression, entering the data from the
baseline condition first, and the data from the interference condition
in the next step (for an example of this approach, see
Salthouse & Meinz, 1995)..In this way, one can estimate
whether the interference score explains additional variance over
and above the variance explained by the baseline condition.
Alternatively, one might use linear regression to predict the
interference condition latencies from the baseline latencies; the
residual latencies can be used as an estimate of the Stroop
effect or, more precisely, as an estimate of the deviation of
the interference effect from what can be expected knowing the
individual’s baseline speed and the mean interference effect
AGING AND THE STROOP EFFECT 125
across participants. An additional advantage of both these methods
may be that in this way the reliability problems associated
with difference scores (Cronbach & Furby, 1970) is circumvented.
If one is merely interested in the question whether group
differences in the interference effect are present, one may resort
to individual Brinley analysis (for applications of this technique,
see Maylor & Rabbit, 1994, and Sliwinski, Buschke, Kuslansky,
Senior, & Scarisbrick, 1994).
Third, when the information-loss model was fitted to the data,
a single equation again proved sufficient to adequately describe
the data. It may be noted, however, that the fit of this model
was only slightly better than that of the linear model, such that
one may assume that a linear model is sufficient for explaining
the data. The old over young decay-rate ratio in the informationloss
model was found to be about 1.4. This means that when
older participants are requested to name colors, they lose about
40% more information in the different processing steps than
younger adults. This ratio is slightly higher than the ratios found
by Myerson et al. (1990) for a variety of tasks, which lie between
1.21 and 1.33.
At first sight, the conclusion of a general age effect in the
Stroop test rather than an exacerbation of age differences in
the interference condition may seem surprising. Most of the
individual studies in this data pool conclude that there are indeed
large age differences in the interference effect, with the increase
in latency in older participants being much larger than in younger
adults. This is true in absolute terms, but we have explained
why we believe this conclusion to be seriously flawed (p. 120),
namely in that these studies fail to take the general or average
slowing present in the data into account. Interestingly, some
results from the primary studies cited already point at this general
effect. For instance, Salthouse (1996) found that all Stroop
conditions (including a compatible color-naming condition, in
which the ink color is identical to the color designated by the
color words) loaded quite high on a single speed factor comprising
a total of 20 measures of perceptual and reaction time speed.
The Stroop interference condition did not have any specific
variance attached to it, showing that it is merely representative
of the general speed factor apparent in this study. Salthouse
and Meinz (1995) found that the age-related variance in the
interference condition was reduced considerably after control
for latency in the baseline condition, namely from .323 to .038,
or a reduction of 88%. Thus, in this study, only a small portion
of the age-related variance in the interference condition could
not be explained through the slowing already present in the
baseline condition.
Our findings have consequences for theories of inhibition as
an age-related process. Obviously, these data do not point at an
age-related breakdown of inhibition or suppression of type
tapped by Stroop interference measures. The concept of a global
age-related deficit in inhibition has recently been challenged in
two comprehensive literature reviews (Burke, 1996; McDowd,
1996). It is quite possible that tasks other than the Stroop colorword
task do indeed demonstrate inhibition problems in older
adults, but then the inhibition account of cognitive aging should
at least be expanded to acknowledge the fact that certain aspects
of inhibition do not break down with advancing age (see
Kramer, Humphrey, Larish, & Logan, 1994, for one study comparing
age effects on different aspects of inhibition and finding
differential effects according to the task used). To the extent
that the Stroop task is indicative of frontal lobe or prefrontal
cortex functioning (Kolb & Whishaw, 1996), our findings also
challenge the frontal lobe account of cognitive aging (West,
1996), suggesting that the absence of adult age differences in the
Stroop interference effect needs to be incorporated into theories
about aging at the neuropsychological level.
