RESEARCH ARTICLE

Association of Education Outside the

Classroom and Pupils’ Psychosocial
Well-Being: Results From a School Year
Implementation
MADS BØLLING, MEd PhDa JANNI NICLASEN, MSc, PhDb PETER BENTSEN, MSc, PhDc GLEN NIELSEN, MSc, PhDd

ABSTRACT
BACKGROUND: Education Outside the Classroom (EOtC) is a teaching method that is gaining traction, aiming to promote
learning and well-being. However, research on the association between EOtC and well-being is limited.
METHODS: This quasi-experimental trial involved pupils (9-13 years) from 16 Danish public schools which implemented EOtC
in some classes. Pupils (N = 511) from 27 classes were regularly exposed to EOtC (2-7 hours per week), and pupils (N = 120)
from 7 parallel comparison classes were much less exposed (less than 2 hours per week). The pupils’ psychosocial well-being
was measured at the beginning and end of the school year using the Strengths and Difficulties Questionnaire. The results were
compared between the groups.
RESULTS: Pupils regularly exposed to EOtC showed the greatest improvement in prosocial behavior. Negative associations
were seen between EOtC and hyperactivity-inattention and peer problems in pupils of low socioeconomic status (SES). The
observed improvements were smaller when the EOtC was spread over a larger number of sessions. Results were independent
of sex.
CONCLUSIONS: Regular exposure to EOtC was found to promote social well-being, especially for pupils of low SES, and was
most beneficial when concentrated in fewer, longer sessions.

Keywords: mental health; outdoor learning; school-based outdoor education; school-based health promotion; udeskole.
Citation: Bølling M, Niclasen J, Bentsen P, Nielsen G. Association of education outside the classroom and pupils’ psychosocial
well-being: results from a school year implementation. J Sch Health. 2019; DOI: 10.1111/josh.12730

E ducation Outside the Classroom (EOtC) is a

teaching method that is gaining traction in
Scandinavia, Germany, UK, United States, and other
Western countries.1 It can broadly be defined as
relocating standard curriculum teaching to places
outside the buildings and walls of the schools for
a single or a few days per week as a supplement
to indoor classroom teaching.2 Places used in EOtC
include forests, school gardens, and museums.3-5
EOtC creates variation in school days and typically
aims to promote learning, positive social relations,
motivation for school, and psychosocial well-being.6,7
EOtC often includes activities described as child-
led, hands-on problem solving and experimentation,
peer collaboration, and inductive learning approaches,
as well as physical activity, games, and play.8
Although EOtC can be regarded as a school-based
health initiative integrated into academic curricular

a PhD Student, (mads.boelling@regionh.dk), Department of Nutrition, Exercise and Sports (NEXS), University of Copenhagen, Nørre Allé 51, DK-2200 Copenhagen N, Denmark.
bAssistant Professor, (janni.niclasen@psy.ku.dk), Department of Psychology, University of Copenhagen, Øster Farimagsgade 2A, 1353 Copenhagen K, Denmark.
c Senior Researcher, (peter.bocz.bentsen@regionh.dk), Health Promotion, Steno Diabetes Center Copenhagen, The Capital Region of Denmark, Niels Steensens Vej 6, DK-2820
Gentofte, Denmark.
d
Associate Professor, (gnielsen@nexs.ku.dk), Department of Nutrition, Exercise and Sports (NEXS), University of Copenhagen, Nørre Allé 51, DK-2200 Copenhagen N, Denmark.
Address correspondence to: Mads Bølling, Postdoc, (mads.boelling@regionh.dk), Health Promotion, Steno Diabetes Center Copenhagen, The Capital Region of Denmark, Niels
Steensens Vej 6, DK-2820 Gentofte, Denmark.
We would like to express our greatest thankfulness to the participating children, parents, teachers, and schools for their cooperation. In addition, we would like to thank our
colleagues at the TEACHOUT project and CHESS research group for their feedback. We appreciate the thorough revision of the final manuscript by Lærke Mygind. This study was
funded by the Danish foundation TrygFonden (ID #102171). The funders have not been involved in the study design, analyses, interpretation, writing, or the decision to submit this
paper. The authors declare that they have no potential conflict of interest in the research.

