Learning (Not) to Talk About Race: When Older Children Underperform
in Social Categorization
Evan P. Apfelbaum, Kristin Pauker, Nalini Ambady,
and Samuel R. Sommers
Michael I. Norton
Harvard Business School
The present research identifies an anomaly in sociocognitive development, whereby younger children (8
and 9 years) outperform their older counterparts (10 and 11 years) in a basic categorization task in which
the acknowledgment of racial difference facilitates performance. Though older children exhibit superior
performance on a race-neutral version of the task, their tendency to avoid acknowledging race hinders
objective success when race is a relevant category. That these findings emerge in late childhood, in a
pattern counter to the normal developmental trajectory of increased cognitive expertise in categorization,
suggests that this anomaly indicates the onset of a critical transition in human social development.
Keywords: social development, categorization, regulatory behavior, stereotype knowledge, social norms
In the typical trajectory of cognitive development, children tend
to improve on problem-solving tasks as they get older due to,
among other factors, superior memory capacity, more effective
information processing, and greater facility with categorization
processes (Case & Okamoto, 1996; Inhelder & Piaget, 1958; Kail,
better than older children on any cognitive task. We report one such
anomaly in which younger children outperform older children on a
social categorization task in which performance efficiency comes at
the expense of acknowledging racial difference.
As any parent who has experienced the social discomfort of a
child unselfconsciously and publicly pointing out a stranger’s race
can attest, talking about race—even in the context of an accurate,
diagnostic physical description—is often considered normatively
inappropriate (Norton, Sommers, Apfelbaum, Pura, & Ariely,
2006). It is clear, however, that children become aware of social
categories such as race at a very young age. Infants as young as 6
months old can perceptually discriminate between racial groups
(Katz & Kofkin, 1997), by preschool most children can accurately
identify others’ membership in racial categories (Aboud, 2003),
and at the age of 5 children begin to demonstrate knowledge of
some social stereotypes (Ambady, Shih, Kim, & Pittinsky, 2001).
Awareness of moral and social conventions, however, develops
somewhat later. By 10 years of age, most children have internal-
ized social and moral norms (Piaget, 1932; Turiel, 2006) and can
effectively take the perspective of others and engage in self-
presentation (Aloise-Young, 1993; Banerjee, 2002; Carpendale &
Chandler, 1996; Selman, 1980). These developments are accom-
panied by a greater understanding of abstract ideas, which in the
domain of race include discrimination and the proscription of
prejudice (Brown & Bigler, 2005; McKown & Weinstein, 2003;
Rutland, 2004; Rutland, Cameron, Milne, & McGeorge, 2005;
Verkuyten, Kinket, & van der Wielen, 1997).
The present research examines age group differences in chil-
dren’s tendency to frankly acknowledge versus strategically avoid
race. We sought to investigate whether, after approximately 10
years of age, differences emerge in children’s performance on a
task in which race is a relevant category. Since such differences
would coincide with the point at which children possess both
knowledge of norms regarding prejudice and the cognitive capac-
ity to regulate behavior, the demonstration of this variation would
suggest the need for developmental research elucidating bases of
this possible transition. Accordingly, we investigated how children
aged 8 to 11 years performed on a social categorization task.
The task was a photo identification game in which describing
people by category memberships (e.g., gender, race) facilitates
task performance. We expected that older children, by virtue of
their being more facile at social categorization, would outperform
younger children, except in one key instance: when the race of the
individuals pictured was a salient category. In this instance, we
expected that older children’s desires to follow prevailing social
and moral conventions in the United States and to appear unprej-
udiced (e.g., Gaertner & Dovidio, 1986) would prompt them to
forgo the use of this category, even at the expense of task perfor-
mance. This expectation followed from recent work demonstrating
that those adults most concerned with appearing unprejudiced tend
to strategically avoid talking about race during social interaction
(Apfelbaum, Sommers, & Norton, in press). In short, we hypoth-
esized that at approximately 10 years of age, social concerns
regarding the acknowledgment of race would override the tradi-
tional age group differences regarding improved performance on
Evan P. Apfelbaum, Kristin Pauker, Nalini Ambady, and Samuel R.
Sommers, Department of Psychology, Tufts University; Michael I. Norton,
Harvard Business School.
