© 2008 by the Board of Trustees of the University of Illinois

The Recess Debate
A Disjuncture between Educational Policy

and Scientific Research
•

Anthony D. Pellegrini

Some devalue recess because they assume it to be a waste of time. There is no theory
or empirical evidence to support this point of view. There is, however, abundant
and clear evidence that recess has beneficial effects on children’s social competence
and academic performance. The author tells how his interest in standardized tests
led him to years of recess study, compares recess survey findings in the United
States to those in the United Kingdom, and summarizes the benefits of recess for
school performance.

Recess has been part of the school day for as long as we can remem-
ber. Typically, most people have considered what children do during recess
as merely “playful.” Adults usually regard it as a break from the serious work
of the day—reading, writing, and arithmetic—while kids often say, perhaps
only half-jokingly, that it is their favorite time. Because what goes on at recess
does not appear serious, some claim it interferes with the “educational” mis-
sion of schools. This perception has led many districts to question the need
for recess.
Since I explored this trend in considerable detail three years ago in Recess:
Its Role in Education and Development, recess has remained under attack in
both the United States and the United Kingdom. The debate over recess began
around the same time (the early 1980s) in both countries and revolves around
similar issues in both places. The onslaughts against recess persist today, even
in the face of significant research supporting its educational value, a lack of
research supporting a contrary view, and a rising awareness of the importance
of play in general. Thus, it is useful to look anew at the arguments for and
against recess and to be reminded of what the evidence does and does not
show.

AMJP 01_2 text.indd 181 9/8/08 4:07:23 PM

182 A m E R I C A N J O u R N A L O F P L A Y   •   F a l l 2 0 0 8

The Argument against Recess

Breaks during the school day, like breaks from work on the factory assembly
lines, have existed for nearly as long as each of those institutions has existed.
Indeed, the rationale for breaks in both is very similar: after a reasonable amount
of work, you need a break, if for no other reason than it may help you to be
more productive. If you have never worked on an assembly line or do not
remember your primary school days, perhaps you can remember driving on a
long trip. You probably recall that the longer you drove the less attentive (and
less safe) you became. If you pulled over for a rest or a break, you were more
attentive (and safer) after you started again. This explains why many states
have laws governing the length of time truckers and airline pilots can drive or
fly without a break.
This rather simple but powerful and widely understood benefit of breaks
has not deterred a group, generally comprised of school administrators, from
reducing recess time or eliminating recess all together from the school day. The
reasons these “no nonsense” school superintendents and principals, as well as
many politicians, most often give are twofold. First, they claim that recess is a
waste of valuable time that could be more profitably used for instruction. Sec-
ond, they claim that during recess kids get bullied and that on the playground
they learn aggression.
Politicians and school administrators often use the first argument—recess
is a waste of instructional time—to demonstrate that they “mean business” in
making schools more effective. A number of years ago, then Atlanta Public
Schools superintendent Benjamin Canada and I discussed the role of recess in
schools on the Good Morning America TV show. I was touted as the “expert” on
recess, whereas Canada had made national news for proudly eliminating recess
in Atlanta schools and replacing it with physical education. He claimed that
by eliminating recess from the whole school system he had raised achievement
scores. Recess, he said, was a waste of time, and kids did not learn by “hanging on
monkey bars.” They could just as easily “blow off steam” in physical education
while at the same time learning useful skills. When pressed by both me and the
TV host for evidence of how achievement had gone up as a result of eliminating
recess, Canada did not provide supporting data, and to my knowledge no one
has ever presented data to uphold such a claim.
The evidence is exactly the opposite of Canada’s claims. As I shall summa-
rize below, in numerous controlled experiments children’s attention to school