Journal of School Health • 2019 • © 2019, American School Health Association • 1

teaching,9,10 EOtC is distinguished by not involving
explicit health education.
Some schools choose EOtC as a method for the
whole school, whereas in others, individual teachers
take it upon themselves to use this method with their
class. Teachers often start practicing EOtC in order
to bridge the gap between high and low achieving
pupils,11 and as an inclusive teaching method for pupils
experiencing emotional and behavioral difficulties,12
because it involves body-based learning approaches,13
and seems a viable approach to meeting the needs of
those pupils that find it hard to cope with the require-
ments of traditional teaching and learning activities.14
Surveys have shown that in Danish public schools
weekly or biweekly provision of EOtC has increased
from at least 14% in 2007 to at least 17.9% in 2014.15
A review of research into EOtC showed that most
research thus far has been qualitative and few studies
have directly related EOtC to pupils’ well-being.1
In these studies, it was argued that time spent in
games and play outside the classroom promoted more
positive social interactions between pupils.3,16 Others
highlighted that the use of practical skills had a
potential to alter pupils’ perceptions of each other
and that the presence of teachers during the breaks
on EOtC days could benefit pupils’ interactions, as
teachers could assist pupils with the challenges in their
social interactions.17
Quantitative studies of the association between a
regular relocation of teaching to places outside the
buildings and walls of the schools and pupils’ psychoso-
cial well-being have been even fewer.1 A Swedish
1-year elementary school quasi-experimental trial
implemented EOtC at least 1 hour a day, and found
a significant positive association with psychosocial
well-being among boys, but not among girls.18
A school reform from 2014 in Denmark intended,
among other aims, to provide pupils with more varied
teaching and more physical movement in order to
improve the inclusion, well-being and motivation of
pupils with differing academic capacities.19 This reform
has contributed to increased attention to EOtC as a
means to enhance pupils’ psychosocial well-being20
despite the scarcity of research on this subject.
At time in which EOtC is becoming an increasingly
common teaching method in Denmark, our aim is to
investigate the association between the psychosocial
well-being of a group of 631 Danish pupils over a
school year and whether their teachers implement
EOtC regularly. We further investigate if this potential
association is modified by sex or socieconomic status
(SES).
METHODS
Participants
This study was part of the TEACHOUT project,21
which was designed as a quasi-experimental trial to
2 • Journal of School Health • 2019 • © 2019, American School Health Association
compare pupils over a school year in classes exposed
regularly to EOtC to pupils in comparison parallel
classes with little or no use of EOtC. Of the 1313
Danish public schools, we contacted 549 schools across
the country to recruit classes from grades 3-6. Most of
the schools were known to have previously practiced
EOtC. Eighteen schools agreed to participate in the
study and to have at least one school class practicing
EOtC during the following school year, and intended
to have at least one class from the same grade level
that would not practice EOtC that year.
From these 18 schools, teachers from a combined
total of 28 school classes agreed to provide at least
5 hours of EOtC per week with each of their classes,
divided over 1 to 2 sessions each week of the school
year. The teachers participated in a 2-day seminar
about the pedagogy and didactics of EOtC to promote
a comprehensive use. Teachers from a further 20
classes from the 18 schools agreed to participate in
the study while teaching as usual. Of the 48 classes, 34
(on 16 schools) provided both baseline and follow-
up Strengths and Difficulties Questionnaire (SDQ)
measures and were included in this study (24 EOtC
and 10 comparisons). None of the participating classes
took part in any other intervention projects during the
intervention period.
In the 34 participating school classes, 911 pupils had
consent to participate. Of the pupils with consent, 733
conducted baseline questionnaires at the beginning of
the school year, and of those, 86.1% (N = 631) also
conducted follow-up questionnaires at the end of the
school year.
Instruments
Exposure to EOtC. Data on the pupils’ exposure
to EOtC were collected throughout the school year
for all EOtC and comparison classes that initially were
eligible for participation. At face-to-face meetings prior
to the study, teachers were instructed and encouraged
to report daily on their implementation of EOtC using
an online instrument.21 The teachers had to report
the session’s length and location each time their class
received more than 45 minutes of continuous teaching
outside the school buildings (including transport,
evaluation, and breaks).
The exposure data were screened for eligibility.21
Reported EOtC sessions that had lasted less than
45 minutes, and/or were located inside the school
buildings, or at a location where the teaching normally
takes place (for instance, physical education (PE)
lessons in the school yard), were recoded as non-EOtC
sessions. To estimate the mean weekly duration and
number of EOtC sessions for each class, an intraclass
mean imputation was performed.22
Psychosocial well-being. At the beginning and at
the end of the school year, the pupils’ psychoso-
cial well-being was measured with the SDQ.23 The
Table 1. Cronbach’s Alpha Values for SDQ Scales

SDQ Scales Baseline/Follow-Up Cronbach’s Alpha Levels

Emotional Conduct Hyperactivity- Peer Problems Prosocial Behavior
Mean Age Problems Problems Inattention (ex. Item 11) (ex. Item 4)
Grade 3 (N= 147) Baseline 9.8years .614 (N= 132) .530 (N= 135) .653 (N= 130) .424 (N= 134) .626 (N= 137)
Follow-up 10.5years .641 (N= 136) .397 (N= 137) .796 (N= 141) .499 (N= 135) .675 (N= 141)
Grade 4 (N= 282) Baseline 10.7years .663 (N= 273) .519 (N= 264) .759 (N= 269) .570 (N= 273) .507 (N= 272)
Follow-up 11.4years .695 (N= 271) .451 (N= 272) .791 (N= 272) .623 (N= 271) .520 (N= 274)
Grade 5-6 (N= 202) Baseline 11.9years .578 (N= 180) .455 (N= 195) .732 (N= 190) .530 (N= 192) .576 (N= 195)
Follow-up 12.6years .596 (N= 194) .484 (N= 197) .731 (N= 194) .539 (N= 195) .618 (N= 196)
Mean α (SD) Baseline .618 (.043) .501 (.041) .715 (.055) .550 (.028)† .570 (.060)
Follow-up .644 (.050) .444 (.044) .773 (.036) .581 (.059)† .604 (.078)
Total .631 (.044) .473 (.049) .744 (.052) .566 (.042)† .587 (.065)
SDQ, Strengths and Difficulties Questionnaire.
Reported levels of Cronbach’s alpha values (α ) was calculated on cases eligible for analysis, ie, if both baseline and follow-up scales consisted of more than 3 items.
† Grade 3 not included, as the Cronbach’s alpha value in this scale was too low for this age group.