This research was supported by a Jacob K. Javits Fellowship, a grant
from the Russell Sage Foundation, and National Science Foundation Grant
BCS-0724416. We are extremely grateful to all the children, parents,
teachers, and administrators who helped make this project possible. We
also thank Richard Lerner, Donald Wertlieb, and Dan Ariely for their
valuable comments on earlier versions of this article. Evan P. Apfelbaum
and Kristin Pauker contributed equally to this research.
Correspondence concerning this article should be addressed to Evan P.
Apfelbaum or Kristin Pauker, Department of Psychology, Tufts University,
490 Boston Avenue, Medford, MA 02155. E-mail: evan.apfelbaum@
tufts.edu or email@example.com
2008, Vol. 44, No. 5, 1513–1518
Copyright 2008 by the American Psychological Association
categorization-based tasks. Such an anomalous pattern of behavior
would suggest a critical milestone in human social development,
requiring subsequent longitudinal research.
Participants and Design
We recruited 101 children from three suburban public elemen-
tary schools that serve middle-class and upper middle-class fam-
ilies outside Boston, Massachusetts. The sample comprised two
contiguous age groups: 8- and 9-year-olds (M ? 9.26 years, SD ?
0.40) and 10- and 11-year-olds (M ? 10.75 years, SD ? 0.38). The
younger age group included 51 children (26 girls), and the older
age group included 50 children (25 girls). Participants were ran-
domly assigned to a race-relevant or race-neutral condition, yield-
ing a 2 (age group: younger vs. older) ? 2 (task type: race-relevant
vs. race-neutral) between-subjects design. Participants were pre-
dominantly White (86.1%)1and none was Black; each child re-
ceived a small gift for participating.
Materials and Procedure
Parents of third, fourth, and fifth graders were informed of the
study via letters sent home by school administrators. Upon receipt
of parental consent, individual children who provided verbal assent
participated in a quiet location, separate from other children. The
only information they received prior to testing was that they would
be “playing matching games.” The experimental task—adapted
from an analogous task used to gauge adults’ willingness to
mention race (Apfelbaum et al., in press; Norton et al., 2006)—
was an interactive photo identification exercise similar to a chil-
dren’s board game. Children sat in front of an array of forty
4 ? 6-in. (10.16 ? 15.24-cm) photographs of people (cropped
above the waist) arranged in 4 rows of 10. Photos in the array
differed in many respects but only varied systematically on 4
dimensions. For the race-relevant condition, these dimensions
were background color (red/blue), gender (male/female), weight
(fat/thin), and race (Black/White). Participants assigned to the
race-neutral condition viewed photos that varied on the same
dimensions, except that all of their photos depicted White individ-
uals. Instead, a dark brown or off-white, head-sized oval sticker
was appended to the bottom-left corner of each photo, reestablish-
ing a fourth dimension on which the array varied systematically.
The sticker colors were not arbitrary—the dark brown was a
composite blend of the Black individuals’ actual skin tone from the
race-relevant array, while off-white stickers displayed the average
White skin tone from the same array. Thus, the race-relevant and
race-neutral photo arrays displayed virtually identical perceptual
information, but only in the former condition was this information
tied to the social category of race (see Figure 1 for sample photos).
An experimenter sat facing the participant, holding a single
target photo from the array. This target photo was one of two
photos, randomly selected from the array at the outset of the study.
Participants were instructed to ask yes/no questions to narrow the
array down to the single target photo held by the experimenter.
Their primary goal—to use the fewest questions necessary—was
explicitly stated twice during the course of the instructions and
then repeated a third time moments before the task begun. Thus,
participants were clearly aware that their objective was to strive for
performance efficiency, an effort facilitated by acknowledgment of
categorical differences in race or sticker color. While the experi-
menter turned on the video camera, children familiarized them-
selves with the array until they felt prepared to begin the interac-
tion. After completing the task, children completed a manipulation
check that asked them to indicate whether they noticed that the
photos in the array varied by race or sticker color.
1Non-White (6 Asian, 6 biracial, and 2 Latino) participants were in-
cluded in the reported analyses. Results did not significantly differ when
analyses were restricted to White participants only.
while stimuli in the race-neutral condition varied by sticker color (bottom
panel). Clockwise from the top left, photographs are from, respectively,
people//man-6-4.htm, and http://flickr.com/photos/technowannabe/
164515893/sizes/o/. The bottom right photo is in the public domain; all other
photos are adapted according to a Creative Commons license
Stimuli in the race-relevant condition varied by race (top panel),
There were two primary outcomes of interest. We examined
between-groups differences in objective performance, based on the
total number of questions needed to complete the task, and the
frequency with which participants asked questions about categor-
ical differences in race or sticker color, depending on the condition
to which they had been assigned.