AMJP 01_2 text.indd 182 9/8/08 4:07:24 PM

tasks decreased the longer they were deprived of a break and, correspondingly,
children were significantly more attentive after recess than before. It is very
much like taking a break on a long highway trip.
Contrary to popular belief, physical education classes do not provide such
a benefit. In 2001, the Council on Physical Education for Children, a national
organization of physical education teachers, denounced the idea of replacing
recess with physical education, although the council had a vested interested
in promoting physical education. As the council members would surely agree,
physical education—like other instructional disciplines—rightfully imposes
rigorous demands on children and adolescents so as to stretch their skills.
Therefore, it seems clear, the demands of a physical education class do not
constitute a break.
The second argument—that during recess, especially playground recess,
kids get bullied—also has flaws. It is true that kids get bullied on playgrounds,
but they get bullied in cafeterias, too, and in hallways, in bathrooms, in locker
rooms, just about anywhere with little or no adult supervision. Even so, the
base rate of aggression on playgrounds is incredibly low. Specifically, of all the
behaviors observed on preschool and primary school playgrounds in many
countries, physical and verbal aggression account for less than 2 percent of the
total (Pellegrini 1995; Smith and Connolly 1980).
The fact that rates of aggression are low at recess does not mean there are
no incidents that damage kids. Aggressive behavior can be intense even when
its rates of occurrence are low, and where there is intense aggression, people get
hurt. However, adult supervision of recess periods, like adult supervision of the
cafeteria and the hallways between classes, has a potent effect on dampening
aggression (Pellegrini 2002).
Contrary to the negative-behavior argument, recess remains one of the only
times during the school day when children have time and opportunities to interact
with their peers on their own terms. Through interaction at recess, children learn
social skills, such as how to cooperate and compromise and how to inhibit ag-
gression. Eliminating or reducing recess destroys these learning opportunities.

Why Study Recess? One Researcher’s Journey

Before examining the research in favor of recess, I should note how I came
to it. As an academic psychologist, I should be concerned with the ways in

T h e R e c e s s D e b a t e 183

AMJP 01_2 text.indd 183 9/8/08 4:07:24 PM

184 A m E R I C A N J O u R N A L O F P L A Y   •   F a l l 2 0 0 8

which children learn and develop in school. Studying recess and how children
become socially competent seems a more legitimate venue for an educational
psychologist. However, having investigated the role of children’s play in their
social and cognitive development for many years, especially play fighting (Pel-
legrini and Smith 1998; Pellegrini 2002, 2003) and the games of boys and girls
on school playgrounds (Pellegrini et al. 2002), the study of recess seemed a
logical extension.
My interest in school recess was really piqued by the debate over the role
of recess in Georgia in the early 1990s (well before Benjamin Canada’s claims
on Good Morning America) and the simultaneous use of standardized tests
as the sole criterion for the promotion of children from kindergarten to first
grade. As part of this—in my view, very questionable—venture, there was talk
of eliminating recess so kids could spend more time on the “important skills”
necessary to pass the tests. The argument went like this: test scores are declining,
and so given the limited number of hours in the school day, it makes sense to
eliminate or minimize a practice that is trivial at best and, in any case, antitheti-
cal to more serious educational enterprise.
My first reaction to the testing question was disbelief. We have known for
decades that kindergarteners are unreliable test takers (Messick 1983). Kids tend
not to perform consistently across time. For example, they could score in the
99th percentile on Tuesday, but if they retook the very same test on Wednesday,
they could score in the 65th percentile. If they took it a third time on Thursday,
they could score in the 99th percentile again. The different scores could be
due to something as simple as a swing in motivation related to a change in the
testing environment. (I observed this particular example in my own daughter’s
experience.)
Because children are unreliable test takers, it is important for educators to
use a number of different assessment strategies. Tests can and should be used,
but in conjunction with other measures, such as attendance, grades, teacher
evaluations, and observations of behavioral competence. When all of these
things are aggregated, we get a more valid picture (Cronbach 1971).
When the testing question arose in Georgia, I had been studying rough and
tumble play on the school playground during recess for several years. As part
of this research I had access to test scores from kindergarten through at least
first grade. I knew that what kids did on the playground required pretty high
levels of social cognitive competence, and I knew that kids were motivated to
implement those skills on the playground because they enjoyed interacting with