SDQ is a widely used, 25-item screening question-
naire covering positive and negative psychosocial
well-being aspects among children and adolescents
assigned to 5 scales: emotional problems, conduct
problems, hyperactivity-inattention, peer problems,
and prosocial behavior.24 Different versions of the SDQ
exist, including parental-, teacher-, and self-report ver-
sions. Participating pupils were asked to complete the
Danish version of the self-report SDQ,25 originally
validated for pupils aged 11-17.26 The class teachers
were instructed to administer the questionnaire during
lessons and to ensure the confidentiality of the data
collection. The questionnaires were accessed online
via unique personal codes.
Socioeconomic status. The pupils’ SES was esti-
mated using the Danish Occupational Social Class
(DOSC) measurement as the highest Social Group level
of their parents or legal guardians (I-VI).27 These data
were obtained in a separate electronic questionnaire
assigned to the parents via unique personal codes. The
social group levels were recoded into a binary (high or
low) SES variable. Levels I, II, and III were coded as
high SES, which include jobs that require a minimum
of one and a half years of practical skills and theoretical
training. Levels IV, V, and VI were coded to low SES,
which include jobs that require up to 1 year of theo-
retical and practical training, in addition to unskilled
workers and economically inactive.
Data Analysis
Internal consistency of SDQ scales. Cronbach’s
alpha analyses were performed for each of the 5 SDQ
scales to assess the consistency of the pupils’ answers
at baseline and follow-up, to improve the reliability
of the sample data. As SDQ measures psychological
constructs based on only a few item scales, and as our
research question included an explorative approach,
we accepted Cronbach’s alpha levels down to 0.50.28,29
Reliability analyses across scales and grades showed
reasonable levels of Cronbach’s alpha. Thus, an item
deletion approach was applied to increase alpha.
Table 1 shows the final levels of Cronbach’s alpha.30
Consequently, SDQ Item 11 was omitted from the peer
problems scale and Item 4 from the prosocial behavior
scale. Ultimately, due to a Cronbach’s alpha below
0.50, the conduct problems scale was not included in
any analysis, which meant that it was not possible
to calculate a total difficulties score. Furthermore, the
analyses of association of EOtC and peer problems did
not include pupils in Grade 3 as Cronbach’s alpha was
below 0.50 for this scale at this grade.
Z-score outlier analysis of the dependent variables
(baseline to follow-up difference in SDQ scale) was
performed after the analysis of internal consistency.
Cases that fell outside a 99.9% range of the normal
distribution (±3.29 SDs) were considered outliers and
deleted.
Only pupils who had answered at least 3 items of a
SDQ scale were included in the analysis on this scale.
Grouping. Other studies have shown that EOtC is
practiced approximately 3 hours/week when teachers
are committed to EOtC.31 An average weekly exposure
to EOtC of 2 hours was used as the cut point to divide
school classes into EOtC and comparison classes.
Modeling. First, the association between EOtC
and improvement in psychosocial well-being was ana-
lyzed as the difference between EOtC and comparison
group on baseline to follow-up development in each
SDQ scale. This was done using a multiple linear mixed
model with the variables EOtC vs comparison group,
mean number of weekly EOtC sessions, age, sex, and base-
line as fixed effects and with classes nested in schools
as random effect. The variable mean number of weekly
EOtC sessions was included in the model because the
total hours of EOtC could have been distributed on
1 or several sessions per week, with high dispersion
having a negative effect important to control for.

Journal of School Health • 2019 • © 2019, American School Health Association • 3

 Second, to investigate whether the potential
associations of regularly exposure to EOtC on aspects of
psychosocial well-being were modified by sex and SES,
interaction terms between EOtC and these variables
were added as fixed effects in 2 additional separate
models (one for each).
Linear model assumptions of homoskedasticity
and normality of residuals were evaluated by visual
inspection.
The between group effect size g were calculated of
all categorical variables based on estimated marginal
means.32 The between group effect sizes d were cal-
culated of continuous variables based on t-statistics.33
All statistical analyses were conducted in Linear
Mixed Models in SPSS version 24. Significance test
were 2-tailed, and levels of p < .050 were considered
as statistically significant.
RESULTS
Exposure to EOtC
Although the teachers of 24 of the original 34
participating classes had agreed to provide a minimum
of 5 hours of EOtC each week, did analysis of the
EOtC exposure data revealed that only 13 of the
classes (54.2%) consistently reached that level during
the year. However, all of the 24 classes were
above the 2 hours average weekly exposure to EOtC
cut point.
In addition, some of the teachers that had not
intended to practice EOtC during the year reported
significant amounts of EOtC for their classes. Three
of the comparison classes (30.0%) were above the
2 hours of average weekly exposure to EOtC cut-point.
These 3 classes were reassigned to the EOtC group.
Only 2 of the comparison classes (20.0%) consistently
reached a null implementation level.
In consequence, the EOtC group consisted of 27
classes (511 pupils) exposed to 2.4-6.7 hours of EOtC
in average per week (mean = 4.8 ± 0.2 hours), and the
comparison group consisted of 7 classes (120 pupils)
exposed to 0.0-1.9 hours of EOtC in average per week
(mean = 0.6 ± 0.3 hours) (Table 2). The difference in
exposure to EOtC between the EOtC and comparison
group was on average 4.1 hours (±0.5 hours) per week
(p < .000).
Characteristics of the EOtC and Comparison Groups
The EOtC and comparison groups did not differ
on sociodemographic descriptors, nor in their class
teacher’s teaching experience, but girls in the
EOtC group had a significant higher baseline level
of emotional problems (Table 2).
Association Between EOtC and Psychosocial Well-Being
Regular exposure to EOtC was significantly
(p = .045) associated, with an increase in prosocial
4 • Journal of School Health • 2019 • © 2019, American School Health Association
behavior (g = 0.25) in models adjusting for the effects
of sex, age, baseline SDQ, and the mean number of weekly
EOtC sessions. No significant associations were found
between regular exposure to EOtC and alteration of
emotional problems, hyperactivity-inattention, or peer
problems in this model (Table 3).
Influence of Sex and Socioeconomic Status
Sex did not significantly modify the associations
between regular exposure to EOtC and any of the SDQ
scales (Table 3).
The associations between regular exposure to EOtC
and reduction in hyperactivity-inattention (p = .035)
and peer problems (p = .045) were significantly
modified by whether pupils were of high or low
SES (Table 3). SES did not significantly modify the
association between regular exposure to EOtC and
emotional symptoms and prosocial behavior (Table 3).
Subsequently, post hoc analysis of the specific alter-
ation of hyperactivity-inattention and peer problems
was performed for each SES level (Table 4). The
reduction of hyperactivity-inattention was greater for
pupils of low SES (g = 0.62) than pupils of high SES
(g = 0.30). Only the reduction for pupils of low SES
was close to being significant (p = .054). Similarly for
peer problems, the reduction was greater for pupils of
low SES (g = 0.30) than pupils of high SES (g = −0.08);
however, the alterations were not significant for any
of the groups (Table 4).
The Influence of the Number of EOtC Sessions
Adjusting for the mean number of weekly EOtC sessions
strengthened the association between EOtC and the
observed improvements in the well-being measures.
In the final models, the mean number of weekly EOtC
sessions was significantly associated with increase in
hyperactivity-inattention (B = 0.571, p = .017), and
significantly associated with a decrease in prosocial
behavior (B = −0.354, p = .008) (Table 3).
Table 5 shows a post hoc analysis of how the
mean number of weekly EOtC sessions was associated to
improvements in hyperactivity-inattention and proso-
cial behavior when adjusting for the mean weekly hours
of EOtC (and vice versa). The mean number of weekly
EOtC sessions was significantly negatively associated
with both improvement of hyperactivity-inattention
and improvement in prosocial behavior. Conversely,
the mean weekly hours of EOtC was significantly
positively associated with a reduction of hyperactivity-
inattention when adjusting for number of sessions,
while the association with an increase in prosocial
behavior was positive, but not significant (Table 5).
DISCUSSION
This study’s finding that exposure to 2-7 hours
of EOtC per week significantly increased prosocial
Table 2. Descriptive Statistics of Participants and School Classes