Scores reflecting the total number of questions asked (with
lower values denoting better performance) were submitted to a
two-way analysis of variance. This analysis indicated a main effect
of task type such that participants in the race-relevant condition
(M ? 7.86, SD ? 1.82) performed relatively less efficiently than
participants in the race-neutral condition (M ? 6.64, SD ? 1.37),
F(1, 97) ? 15.53, p ? .001, r ? .37. The main effect of age group
was not significant (F ? 1).
As displayed in Figure 2, these effects were qualified by a
significant Age Group ? Task Type interaction, F(1, 97) ? 8.17,
p ? .005, r ? .28. In the race-neutral condition, as expected,
participants in the older group (M ? 6.17, SD ? 1.07) outper-
formed participants in the younger group (M ? 7.04, SD ? 1.48),
t(97) ? 1.94, p ? .05, r ? .19. In the race-relevant condition,
however, this pattern was reversed as participants in the younger
group (M ? 7.38, SD ? 1.10) outperformed participants in the
older group (M ? 8.30, SD ? 2.22), t(97) ? 2.10, p ? .04, r ? .21.
Such a pattern is remarkable considering that, by virtue of being
older and thus, more cognitively mature, older children ought to
outperform their younger counterparts (as they did in the race-
neutral condition). Not only did presentation of a race-relevant task
negate this inherent performance gap, but it reversed it. These
results support the hypothesis that, at approximately 10 years of
age, children’s tendency to regulate the appearance of prejudice is
powerful enough to undermine performance on a task rooted in
basic cognitive skills.
Acknowledgments of Difference
We further investigated this shift in behavior by assessing the
frequency with which participants acknowledged race or sticker
color. Our a priori definition for the acknowledgment of race
included specific references to racial groups such as Black, African
American, White, or Caucasian, as well as race-related terminol-
ogy (e.g., “Does your person have brown skin?”). Similarly, we
recognized any acknowledgment of a perceptual difference be-
tween sticker types (e.g., “Does your person have a brown
sticker?”). We then recoded these measures as a single variable—
acknowledgment of difference—to allow for analysis across ex-
As displayed in Figure 3, in the race-neutral condition, chi-
square analysis indicated that the frequency with which partic-
ipants acknowledged difference (i.e., in sticker color) was high
and virtually identical among younger children (M ? 77.8%,
SD ? 42.4) and older children (M ? 78.3%, SD ? 42.2),
suggesting that participants believed that color was a useful
category for completing the task. However, in the race-relevant
condition, while a full 76.5% (SD ? 42.8) of children in the
younger group acknowledged difference (i.e., race), only 37.0%
(SD ? 45.8) of the older group did so, ?2(3, N ? 101) ? 14.23,
p ? .005, ? ? .38. A planned contrast confirmed that the older
group’s reluctance to ask about race represented a marked
divergence in behavior relative to the general pattern displayed
by participants in the other three conditions, ?2(1, N ? 101) ?
14.48, p ? .0001, ? ? .38. When acknowledgment of race
was measured with even more conservative coding
criteria—explicit use of the terms Black and African Ameri-
can—an equally compelling result emerged: While 33.3%
(SD ? 46.3) of participants in the younger group asked ques-
tions including the terms Black and African American, not a
single participant in the older age group (0%) did so, ?2(1, n ?
51) ? 10.67, p ? .001, ? ? ?.46.2The final self-report
measures confirmed that every participant reported noticing
differences (either based on race or sticker color) in the array.
In sum, these findings lend support to the proposed performance
difference between children younger and older than 10 years of
age, as the majority of older children in a race-relevant setting
avoided mentioning race, even though doing so undermined their
2As a comparison, we investigated children’s mention of two other
stigmatized social categories depicted in the photo array: gender and
weight. Chi-square analyses (collapsing across race-relevant and race-
neutral conditions) indicated no significant differences between
younger and older age groups for mention of gender, ?2(1, N ? 101) ?