AMJP 01_2 text.indd 184 9/8/08 4:07:24 PM

their peers. So I wondered if what kindergarten children did on the playground
could be a valid predictor of their first-grade achievement, as measured by a
standardized test. That is, does kindergarten playground behavior predict first-
grade test scores, even after we control statistically for academic achievement
in kindergarten? In essence, I wanted to know if there was predictive academic
value in what kindergarteners did at recess, beyond that information provided
in their kindergarten academic achievement, as measured by a standardized
test score. How much did recess activities tell us, beyond test scores, about how
well kindergarteners would do in first grade?
My hypothesis was that the recess behavior would tell us a great deal. After
all, when kids are on the playground they are typically interacting with their
peers, and to do so takes some pretty sophisticated skills. For example, to play
cooperatively with their peers, children have to be able and willing to see things
from the perspectives of their peers, use compromise to resolve conflicts, follow
the rules of play and games, and use language to negotiate all of this. Indeed, we
know that the types of language kids use to negotiate conflicts and compromise
are very similar to the language of school instruction (Heath 1983) and the
language of literacy (Pellegrini and Galda 1982).
Further, when kids manipulate and build with playground materials and
when they play games—such as tag—with their peers, they are motivated to
marshal their social cognitive resources. Children generally like to interact
with their peers at recess, so they try their best to initiate and sustain play.
For instance, one may have to compromise (share a toy or a turn) in order to
continue to play with one’s best friend. One typically does this because one is
motivated to do so, perhaps more so than to perform on an achievement test.
Tests, at least for most young kids, are not very motivating.
These kindergarten behavioral measures that I developed and adminis-
tered did indeed predict first-grade achievement, beyond the kindergarten test
scores. That is, these playground behaviors were correlated with first-grade test
scores, even after kindergarten test scores were statistically controlled. This
reinforces the notion that multiple measures should be used in “high-stakes”
assessments.
In an effort to change policy in the state of Georgia, my friend and colleague
Carl Glickman and I wrote articles for such publications as the Atlanta Journal
Constitution and Principal to publicize our finding to the general public and
educators of young children. Afterward, testing policies changed in Georgia, but
efforts to minimize or eliminate recess continued to grow, both in the United

T h e R e c e s s D e b a t e 185

AMJP 01_2 text.indd 185 9/8/08 4:07:24 PM

186 A m E R I C A N J O u R N A L O F P L A Y   •   F a l l 2 0 0 8

States and the United Kingdom, where I was also conducting research. Policy
makers, teachers, parents, newspapers, and radio and television stations in both
countries began contacting me and asking about recess.

The Reduction of Recess in the united States
and the united Kingdom

An important barometer of prevailing perceptions of the importance of recess
is the way in which recess time has eroded across the last fifteen years. One of
the first surveys of recess in the United States was conducted in 1989 by the
National Association of Elementary School Principals (NAESP), which kindly
sent me their findings. The survey went to school superintendents in all fifty
states and the District of Columbia. Responses were received from forty-seven
states and showed that recess existed, in some form, in 90 percent of all school
districts. Generally, individual schools (87 percent of those reporting) set re-
cess policy. Consequently, there was significant variation both within school
districts and within states. Ninety-six percent of the schools with recess had it
once or twice per day. In 75 percent of the schools with recess, it lasted fifteen
to twenty minutes. The survey did not report what form that recess took or
whether organized physical education was counted as recess. Indeed, about
one-half of the districts with recess had “structured” times.
Regarding recess supervision, the survey indicated that teachers assumed
responsibility in 50 percent of the cases and teachers’ aides in 36 percent. Among
the aides, 86 percent had no formal training for supervision. This is not a trivial
finding. A well-trained supervisor can both support the positive social interac-
tions of children and guard against aggression and bullying.
Ten years later, the U.S. Department of Education surveyed recess in kin-
dergarten. According to a summary provided to the author by Ithel Jones, As-
sociate Professor of Early Childhood Education at Florida State University, 71
percent of surveyed kindergartens reported having a daily recess period; 14.6
percent had recess three to four times per week; 6.7 percent had recess one to
two times per week; and 7.7 percent had no recess. Regarding the duration of
recess, 27 percent had thirty minutes; 67 percent had sixteen to thirty minutes;
and 6 percent had less than fifteen minutes. Children attending private kinder-
gartens were twice as likely to have recess as children attending public schools:
48.3 percent vs. 22.2 percent.