N EOtC Group Comparison p

Pupil background (N= 631)
||
Sex 511/120 53.4% girls 56.7% girls
¶
Age 511/120 10.9 (.9) years 10.8 (.6) years
¶
Affiliation with the current class 3.4 (1.5) years 3.5 (1.3) years
||
Behavioral and emotional disorders 348/60 2.0% 1.7%
§
SES 340/58
Social Group I 22.1% 24.1%
Social Group II 35.9% 32.8%
Social Group III 16.5% 15.5%
Social Group IV 20.9% 25.9%
Social Group V 2.9% 1.7%
Social Group VI 1.8% .0%
SDQ baseline scores (N= 631)
Emotional problems (girls/boys) 503/117 3.0 (2.1)/2.5 (2.1) 2.6 (2.0)/1.7 (1.8) ∗∗ ¶/¶

Conduct problems (girls/boys) 501/117 1.2 (1.3)/1.8 (1.6) 1.2 (1.6)/1.6 (1.5) ¶/¶
¶¶
Hyperactivity-inattention (girls/boys) 506/118 2.8 (2.1)/3.6 (2.3) 2.8 (2.3)/3.1 (2.2) /
Peer problems † (girls/boys) 369/105 1.5 (1.5)/1.9 (1.6) 1.4 (1.7)/1.7 (1.6) ¶¶
/
Prosocial behavior ‡ (girls/boys) 503/118 6.8 (1.2)/6.4 (1.3) 6.7 (1.0)/6.2 (1.5) ¶¶
/
Principle teacher background (N= 34)
¶
Teaching experience 21/6 12.4 (7.6) years 14.0 (11.2) years
¶
EOtC teaching experience 19/5 3.0 (4.8) years 2.0 (4.5) years
Further education on EOtC 23/6 34.8% 16.7% §

Exposure to EOtC at class level (N= 34)

Mean hours of EOtC per week 27/7 4.8 (1.3) hours .6 (.8) hours ∗∗∗ ¶

Mean number of weekly EOtC sessions 27/7 1.3 (.5) sessions .3 (.4) sessions ∗∗∗ ¶

Mean weekly use of places in EOtC

∗∗∗ ¶
Natural environments and nature schools 27/7 1.8 (1.0) hours .2 (.2) hours
∗∗∗ ¶
Cultural institutions 27/7 1.1 (.6) hours .1 (.1) hours
∗∗ ¶
School ground 27/7 1.0 (.7) hours .2 (.4) hours
∗ ¶
Companies/factories 27/7 0.1 (.2) hours <.0 (<.0) hours
∗ p < .100.
∗∗ p < .050.
∗∗∗ p < .010.
EOtC, Education Outside the Classroom; SDQ, Strengths and Difficulties Questionnaire; SES, socioeconomic status.
N = 631. N, sample size EOtC group/comparison. Values are mean and SD in brackets, or percent.
† Item 11 and Grade 3 not included as the Cronbach’s alpha value on this scale was too low for this grade and to reach an acceptable Cronbach’s alpha for the scale on other
age-groups.
‡ SDQ Item 4 not included in analysis to reach an acceptable Cronbach’s alpha for the scale.
§ Gamma test.
|| Chi-square test.
¶ Independent samples t-test.

behavior for our sample of 631 Danish pupils adds
to the limited body of knowledge on the association
of EOtC to pupils’ psychosocial well-being. To our
knowledge, the only similar study of the association
between at least a yearlong regular exposure to EOtC
and aspects of pupils’ psychosocial well-being found no
main association but a positive association on overall
well-being for boys.18 Our study is therefore the first
to document an association between exposure to EOtC
and psychosocial well-being that is not sex dependent.
Our study is also the first to report whether the
association between regular exposure to EOtC and
psychosocial well-being is modified by SES. These
findings lend support the claim that EOtC is a means
to strengthen inclusion of pupils of low SES, ie,
pupils at risk of exclusion in school work and school
life,11,14,18 although further studies with larger samples
are needed in this area.
EOtC often involves peer collaboration and play,8
which are expected to foster pupils’ proximity,
familiarity, and shared interests, which are potentially
supportive for positive social relations.34 Also, EOtC
has been found to promote physical activity,35
which may affect reduction of hyperactivity and
inattention positively.36 Furthermore, in this study,
EOtC was predominantly implemented in natural
environments which have been shown to promote
children’s attention,37,38 as well as having a potential
of promoting impulse control.39
The results also highlight the importance of the
duration of EOtC sessions, with a higher number
of shorter EOtC sessions decreasing the positive