.21, p ? .65, or weight, ?2(1, N ? 101) ? .20, p ? .74. Thus, this
strategy of avoidance seems particularly relevant to race-related con-
siderations, a conclusion likely driven by the extreme social undesir-
ability of being stigmatized as a racist in the United States (Sommers &
8- and 9-year-olds
10- and 11-year-olds
Total questions asked
task type and age group, where fewer questions asked indicates better
performance. Error bars represent standard error.
Total questions needed to complete the task as a function of
task performance. Indeed, analyses provided evidence that ac-
knowledgment of difference in the race-relevant setting mediated
the performance decrements exhibited by older versus younger
children.3And more generally, the negative relationship between
acknowledgment of difference and total questions asked across
conditions, r(99) ? ?.27, p ? .01, indicates that participants who
avoided acknowledging race or sticker color did so at the expense
of objective success on the task.
infants to outperform older infants in several domains, including
music perception (Hannon & Trehub, 2005), and face perception
(Kelly et al., 2005; Pascalis, de Haan, & Nelson, 2002), such a
counterintuitive pattern of behavior has seldom been documented in
older children. The most notable exception has been demonstrated in
the classic problem-solving exercise involving functional fixedness,
wherein adults and older children perform worse than younger chil-
dren on tasks that require them to utilize tools in a manner atypical of
their traditional function (Duncker, 1945; German & Defeyter, 2000).
We suspect that just as increased knowledge of the typical function of
tools sets the stage for the anomaly of functional fixedness, so, too,
may increased knowledge regarding racial stereotypes and norms
predict the tendency for older children to avoid race. Indeed, in a
follow-up task administered to the present sample, participants’ re-
sponses to a measure of stereotype knowledge indicated that their
awareness of racial stereotypes was negatively associated with the
tendency to acknowledge race in the earlier race-relevant photo task,
r(44) ? ?.33, p ? .03; no such relationship existed among partici-
pants in the race-neutral version, r(48) ? .13, p ? .37.4In other
words, children who exhibited greater awareness of racial stereotypes
tended to be the ones who were most likely to avoid mentioning race
in the categorization task. To the extent that such knowledge of racial
stereotypes relates to a more general understanding of contemporary
race-related norms, it may represent one factor that sets the stage for
the behavioral effects presented here.
In short, though cross-sectional, our work is consistent with the
presence of a developmental anomaly: that a consequence of
increased understanding of norms pertaining to race is the ten-
dency to avoid acknowledging race altogether. To the extent that
these results are pertinent to age changes as well as age group
differences, they suggest that at some point after mastering cate-
gorization and generalization across classes of stimuli, children
learn that by applying these same principles to people—describing
others on the basis of skin color—they risk appearing prejudiced
and might receive social sanctions (Pollock, 2004; Schofield,
1989). While such labeling of people may be both correct as a
category extension and useful for distinguishing people from one
another, mentioning race is often considered inappropriate in so-
cial discourse (Norton et al., 2006). As a result, it may be that
younger children who have not yet internalized such social con-
ventions and concerns about appearing prejudiced are able to
outperform older children on a task for which the acknowledgment
of race is a relevant consideration. Though these findings cannot
3We used a series of regression analyses (Baron & Kenny, 1986) to
test the prediction that acknowledgement of difference would mediate
the relationship between age group and performance in the race-related
condition. Given our relatively small sample size for this type of
analysis, the most appropriate model was a bias-corrected bootstrap
mediation (Efron & Tibshirani, 1993; Preacher & Hayes, 2004; Shrout
& Bolger, 2002). Using 1,000 resamples, these analyses revealed that
age group was a positive predictor of total questions (? ? .26, p ? .07),
a relationship that was reduced when acknowledgment of difference
was added as a predictor (? ? .16, p ? .28). This indirect effect was
statistically significant (95% confidence interval: .0045, 1.11), indicat-
ing that acknowledgment of difference mediated the relationship be-
tween age group and task performance.
4After the categorization task described above, participants were led
of male children. In consecutive fashion, participants were asked to select the
child in each pair most likely to have exhibited 1 of 13 narrated behavioral
episodes. Nine target episodes described behaviors that typify prevalent neg-
ative racial stereotypes, while four nonstereotypical fillers described race-
neutral behaviors. For stereotypical episodes, we always displayed one photo
of a child belonging to the racial group targeted by the stereotype, and the
second photo was randomly selected. The second photo was chosen from
among three racial groups (Black, Asian, or White), with the stipulation that
the photo chosen did not belong to the same racial group as the target photo.