AMJP 01_2 text.indd 186 9/8/08 4:07:24 PM

While a direct comparison with the 1989 survey is not possible, there are
some interesting points to consider. Most interesting is that in kindergarten only
70 percent of the children had daily recess. If there is one grade where we would
assume that all children would have recess daily, it would be kindergarten.
In the late 1990s, British psychologist Peter Blatchford and colleagues (Blatch-
ford and Sumpner 1998) conducted a national survey of recess (called “break
time” in England) in primary and secondary schools across the United Kingdom.
Their 60 percent return rate produced a sample of 6 percent of all English schools.
Importantly, recess in the United Kingdom is uniform compared to recess in the
United States. In the United Kingdom, schools have a morning, lunch, and after-
noon break. The Blatchford survey showed that while students across all grades
had breaks, the duration decreased with age. Children in infant school (five to
seven years of age) had ninety-three minutes; children in junior school (seven to
eleven years of age) had eighty-three minutes; and children in secondary school
(eleven to sixteen years of age) had seventy-seven minutes. Clearly, English chil-
dren had much more recess than their American counterparts, and the duration
of the periods seemed more sensitive to the maturity of the students.
There is, however, a movement against recess in the United Kingdom as
well. The issues propelling this movement are very similar to those in the United
States and have been very evident in the media. There, too, pressure has resulted
in a reduction in break time. Within the five-year period from 1990–1991 to
1995–1996, 38 percent and 35 percent, respectively, of junior and secondary
schools reduced the lunch break. Among infant schools, 26 percent reduced
the lunch break and 12 percent eliminated the afternoon break. Twenty-seven
percent of the junior schools and 14 percent of the secondary schools eliminated
the afternoon break.
One would think that such drastic change should be directed by empirical
support, but, no, on the contrary, research supports keeping recess in schools.

Benefits of Recess for School Performance

There are two main arguments for the continued presence of recess in pri-
mary schools. The first is evidence of how learning benefits from “distributed
practice” (like the example of taking a break during highway driving noted
earlier), which recess affords. The second concerns the development of cogni-
tive efficiency and how recess may especially facilitate learning in younger and

T h e R e c e s s D e b a t e 187

AMJP 01_2 text.indd 187 9/8/08 4:07:25 PM

188 A m E R I C A N J O u R N A L O F P L A Y   •   F a l l 2 0 0 8

cognitively immature children. Both of these arguments propose that benefits
associated with recess are immediate, that is, they occur almost simultaneously
with the recess behaviors themselves.