Journal of School Health • 2019 • © 2019, American School Health Association • 5

Table 3. Associations Between EOtC and SDQ Scales

Estimated Marginal Means

Fixed Effect EOtC Group Comparison
N F p Mean (95% CI) SD Mean (95% CI) SD g
Emotional problems
1
EOtC 503/117 (1, 28) = .57 .456 −.233 (−.404, −.061) 1.86 −.055 (−.471, .362) 2.21 .09
2
EOtC× Sex 503/117 (1, 60) = .85 .356
3
EOtC× SES high/low 335/56 (1, 376) = .14 .706
Hyperactivity-inattention
1
EOtC 506/118 (1, 30) = 2.26 .143 −.094 (−.358, .169) 2.90 .446 (−.186, 1.077) 3.37 .18
2
EOtC× Sex 506/118 (1, 602) = .10 .747
3
EOtC× SES high/low 337/57 (1, 376) = 4.49 .035
Peer problems†,‡,||
1 EOtC 369/105 (1, 22) = .70 .411 −.122 (−.313, .069) 1.77 −.327 (−.751, .097) 2.09 −.11
2 EOtC× Sex 369/105 (1, 456) < .00 .991
3 EOtC× SES high/low 234/50 (1, 271) = 4.25 .038
Prosocial behavior§
1 EOtC 503/118 (1, 28) = 4.39 .045 .063 (−.082, .207) 1.59 −.350 (−.697, −.002) 1.85 .25
2 EOtC× Sex 503/188 (1, 602) = 1.54 .216
3 EOtC× SES high/low 338/57 (1, 378) = .15 .697
EOtC, Education Outside the Classroom; SDQ, Strengths and Difficulties Questionnaire; SES, socioeconomic status.
N = 631. N, sample size EOtC group/comparison; CI, confidence interval.
1 Model 1. 2 Model 2. 3 Model 3. Positive effect size g indicates a positive effect in favor of the EOtC group (ie, a decrease in emotional problems, hyperactivity-inattention or
peer problems, or an increase in prosocial behavior. Models are adjusted for sex, age, baseline, and mean number of weekly EOtC sessions. Classes nested in schools are included
in the models as random effect.
† Grade 3 not included in analysis as the Cronbach’s alpha value on this scale was too low for this age group.
‡ SDQ Item 11 not included in analysis to reach an acceptable Cronbach’s alpha for the scale.
§ SDQ Item 4 not included in analysis to reach an acceptable Cronbach’s alpha for the scale.
|| Test for equality of variance exposed that the assumption of homoskedasticity in the peer problems moderation analysis, F (7, 276) = 2.11, p = .045, was violated. Thus,
F-statistics for EOtC × SES high/low moderation were computed under the condition of heteroskedasticity.

Table 4. Post Hoc Analysis of the Associations Between EOtC and SDQ Scales for High and Low SES Separately

Estimated Marginal Means

Fixed Effect EOtC Group Comparison
N F p Mean (95% CI) SD Mean (95% CI) SD g
Hyperactivity-inattention
High SES (I, II, III) 251/41 (1, 25) = 2.63 .118 −.204 (−.506, .099) 2.33 .514 (−.314, 1.341) 2.57 .30
Low SES (IV, V, VI) 86/16 (1, 25) = 4.09 .054 .165 (−.308, .639) 2.14 1.523 (.279, 2.768) 2.42 .62
Peer problems†,‡
High SES (I, II, III) 184/36 (1, 11) = .17 .691 −.216 (−.450, .018) 1.46 −.340 (−.938, .258) 1.64 −.08
Low SES (IV, V, VI) 50/14 (1, 59) = 2.15 .148 −.297 (−.795, .201) 1.76 .612 (−.446, 1.671) 1.98 .50
EOtC, Education Outside the Classroom; SDQ, Strengths and Difficulties Questionnaire; SES, socioeconomic status.
N = 631. N, sample size EOtC group/comparison; CI, confidence interval.
Positive effect size g indicates a positive effect in favor of the EOtC group, ie, a decrease in hyperactivity-inattention or peer problem. Models are adjusted for sex, age, baseline,
and mean number of weekly EOtC sessions. Classes nested in schools are included in the models as random effect.
† Grade 3 not included in analysis as the Cronbach’s alpha value on this scale was too low for this age group.
‡ SDQ Item 11 not included in analysis to reach an acceptable Cronbach’s alpha for the scale.

impacts of EOtC. This finding stresses a need to
be sensitive to the volume and frequency of EOtC
practiced, as too many short EOtC sessions appeared to
increase pupils’ hyperactivity-inattention and reduce
prosocial behavior in this study. For pupils with
attention deficit hyperactivity disorder, interruptions
in school scheduling are unfavorable.40 Also, pupils
propinquity to each other is a fundament for a
positive social climate.41 It may therefore be expected
that too many interruptions could have a negative
impact on hyperactivity-inattention reduction and
also reduce pupils’ possibility of being close together
since the start-up and end phases of each EOtC
session potentially eliminate time from group work
and breaks.
Limitations
This study’s main strength is that the analysis
and findings are based on a detailed assessment of the
actual exposure to EOtC, rather than the intended