This ensured that participants who belonged to one of the racial groups in the
pair were not always making a decision that involved an ingroup/outgroup
choice. Results did not differ among children who saw two outgroup members
or one ingroup and one outgroup member. For filler episodes, the racial
composition of the pairs was randomized. Participants’ selections regarding
each of the nine target episodes were then combined to form a single index of
negative stereotype knowledge. No significant difference emerged in stereo-
type knowledge between age groups, t(94) ? 0.60, p ? .56. In addition to
significant negative relationship between the acknowledgment of race and
stereotype knowledge reported in the text, an even more robust relationship
emerged when considering just the terms Black or African American, r(44) ?
–.44, p ? .005, as the tendency to avoid these terms was even more strongly
associated with knowledge of racial stereotypes.
8- and 9-year-olds
10- and 11-year-olds
was acknowledged (%)
function of task type and age group. Error bars represent standard error.
Frequency with which color difference was acknowledged as a
be extrapolated to behavior exhibited by Black children—an im-
portant consideration for future study—the data presented here are
sufficiently provocative to merit longitudinal research investigat-
ing such interpretations and, more generally, the possibility of
substantive transitions over the course of development.
The present findings also speak to broader issues regarding the
development of intergroup attitudes and relations. The extant litera-
ture typically reports a decrease in biased intergroup attitudes as
children increase in age, yet there is no agreed upon explanation for
why this pattern emerges. Some suggest that this decline reflects a
genuine decrease in prejudice resulting from increased perspective
taking during the acquisition of concrete operations (Aboud, 1988;
Doyle & Aboud, 1995; Doyle, Beaudet, & Aboud, 1988), yet others
suggest it represents an increasing desire to adhere to societal norms
and engage in self-presentation (Nesdale, 2004; Rutland, 2004). To
the extent that our findings indicate a possible transition at 10 years of
age—by some accounts the earliest point at which children possess
both the cognitive skills to perspective take and autonomous motiva-
tion to respond to social norms (Brown & Bigler, 2005; Crandall,
Eshelman, & O’Brien, 2002; Rutland et al., 2005)—they suggest an
integrative account by which some degree of both capacities are
necessary to explain the decrease in biased intergroup attitudes. In-
deed, while children younger than 10 years of age may already
possess the cognitive tools needed to engage in such regulation, the
present study suggests that only when integrated with emerging mo-
tivations to adhere to social norms can these skills be harnessed for
strategic purposes. Such an interpretation carries with it important
practical implications, namely that interventions aimed at improving
intergroup relations by promoting genuine egalitarianism (as distinct
from simply controlling prejudice) might be particularly relevant to
children around 10 years of age.
In closing, some may herald the emergence of the tendency to
avoid acknowledging race as evidence of progress made toward
social equality—the realization that race doesn’t matter. Chil-
dren do notice race, however, and failing to acknowledge as
much may have negative social consequences in addition to the
performance deficits identified herein (Katz, 2003; Richeson &
Nussbaum, 2004; Schofield, 2007). Regardless, the anomaly in
task performance demonstrated in the present study may point
to the onset of an important transition in human social devel-
opment at 10 years of age, when internalized social and moral
norms begin to regulate behavior, even when such regulation
comes at a cost.
Aboud, F. (1988). Children and prejudice. Oxford, England: Blackwell.
Aboud, F. R. (2003). The formation of in-group favoritism and out-group
prejudice in young children: Are they distinct attitudes? Developmental
Psychology, 39, 48–60.
Aloise-Young, P. A. (1993). The development of self-presentation: Self-
promotion in 6- to 10-year-old children. Social Cognition, 11, 201–222.
Ambady, N., Shih, M., Kim, A., & Pittinsky, T. L. (2001). Stereotype
susceptibility in children: Effects of identity activation on quantitative
performance. Psychological Science, 12, 385–390.
Apfelbaum, E. P., Sommers, S. R., & Norton, M. I. (in press). Seeing race
and seeming racist? Evaluating strategic colorblindness in social inter-
action. Journal of Personality and Social Psychology.
Banerjee, R. (2002). Audience effects on self-presentation in childhood.