At least two notes of caution are in order. First, although, to
the best of our knowledge, the present data set comprises all
relevant studies, the number of studies included here and thus
the power for finding interactions are quite limited. Also, the
fact that the mean latencies of the baseline and interference
conditions do not overlap (as can be seen by the lack of overlap
in the data clouds in Figure 1) may limit our ability to find
interactions (Fisk, Fisher, & Rogers, 1992; Kliegl, 1994). However,
the facts that the mean slope of the individual studies
matches the slope of the regression line quite well and that the
interaction analysis does not point at the presence of two distinct
theoretically acceptable functions suggest that the failure to find
a deviation from a general slowing pattern is not simply due to
a lack of statistical power. The reader may note that reliability
does not seem to be a problem in these data either: Salthouse
(1996) and Salthouse and Meinz (1995) found reliability of the
different color Stroop conditions to be .89 or higher.
Another note of caution concerns contrary evidence advanced
by Spieler et al. (1996). These authors found that the shape
of the latency distribution changes from the baseline to the
interference condition, and more so for older than for younger
adults. The change was especially notable in the right tail of
the distribution, indicating that older adults make more extremely
slow responses in the interference condition. This suggests
that in older adults, a qualitative shift in processing occurs
from the baseline to the interference condition, whereas a general
slowing account would predict a merely quantative change.
(Note, however, that we do not know whether the shift pattern
observed is really atypical; the field lacks empirical data on
what usually happens to age differences in the distribution of
latency scores when mean latency increases.) In a process-dissociation
procedure, Spieler et al. also noted an increased influence
of the word-naming process with increasing age and no age
difference in the color naming process. Thus, this study does
seem to point at qualitative differences between age groups in
the interference condition. Clearly, more and certainly more detailed
research is needed before strong final conclusions can be
reached.
Summarized, our meta-analyses strongly suggest that the presumed
age relatedness of the Stroop interference effect is merely
an artifact of general slowing. This implies that the type of
inhibition tapped by the Stroop interference effect is not vulnerable
to the effect of aging.
References
References marked with an asterisk indicate studies included
in the meta-analysis.
Berry, W. D., & Feldman. S. (1985). Multiple regression in practice.
Sage University Paper Series on Quantitative Applications in the
Social Sciences, SO (Serial No. 07-050).
Burke, D. M. (1996, April). Age-linked decline in inhibitory processes:
126 VERHAEGHEN AND DE MEERSMAN
A distributed or localized deficit? Paper presented at the Cognitive
Aging Conference, Atlanta, GA.
Cerella, J. (1990). Aging and information processing rate. In J. E. Birren
& K. W. Schaie (Eds.), Handbook of the psychology of aging
(3rd. ed., pp. 201-221). San Diego, CA: Academic Press.
Cohen, J. D, Dunbar, K., & McClelland, J. L. (1990). On net control
of automatic processes: A parallel distributed processing account of
the Stroop effect. Psychological Review, 97, 332-361.
“Conn, N. B., Dustman, R. E., & Bradford, D. C. (1984). Age-related
decrements in Stroop color test performance. Journal of Clinical
Psychology, 40, 1244-1250.
*Comalli, P. E., Wapner, S., & Werner, H. (1962). Interference effects
of Stroop color-word test on childhood, adulthood, and aging. Journal
of Genetic Psychology, 100, 47-53.
Cronbach, L., & Furby, L. (1970). How should we measure change—
or should we? Psychological Bulletin, 74, 68-80.
*Dulaney, C. L., & Rogers, W. A. (1994). Mechanisms underlying reduction
in Stroop interference with practice for young and old adults.
Journal of Experimental Psychology: Learning, Memory, and Cognition,
20, 470-484.
Dunlap, W. P., Cortina, J. M., Vaslow, J. B., & Burke, M.J. (1996).
Meta-analysis of experiments with matched groups or repeated measures
designs. Psychological Methods, 1, 170-177.
Fisk, A. D., & Fisher, D. L. (1994). Brinley plots and theories of aging:
The explicit, muddled, and implicit debates. Journal of Gerontology:
Psychological Sciences, 49, P81-89.
Fisk, A. D., Fisher, D. L., & Rogers, W. A. (1992). General slowing
alone cannot explain age-related search effects: A reply to Cerella.
Journal of Experimental Psychology: General, 121, 73-78.
•Hartman, M., & Hasher, L. (1991). Aging and suppression: Memory