6 • Journal of School Health • 2019 • © 2019, American School Health Association

Table 5. Analysis of the Influence of Average Weekly Duration and Dispersion of EOtC on Hyperactivity-Inattention and Prosocial
Behavior
Fixed Effect
N F
Reduction in hyperactivity-inattention
Mean hours of EOtC per week† 624 (1, 32) = 4.23
Mean number of weekly EOtC sessions‡ 624 (1, 33) = 8.29
Increase in prosocial behavior c
Mean hours of EOtC per week † 621 (1, 30) = 2.52
Mean number of weekly EOtC sessions‡ 621 (1, 30) = 5.85
EOtC, Education Outside the Classroom.
N = 631. N, sample size; CI, confidence interval; S.E., standard error.
Positive effect size d indicates a positive effect, ie, reduction in hyperactivity-inattention and increase in prosocial behavior. Models are adjusted for sex and age with classes
nested in schools as random effect.
† Model adjusted for mean number of weekly EOtC sessions.
‡ Model adjusted for mean hours of EOtC per week.
exposure that teachers agreed to at the study’s onset.
Randomized controlled trials are often considered
to be the most valid way to assess the effects
of an intervention. However, the current EOtC
practice in Denmark is predominately teacher-driven
and exposure to EOtC ultimately depends on the
individual teacher’s actual implementation. The only
previous large-scale controlled experimental study
on the association between regular exposure to EOtC
and pupil’s psychosocial well-being did not report
a precise estimate of the extent of implementation
delivery.18 Results based on exposure assessment are
an apparent need in the field of outdoor teaching
and learning specifically,42 and in school-based health
promotion in general.43,44
The original self-report version of SDQ was
developed for children and adolescents between 11
and 16 years of age.26 However, our study included
pupils from 9 to 13 years of age. Studies using the
self-reported version of SDQ on pupils just below the
intended age span seldom exceed levels of Cronbach’s
alpha above 0.70 and often just above 0.50.44-46 We
did not perform analysis on the conduct problems
scale due to levels of Cronbach’s alpha below 0.50.
Others have documented similar reliability issues in
the conduct problems scale with this age group.45,47
Initially, classes were predominantly sampled pair-
wise, ie, classes with regular exposure to EOtC and
parallel classes with rare exposure) to reduce bias from
confounding variables, such as parental background,
local area, and overall school resources, as the distri-
bution of pupils into classes within schools and school
years is random in the Danish public school system.
Only 3 of the classes that agreed to provide teaching
as usual were relocated to the EOtC group. How-
ever, the EOtC and comparison groups will still differ
from having different teachers. As school classes in
Denmark usually have the same teachers over sev-
eral years, some of these differences will already have
Journal of School Health
Parameter Estimates
p B (95% CI) S.E. t d
.048 −.180 (−.359, −.002) .088 −2.057 .12
.007 .809 (.237, 1.380) .281 2.897 −.16
.123 .080 (−.023, .182) .050 1.587 .09
.022 −.389 (−.718, −.060) .161 −2.418 −.14
had influence at baseline and are therefore accounted
for when adjusting for the baseline values of the
outcome being compared. However, some of the differ-
ences, ie, competence to teach using EOtC, may have
affected the pupils during the duration of the study
and may therefore have introduced a selection bias
on the results. The large degree of autonomy of Dan-
ish teachers48 and the variation in quality of teaching
relating to teachers’ enthusiasm and preparedness for a
lesson49 mean that EOtC was necessarily performed in
different ways and with varying quality by the teachers
in our study. Although participating teachers received
an introduction to EOtC, this study may be criticized
for not basing the analysis on any measures of the
quality of delivery and extent of specific EOtC char-
acteristics (such as places and teaching approaches) to
assess mechanisms of action.
Our sampling aimed to involve classes at schools
which were known to use EOtC previously. Far from
all the contacted schools participated with a class in our
study and not all pupils in each class provided baseline
and follow-up measures. The generalizability is most
reliable for pupils that are highly exposed to EOtC
carried out by teachers enthusiastic to use the method,
and at schools with an EOtC-supportive management.
The validity of the findings of this study should
be considered high in comparison to previous similar
studies, as the statistical analysis presented here is
based on a detailed assessment of EOtC exposure.
In addition, the identified associations between EOtC
exposure and aspects of psychosocial well-being were
significant despite the likely variation in EOtC practice,
and the lower reliability of the SDQ data, suggesting
that the results are robust.
Conclusion
In this study, exposure to 2-7 hours of EOtC per
week was shown to improve psychosocial well-being
of Danish public school pupils in terms of increased
• 2019 • © 2019, American School Health Association • 7
prosocial behavior. These findings support previ-
ous qualitative and longitudinal experiment studies
in showing that regular exposure to EOtC pro-
motes pupils’ social well-being. Interestingly, analysis
of detailed data on the duration and frequency of expo-
sure of each class to EOtC revealed that when the hours
of EOtC practiced are spread out over too many ses-
sions (thus resulting in many interruptions), it can
have a negative impact on hyperactivity-inattention
and prosocial behavior. In addition to including such
exposure data, future studies should aim to also