Social Development, 11, 487–507.
Baron, R. M., & Kenny, D. A. (1986). The moderator–mediator variable
distinction in social psychological research: Conceptual, strategic, and
statistical considerations. Journal of Personality and Social Psychology,
Brown, C. S., & Bigler, R. S. (2005). Children’s perceptions of discrimi-
nation: A developmental model. Child Development, 76, 533–553.
Carpendale, J. I., & Chandler, M. J. (1996). On the distinction between
false belief understanding and subscribing to an interpretive theory of
mind. Child Development, 67, 1686–1706.
Case, R., & Okamoto, Y. (1996). The role of central conceptual structures
in the development of children’s thought. Monographs of the Society for
Research in Child Development, 61(1–2, Serial No. 246).
Crandall, C., Eshelman, A., & O’Brien, L. (2002). Social norms and the
expression and suppression of prejudice: The struggle for internalization.
Journal of Personality and Social Psychology, 82, 359–378.
Doyle, A. B., & Aboud, F. E. (1995). A longitudinal study of White
children’s racial prejudice as a social–cognitive development. Merrill-
Palmer Quarterly, 41, 209–228.
Doyle, A. B., Beaudet, J., & Aboud, F. E. (1988). Developmental patterns
in the flexibility of children’s ethnic attitudes. Journal of Cross-Cultural
Psychology, 19, 3–18.
Duncker, K. (1945). On problem-solving. Psychological Monographs, 58,
Efron, B., & Tibshirani, R. J. (1993). An introduction to the bootstrap.
Boca Raton, FL: Chapman & Hall.
Gaertner, S. L., & Dovidio, J. F. (1986). The aversive form of racism. In
J. F. Dovidio & S. L. Gaertner (Eds.), Prejudice, discrimination, and
racism (pp. 61–89). San Diego, CA: Academic Press.
German, T. P., & Defeyter, M. A. (2000). Immunity to functional fixedness
in young children. Psychonomic Bulletin and Review, 7, 707–712.
Hannon, E. E., & Trehub, S. E. (2005). Tuning in to musical rhythms:
Infants learn more readily than adults. Proceedings of the National
Academy of Sciences, USA, 102, 12639–12643.
Inhelder, B., & Piaget, J. (1958). The growth of logical thinking: From
childhood to adolescence. New York: Basic Books.
Kail, R. (2004). Cognitive development includes global and domain-
specific processes. Merrill-Palmer Quarterly, 50, 445–455.
Katz, P. A. (2003). Racists or tolerant multiculturalists? How do they
begin? American Psychologist, 58, 897–909.
Katz, P. A., & Kofkin, J. A. (1997). Race, gender, and young children. In
S. S. Luthar, J. A. Burack, D. Cicchetti, & J. R. Weisz (Eds.), Devel-
opmental psychopathology: Perspectives on adjustment, risk, and dis-
order (pp. 51–74). New York: Cambridge University Press.
Kelly, D. J., Quinn, P. C., Slater, A. M., Lee, K., Gibson, A., Smith, M.,
et al. (2005). Three-month-olds, but not newborns, prefer own-race
faces. Developmental Science, 8, F31–F36.
McKown, C., & Weinstein, R. S. (2003). The development and conse-
quences of stereotype consciousness in middle childhood. Child Devel-
opment, 74, 498–515.
Nesdale, D. (2004). Social identity processes and children’s ethnic preju-
dice. In M. Bennett & F. Sani (Eds.), The development of the social self
(pp. 219–245). New York: Psychology Press.
Norton, M. I., Sommers, S. R., Apfelbaum, E. P., Pura, N., & Ariely, D.
(2006). Colorblindness and interracial interaction: Playing the political
correctness game. Psychological Science, 17, 949–953.
Pascalis, O., de Haan, M., & Nelson, C. A. (2002, May). Is face processing
species-specific during the first year of life? Science, 296, 1321–1323.
Piaget, J. (1932). The moral judgment of the child. Oxford, England:
Pollock, M. (2004). Colormute: Race talk dilemmas in an American
school. Princeton, NJ: Princeton University Press.
Preacher, K. J., & Hayes, A. F. (2004). SPSS and SAS procedures for
estimating indirect effects in simple mediation models. Behavior Re-
search Methods, Instruments, and Computers, 36, 717–731.