for previously relevant information. Psychology and Aging, 6, 587-
594.
Hasher, L., & Zacks, R. T. (1988). Working memory, comprehension,
and aging: A review and a new view. In G. H. Bower (Ed.), The
psychology of learning and motivation (Vol. 22, pp. 193-225). San
Diego, CA: Academic Press.
Hedges, L. V., & Olkin, I. (1985). Statistical methods for meta-analysis.
Orlando, FL: Academic Press.
*Houx, P. J., Jolles, J., & Vreeling, F. W. (1993). Stroop interference:
Aging effects assessed with Stroop color-word test. Experimental
Aging Research, 19, 209-224.
“Kieley, J. M., & Hartley, A. A. (1997). Age-related equivalence of
identity suppression in the Stroop color-word task. Psychology and
Aging, 12, 22-29.
Kliegl, R. (1994, April). Individual differences in cognitive aging: An
experimental perspective. Paper presented at the Cognitive Aging Conference,
Atlanta, GA.
Kolb, B., & Whishaw, I. Q. (1996). Fundamentals of human neuropsychology
(4th ed.). New “fork: Freeman.
Kramer, A. E., Humphrey, D. G., Larish, J. F., & Logan, G. D. (1994).
Aging and inhibition: Beyond a unitary view of inhibitory processing
in attention. Psychology and Aging, 9, 491-512.
•Kwong See, S. T., & Ryan, E. B. (1995). Cognitive mediation of adult
age differences in language performance. Psychology and Aging, W,
458-468.
Laver, G. D., & Burke, D. M. ( 1993). Why do semantic priming effects
increase in old age? A meta-analysis. Psychology and Aging, 8, 34-
43.
*Li, K. Z. H., & Bosman, E. A. (1996). Age differences in Stroop-like
interference as a function of semantic relatedness. Aging, Neuropsychology,
and Cognition, 3, 272-284.
Lima, S. D., Hale, S., & Myerson, J. (1991). How general is general
slowing? Evidence from the lexical domain. Psychology and Aging,
6, 416-425.
Lindsay, D. S., & Jacoby, L. L. (1994). Stroop process dissociations:
The relation between facilitation and interference. Journal of Experimental
Psychology: Human Perception and Performance, 20, 219-
234.
MacLeod, C. M. (1991). Half a century of research on the Stroop effect:
An integrative review. Psychological Bulletin, 109, 163-203.
Maylor, E. A., & Rabbitt, P. M. A. (1994). Applying Brinley plots to
individuals: Effects of aging on performance distributions in two
speeded tasks. Psychology and Aging, 9, 224-230.
McDowd, J. M. (1996, April). Inhibition in attention and aging: Paradigms
lost? Paper presented at the Cognitive Aging Conference,
Atlanta, GA.
Monsell, S. (1996). Control of mental processes. In V. Bruce (Ed.),
Unsolved mysteries of the mind: Tutorial essays in cognition (pp.
93-148). Hove, UK: Erlbaum.
Myerson, J., Hale, S., Wagstaff, D., Poon, L. W., & Smith, G. A. (1990).
The information-loss model: A mathematical theory of age-related
cognitive slowing. Psychological Review, 97, 475-487.
Myerson, J., Wagstaff, D., & Hale, S. (1994). Brinley plots, explained
variance, and the analysis of age differences in response latencies.
Journal of Gerontology: Psychological Sciences, 49, P72-80.
*Panek, P. E., Rush, M. C., & Slade, L. A. (1984). Locus of the age-
Stroop interference relationship. Journal of Genetic Psychology, 145,
209-216.
*Park, D. C., Smith, A. D., Lautenschlager, G., Earles, J., Frieske, D.,
Zwahr, M., & Gaines, C. (1996). Mediators of long-term memory
performance across the life span. Psychology and Aging, 11, 621-
637.
Perfect, T. J. (1994). What can Brinley plots tell us about cognitive
aging? Journal of Gerontology: Psychological Sciences, 49, P60-64.
•Salthouse, T. A. (1996). General and specific speed mediation of adults
age differences in memory. Journal of Gerontology: Psychological
Sciences, 51, P30-42.
•Salthouse, T. A., & Meinz, E. J. (1995). Aging, inhibition, working
memory, and speed. Journals of Gerontology: Psychological Sciences,
SO, P297-P306.
Sliwinski, M., Buschke, H., Kuslansky, G., Senior, G., & Scarisbrick,
D. (1994). Proportional slowing and addition speed in old and young
adults. Psychology and Aging, 9, 72-80.
•Spieler, D. H., Balota, D. A., & Faust, M. E. (1996). Stroop performance
in healthy younger and older adults and in individuals with
dementia of the Alzheimer’s type. Journal of Experimental Psychology:
Human Perception and Performance, 22, 461 -479.
SPSS Inc. (1996). Statistical package for the social sciences (Version
7.5.1) [Computer software]. Chicago: Author.
Stroop, J. R. (1935). Studies of interference in serial verbal reactions.
Journal of Experimental Psychology, 18, 643-661.
*Weir, C., Bruun, C., & Barber, T. (1997). Are backward words special
for older adults? Psychology and Aging, 12, 145-149.
West, R. L. (1996). An application of prefrontal cortex function theory
to cognitive aging. Psychological Bulletin, 120, 272-292.
Received January 24, 1997
Revision received July 16, 1997
Accepted July 16, 1997 •