Massed vs. distributed practice
We have known for many years (e.g., Ebinghaus 1885; James 1901) that children
learn better and more quickly when their efforts toward a task are distributed
rather than concentrated or when they are given breaks during tasks (Hunter
1929). As psychologist Frank Dempster pointed out (1988), the positive effects
of distributed effort have been seen specifically in the ways children learn how
to conduct numerous school-like tasks, such as mastering native- and foreign-
language vocabularies, text materials, and math facts. Laboratory studies have
yielded reliable and robust findings, documenting the efficacy of task spacing
on learning. Indeed, the theory has been supported by research with humans
across the life span and with a variety of other animals.
Classroom studies have been less frequent, and generally the results less
supportive of the theory. Factors associated with the nature of a task (e.g., simple
vs. complex) seem to influence the effects of distributed practice on classroom
learning. However, when the nature of the criterion variable is changed from
material learned to attention to the task at hand, the results of the classroom
research match those of the laboratory. Spacing of tasks may make them less
boring and correspondingly facilitate attention. Attention to a task, in turn,
may be important to subsequent learning (Dempster 1988).
Given the positive effects of distributed practice on children’s attention to
school tasks, it seems puzzling that it has not been more readily used in class-
rooms. One possibility, as suggested by Dempster (1988), is that the complicated
contingencies of running a school may not readily accommodate the added
complexities of a distributed practice regimen. The solution to this conundrum
is simple—use a well-established school institution, recess. Recess provides a
break between school tasks, thus distributing practice.

Developmental differences in cognitive efficiency
Psychologist David F. Bjorklund and I have suggested previously (Pellegrini
and Bjorklund 1997), based on Bjorklund’s theory of “cognitive immaturity”
(Bjorklund and Green 1992), that the facilitative effects of breaks between peri-
ods of intense work should be greater for younger than for older children. From
our position, young children do not process most information as effectively

AMJP 01_2 text.indd 188 9/8/08 4:07:25 PM

as older children. The immaturity of their nervous systems and their lack of
experiences render them unable to perform higher-level cognitive tasks with the
same efficiency as older children and adults, and this directly influences their
educability. As a result, young children are especially susceptible to the effects
of interference and should experience the greatest gains from breaks between
focused intellectual activities, which recess provides.
Evidence in support of this hypothesis can be found in the literature on
memory and cognitive inhibition. Research using a wide range of tasks has
shown that children are increasingly able, as they get older, to inhibit task-ir-
relevant thoughts and to resist interference from task-irrelevant stimuli, and
that such skills contribute significantly to overall cognitive functioning (e.g.,
Bjorklund and Harnishfeger 1990). Inhibition abilities have been proposed to
play a significant role in attention, permitting children to focus on task-relevant
information and not to be distracted by task-irrelevant, peripheral information.
Such abilities have also been proposed to be of central importance to functional
working-memory capacity. Young children have a difficult time keeping extra-
neous information from entering short-term store. As a result, their working
memories are often cluttered with irrelevant information, leaving less mental
space for task-relevant information or for the execution of cognitive strategies
(Bjorklund and Harnishfeger 1990).
From this perspective, there may be a general increase in interference when
children perform a series of highly focused tasks, regardless of the nature of
those tasks. Although one would predict that changing from one type of focused
activity to another would yield some cognitive benefit, children (especially
young children) may experience a continued buildup of interference with re-
peated performance of even different highly focused tasks, and thus experi-
ence greater benefit from a drastic change in activity, such as is afforded by
recess. This is consistent with the evidence that younger children may require
a greater change in activity or stimulus materials before they experience a re-
lease from interference (e.g., Pellegrini and Bjorklund 1996). This should make
school learning particularly difficult for young elementary school children, and
opportunities to engage in non-focused, nonintellectual activities should af-
ford them the needed respite to re-energize their nervous systems so that they
can continue to learn in school. Consistent with this reasoning, recess periods
across the school day should minimize cognitive interference. Importantly,
instructional regimens, such as physical education, would not serve the same
purpose.

T h e R e c e s s D e b a t e 189

AMJP 01_2 text.indd 189 9/8/08 4:07:25 PM

190 A m E R I C A N J O u R N A L O F P L A Y   •   F a l l 2 0 0 8

Conclusion

Some devalue recess because they assume it to be—as they assume play in
young children to be—a waste of time, time that could be otherwise more ef-
ficiently spent. There is no theory or empirical evidence to support this point
of view. The counter-argument, that recess is good, is backed by a large body
of theory and empirical research. Those who advocate the elimination of recess
should present sound theoretical and empirical support for their arguments
or give them up and recognize the abundant and clear evidence that recess
has beneficial effects on children’s social competence and academic perfor-
mance.