include measures of the quality of EOtC implemented
to strengthen conclusions on this topic.
IMPLICATIONS FOR SCHOOL HEALTH
The results of this study provides an argument
for school managements and teachers to integrate
EOtC as a regular part of the academic teaching
in primary schools in order to promote psychosocial
well-being, an important aspect of pupil’s mental
health.1 In comparison to other universal school-
based health initiatives integrated into academic
curricular teaching,9,10 EOtC has the advantage that it
has minimal interference with delivery of the core
curriculum. Several initiatives should be taken to
promote pupils’ psychosocial well-being using EOtC.
• Schools should map favorable places for EOtC in
their local environment, especially natural settings,
which can help teachers organize pupil group work,
exploration, and problem solving across subjects.
• In order to counteract the negative impacts that
the unnecessary interruptions associated with many
brief EOtC sessions have on pupils’ well-being,
schools should allow flexible timetables to enable
multiple successive subject-lessons, which will
enable the use of fewer longer EOtC sessions.
Human Subjects Approval Statement
All teachers were informed of the aim of the study.
A combined TEACHOUT parental/legal guardian form
provided written consent for their child’s participation
in the full project, based on written and oral
information.21 All data were handled confidentially.
The study was permitted by the Danish Data
Authorities (ref. number: 2014-54-0638). By Danish
law, no ethical approval was required for the TEA-
CHOUT project as it was not considered as biomedical
research, confirmed by Regional Committee on Health
Research Ethics (Protocol nr.: H-4-2014-FSP).
REFERENCES
1. Becker C, Lauterbach G, Spengler S, Dettweiler U, Mess F.
Effects of regular classes in outdoor education settings: a
systematic review on students’ learning, social and health
dimensions. Int J Environ Res Public Health. 2017;14(5):485.
8 • Journal of School Health • 2019 • © 2019, American School Health Association
2. Bentsen P, Mygind E, Randrup TB. Towards an understanding of
udeskole: education outside the classroom in a Danish context.
Education 3-13. 2009;37(1):29-44.
3. Mygind E. A comparison of children’s statements about social
relations and teaching in the classroom and in the outdoor
environment. J Adv Educ Outdoor Learn. 2009;9(2):151-169.
4. Turner L, Eliason M, Sandoval A, Chaloupka FJ. Increasing
prevalence of US elementary school gardens, but disparities
reduce opportunities for disadvantaged students. J Sch Health.
2016;86(12):906-912.
5. Braund M, Reiss M. Towards a more authentic science
curriculum: the contribution of out-of-school learning. Int J Sci
Educ. 2006;28(12):1373-1388.
6. Beames S, Higgins P, Nicol R. Learning Outside the Classroom:
Theory and Guidelines for Practice. New York, NY, Oxon, UK:
Routledge; 2012.
7. Rickinson M, Dillon J, Teamey K, et al. A Review of Research
on Outdoor Learning. London, UK: National Foundation for
Educational Research, King’s College London; 2004.
8. Waite S, Bølling M, Bentsen P. Comparing apples and pears?:
a conceptual framework for understanding forms of outdoor
learning through comparison of English Forest Schools and
Danish udeskole. Environ Educ Res. 2016;22(6):868-892.
9. Tancred T, Fletcher A, Melendez-Torres G, Thomas J,
Campbell R, Bonell C. Integrating health education in academic
lessons: is this the future of health education in schools? J Sch
Health. 2017;87(11):807-810.
10. Kase C, Hoover S, Boyd G, et al. Educational outcomes
associated with school behavioral health interventions: a review
of the literature. J Sch Health. 2017;87(7):554-562.
11. Fägerstam E. High school teachers’ experience of the educa-
tional potential of outdoor teaching and learning. J Adv Educ
Outdoor Learn. 2014;14(1):56-81.
12. Fox P, Avramidis E. An evaluation of an outdoor education
programme for students with emotional and behavioural
difficulties. Emot Behav Difficulties. 2003;8(4):267-283.
13. Szczepanski A, Dahlgren LO. Lärares uppfattningar av lärande
och undervisning utomhus (Teachers’ perspectives of learning
and teaching outdoors). Didakt Tidskr. 2011;20(1):21-48.
14. Norðdahl K, Jóhannesson IÁ. ‘Let’s go outside’: Icelandic teach-
ers’ views of using the outdoors. Education 3-13. 2014;44(4):
391-406.
15. Barfod K, Ejbye-Ernst N, Mygind L, Bentsen P. Increased
provision of udeskole in Danish schools: an updated national
population survey. Urban For Urban Green. 2016;20(Suppl
C):277-281.
16. Fägerstam E, Blom J. Learning biology and mathematics
outdoors: effects and attitudes in a Swedish high school context.
J Adv Educ Outdoor Learn. 2013;13(1):56-75.
17. Hartmeyer R, Mygind E. A retrospective study of social relations
in a Danish primary school class taught in ‘udeskole’. J Adv Educ
Outdoor Learn. 2015;16(1):78-89.
18. Gustafsson PE, Szczepanski A, Nelson N, Gustafsson PA. Effects
of an outdoor education intervention on the mental health of
schoolchildren. J Adv Educ Outdoor Learn. 2012;12(1):63-79.
19. The Danish Ministry of Education. Improving the Public
School - Overview of Reform of Standards in the Danish
Public School (Primary and Lower Secondary Education).
Copenhagen, Denmark: The Danish Ministry of Education;
2014 Available at: http://www.uvm.dk/-/media/filer/uvm/
publikationer/engelsksprogede/2014-improving-the-public-
schools.pdf. Accessed January 18, 2017.
20. Ejbye-Ernst N, Mygind L, Bentsen P. Inspirationsguide til God
Udeskolepraksis Målrettet Skoleledere, Lærere og Pædagoger
(Inspirational Guide to Good Udeskolepractice Targeted School
Principals, Teachers and Educators). Copenhagen, Denmark:
Ministry of Children, Education and Gender Equality, Ministry
of Environment and Food, Steno Diabetes Center, VIA
 University College; 2016. Available at: https://www.emu.dk/
sites/default/files/Inspirationsguide_0.pdf. Accessed January
18, 2017.
21. Nielsen G, Mygind E, Bølling M, et al. A quasi-experimental
cross-disciplinary evaluation of the impacts of education outside
the classroom on pupils’ physical activity, well-being and
learning: the TEACHOUT study protocol. BMC Public Health.
2016;16(1):1117.
22. Roth PL, Switzer FS, Switzer DM. Missing data in multiple item
scales: a Monte Carlo analysis of missing data techniques. Organ
Res Methods. 1999;2(3):211-232.
23. Goodman R. The strengths and difficulties questionnaire: a
research note. J Child Psychol Psychiatry. 1997;38(5):581-586.
24. Goodman R, Ford T, Simmons H, Gatward R, Meltzer H. Using
the Strengths and Difficulties Questionnaire (SDQ) to screen
for child psychiatric disorders in a community sample. Br
J Psychiatry. 2000;177(6):534-539.
25. Niclasen J, Skovgaard AM, Andersen A-MN, Sømhovd MJ,
Obel C. A confirmatory approach to examining the factor
structure of the Strengths and Difficulties Questionnaire (SDQ):
a large scale cohort study. J Abnorm Child Psychol. 2013;41(3):
355-365.
26. Goodman R, Meltzer H, Bailey V. The Strengths and Difficulties
Questionnaire: a pilot study on the validity of the self-report
version. Eur Child Adolesc Psychiatry. 1998;7(3):125-130.
27. Christensen U, Krølner R, Nilsson CJ, et al. Addressing social
inequality in aging by the Danish occupational social class
measurement. J Aging Health. 2014;26(1):106-127.
28. Nunnally JC, Bernstein IH. Psychometric Theory. 3rd ed. New
York, NY: McGraw-Hill Education; 1994.
29. Ponterotto JG, Ruckdeschel DE. An overview of coefficient
alpha and a reliability matrix for estimating adequacy of internal
consistency coefficients with psychological research measures.
Percept Mot Skills. 2007;105(3):997-1014.
30. Peterson RA. A meta-analysis of Cronbach’s coefficient alpha.
J Consum Res. 1994;21(2):381-391.
31. Bentsen P, Jensen FS. The nature of udeskole: outdoor learning
theory and practice in Danish schools. J Adv Educ Outdoor Learn.
2012;12(3):199-219.
32. Cohen J. Statistical Power Analysis for the Behavioural Sciences.
Hillsdale, NJ: Lawrence Erlbaum; 1988.
33. Morris SB, DeShon RP. Combining effect size estimates in
meta-analysis with repeated measures and independent-groups
designs. Psychol Methods. 2002;7(1):105.
34. Gifford-Smith ME, Brownell CA. Childhood peer relationships:
social acceptance, friendships, and peer networks. J Sch Psychol.
2003;41(4):235-284.
35. Mygind E. A comparison between children’s physical activity
levels at school and learning in an outdoor environment. J Adv
Educ Outdoor Learn. 2007;7(2):161-176.
Journal of School Health
36. Gapin JI, Labban JD, Etnier JL. The effects of physical activity on
attention deficit hyperactivity disorder symptoms: the evidence.
Prev Med. 2011;52(Suppl 1):70-74.
37. Mårtensson F, Boldemann C, Söderström M, Blennow M,
Englund J-E, Grahn P. Outdoor environmental assessment of
attention promoting settings for preschool children. Health Place.
2009;15(4):1149-1157.
38. Taylor AF, Kuo FE. Children with attention deficits concentrate
better after walk in the park. J Atten Disord. 2009;12(5):402-409.
39. Faber Taylor A, Kuo (Ming) FE. Could exposure to everyday
green spaces help treat ADHD? evidence from children’s play
settings. Appl Psychol Health Well-Being. 2011;3(3):281-303.
40. Ougrin D, Chatterton S, Banarsee R. Attention deficit
hyperactivity disorder (ADHD): review for primary care
clinicians. Lond J Prim Care. 2010;3(1):45-51.
41. Cairns R, Xie H, Leung M-C. The popularity of friendship and
the neglect of social networks: toward a new balance. New Dir
Child Adolesc Dev. 1998;1998(81):25-53.
42. Fiennes C, Oliver E, Dickson K, Escobar D, Romans A, Oliver S.
The Existing Evidence-Base about the Effectiveness of Outdoor
Learning. London, UK: Giving Evidence, UCL Institute of
Education, University College London, The Institute for Out-
door Learning; 2015. Available at: https://givingevidence.files
.wordpress.com/2015/03/outdoor-learning-giving-evidence-
revised-final-report-nov-2015-etc-v21.pdf. Accessed January
18, 2017.
43. Bruhn AL, Hirsch SE, Lloyd JW. Treatment integrity in school-
wide programs: a review of the literature (1993-2012). J Prim
Prev. 2015;36(5):335-349.
44. Gresham FM, Gansle KA, Noell GH, Cohen S, et al. Treatment
integrity of school-based behavioral intervention studies:
1980-1990. Sch Psychol Rev. 1993;22(2):254-272.
45. Koskelainen M, Sourander A, Kaljonen A. The Strengths and
Difficulties Questionnaire among Finnish school-aged children
and adolescents. Eur Child Adolesc Psychiatry. 2000;9(4):277-284.
46. Mellor D. Furthering the use of the Strengths and Difficulties
Questionnaire: reliability with younger child respondents.
Psychol Assess. 2004;16(4):396.
47. Muris P, Meesters C, Eijkelenboom A, Vincken M. The self-
report version of the Strengths and Difficulties Questionnaire:
its psychometric properties in 8- to 13-year-old non-clinical
children. Br J Clin Psychol. 2004;43(4):437-448.
48. Didaktik WI. Curriculum theory: are they the two sides of the
same coin? In: Hopmann S, Riquarts K, eds. Didaktik and/or
Curriculum. Kiel, Germany: Institute für die Pädagogik der
Naturwissenschaften (IPN); 1995:233-263.
49. Schulte AC, Easton JE, Parker J. Advances in treatment integrity
research: multidisciplinary perspectives on the conceptualiza-
tion, measurement, and enhancement of treatment integrity.
Sch Psychol Rev. 2009;38(4):460-475.
• 2019 • © 2019, American School Health Association • 9