---

Get Professional Assignment Help Cheaply

Are you busy and do not have time to handle your assignment? Are you scared that your paper will not make the grade? Do you have responsibilities that may hinder you from turning in your assignment on time? Are you tired and can barely handle your assignment? Are your grades inconsistent?

Whichever your reason is, it is valid! You can get professional academic help from our service at affordable rates. We have a team of professional academic writers who can handle all your assignments.

Why Choose Our Academic Writing Service?

• Plagiarism free papers
• Timely delivery
• Any deadline
• Skilled, Experienced Native English Writers
• Subject-relevant academic writer
• Adherence to paper instructions
• Ability to tackle bulk assignments
• Reasonable prices
• 24/7 Customer Support
• Get superb grades consistently

Online Academic Help With Different Subjects

Literature

Students barely have time to read. We got you! Have your literature essay or book review written without having the hassle of reading the book. You can get your literature paper custom-written for you by our literature specialists.

Finance

Do you struggle with finance? No need to torture yourself if finance is not your cup of tea. You can order your finance paper from our academic writing service and get 100% original work from competent finance experts.

Computer science

Computer science is a tough subject. Fortunately, our computer science experts are up to the match. No need to stress and have sleepless nights. Our academic writers will tackle all your computer science assignments and deliver them on time. Let us handle all your python, java, ruby, JavaScript, php , C+ assignments!

Psychology

While psychology may be an interesting subject, you may lack sufficient time to handle your assignments. Don’t despair; by using our academic writing service, you can be assured of perfect grades. Moreover, your grades will be consistent.

Engineering

Engineering is quite a demanding subject. Students face a lot of pressure and barely have enough time to do what they love to do. Our academic writing service got you covered! Our engineering specialists follow the paper instructions and ensure timely delivery of the paper.

Nursing

In the nursing course, you may have difficulties with literature reviews, annotated bibliographies, critical essays, and other assignments. Our nursing assignment writers will offer you professional nursing paper help at low prices.

Sociology

Truth be told, sociology papers can be quite exhausting. Our academic writing service relieves you of fatigue, pressure, and stress. You can relax and have peace of mind as our academic writers handle your sociology assignment.

Business

We take pride in having some of the best business writers in the industry. Our business writers have a lot of experience in the field. They are reliable, and you can be assured of a high-grade paper. They are able to handle business papers of any subject, length, deadline, and difficulty!

Statistics

We boast of having some of the most experienced statistics experts in the industry. Our statistics experts have diverse skills, expertise, and knowledge to handle any kind of assignment. They have access to all kinds of software to get your assignment done.