References

Bjorklund, David F. 1978. Negative transfer in children’s recall of categorized materials.
Journal of Experimental Child Psychology 26:299–307.

———. 2004, first published 1989. Children’s thinking: Developmental function and
individual differences. 4th ed.

Bjorklund, David F., and Brandi L. Green. 1992. The adaptive nature of cognitive im-
maturity. American Psychologist 47:46–54.

Bjorklund, David F., and Katherine K. Harnishfeger. 1990. The resources construct
in cognitive development: Diverse sources of evidence and a theory of inefficient
inhibition. Developmental Review 10:48–71.

Bjorklund, David F., and Anthony D. Pellegrini. 2002. The origins of human nature:
Evolutionary developmental psychology.

Blatchford, Peter, and Clare Sumpner. 1998. What do we know about breaktime? Re-
sults from a national survey of breaktime and lunchtime in primary and secondary
schools. British Educational Research Journal 24:79–94.

Cronbach, Lee J. 1971. Test validation. In Educational measurement, ed. Robert L.
Thorndike, 443–507.

Dempster, Frank N. 1988. The spacing effect. American Psychologist 43:627–34.
Ebbinghaus, Hermann. 1964, first published 1885. Memory: A contribution to experi-

mental psychology.
Heath, Shirley. 1983. Ways with words: Language, life, and work in communities and

classrooms.
Hunter, Walter S. 1929. Learning II: Experimental studies of learning. In The founda-

tions of experimental psychology, ed. Carl Murchison, 564–627.
James, William. 1901. Talks to teachers on psychology, and to students on some of life’s

ideals.

AMJP 01_2 text.indd 190 9/8/08 4:07:25 PM

Messick, Sam. 1983. Assessment of children. In Handbook of child psychology. Vol. 1,
History, theory, and methods, ed. William Kessen, 477–526.

Pellegrini, Anthony D. 1990. Elementary school children’s playground behavior: Im-
plications for children’s social-cognitive development. Children’s Environments
Quarterly 7:8–16.

———. 1991. Outdoor recess: Is it really necessary? Principal 70:40.
———. 1992. Kindergarten children’s social-cognitive status as a predictor of first-grade

success. Early Childhood Research Quarterly 7:565–77.
———. 1995. School recess and playground behavior: Educational and developmental

roles.
———. 2002. Bullying, victimization, and sexual harassment during the transition to

middle school. Educational Psychologist 37:151–63.
———. 2003. Perceptions and functions of play and real fighting in early adolescence.

Child Development 74:1522–33.
———. 2005. Recess: Its role in education and development.
Pellegrini, Anthony D., and David F. Bjorklund. 1996. The place of recess in school:

Issues in the role of recess in children’s education and development: An introduc-
tion to the theme of the Special Issue. Journal of Research in Childhood Education
11:5–13.

———. 1997. The role of recess in children’s cognitive performance. Educational Psy-
chologist 32:35–40.

Pellegrini, Anthony D., and Lee Galda. 1982. The effects of thematic-fantasy play train-
ing on the development of children’s story comprehension. American Educational
Research Journal 19:443–52.

Pellegrini, Anthony D., Kentaro Kato, Peter Blatchford, and Ed Baines. 2002. A short-
term longitudinal study of children’s playground games across the first year of
school: Implications for social competence and adjustment to school. American
Educational Research Journal 39:991–1015.

Pellegrini, Anthony D., and Peter K. Smith. 1998. Physical activity play: The nature and
function of a neglected aspect of play. Child Development 69:577–98.

Smith, Peter K., and Kevin Connolly. 1980. The ecology of preschool behavior.

T h e R e c e s s D e b a t e 191

AMJP 01_2 text.indd 191 9/8/08 4:07:26 PM