Law

Writing a law essay may prove to be an insurmountable obstacle, especially when you need to know the peculiarities of the legislative framework. Take advantage of our top-notch law specialists and get superb grades and 100% satisfaction.

What discipline/subjects do you deal in?

We have highlighted some of the most popular subjects we handle above. Those are just a tip of the iceberg. We deal in all academic disciplines since our writers are as diverse. They have been drawn from across all disciplines, and orders are assigned to those writers believed to be the best in the field. In a nutshell, there is no task we cannot handle; all you need to do is place your order with us. As long as your instructions are clear, just trust we shall deliver irrespective of the discipline.

Are your writers competent enough to handle my paper?

Our essay writers are graduates with bachelor's, masters, Ph.D., and doctorate degrees in various subjects. The minimum requirement to be an essay writer with our essay writing service is to have a college degree. All our academic writers have a minimum of two years of academic writing. We have a stringent recruitment process to ensure that we get only the most competent essay writers in the industry. We also ensure that the writers are handsomely compensated for their value. The majority of our writers are native English speakers. As such, the fluency of language and grammar is impeccable.

What if I don’t like the paper?

There is a very low likelihood that you won’t like the paper.

Reasons being:

• When assigning your order, we match the paper’s discipline with the writer’s field/specialization. Since all our writers are graduates, we match the paper’s subject with the field the writer studied. For instance, if it’s a nursing paper, only a nursing graduate and writer will handle it. Furthermore, all our writers have academic writing experience and top-notch research skills.
• We have a quality assurance that reviews the paper before it gets to you. As such, we ensure that you get a paper that meets the required standard and will most definitely make the grade.

In the event that you don’t like your paper:

• The writer will revise the paper up to your pleasing. You have unlimited revisions. You simply need to highlight what specifically you don’t like about the paper, and the writer will make the amendments. The paper will be revised until you are satisfied. Revisions are free of charge
• We will have a different writer write the paper from scratch.
• Last resort, if the above does not work, we will refund your money.

Will the professor find out I didn’t write the paper myself?

Not at all. All papers are written from scratch. There is no way your tutor or instructor will realize that you did not write the paper yourself. In fact, we recommend using our assignment help services for consistent results.

What if the paper is plagiarized?

We check all papers for plagiarism before we submit them. We use powerful plagiarism checking software such as SafeAssign, LopesWrite, and Turnitin. We also upload the plagiarism report so that you can review it. We understand that plagiarism is academic suicide. We would not take the risk of submitting plagiarized work and jeopardize your academic journey. Furthermore, we do not sell or use prewritten papers, and each paper is written from scratch.

When will I get my paper?

You determine when you get the paper by setting the deadline when placing the order. All papers are delivered within the deadline. We are well aware that we operate in a time-sensitive industry. As such, we have laid out strategies to ensure that the client receives the paper on time and they never miss the deadline. We understand that papers that are submitted late have some points deducted. We do not want you to miss any points due to late submission. We work on beating deadlines by huge margins in order to ensure that you have ample time to review the paper before you submit it.

Will anyone find out that I used your services?

We have a privacy and confidentiality policy that guides our work. We NEVER share any customer information with third parties. Noone will ever know that you used our assignment help services. It’s only between you and us. We are bound by our policies to protect the customer’s identity and information. All your information, such as your names, phone number, email, order information, and so on, are protected. We have robust security systems that ensure that your data is protected. Hacking our systems is close to impossible, and it has never happened.

How our Assignment Help Service Works

1. Place an order

You fill all the paper instructions in the order form. Make sure you include all the helpful materials so that our academic writers can deliver the perfect paper. It will also help to eliminate unnecessary revisions.

2. Pay for the order

Proceed to pay for the paper so that it can be assigned to one of our expert academic writers. The paper subject is matched with the writer’s area of specialization.

3. Track the progress

You communicate with the writer and know about the progress of the paper. The client can ask the writer for drafts of the paper. The client can upload extra material and include additional instructions from the lecturer. Receive a paper.

4. Download the paper

The paper is sent to your email and uploaded to your personal account. You also get a plagiarism report attached to your paper.

GET A PERFECT SCORE!!!

---