OBJECTIVE:

To investigate whether increasing risk and challenge in primary school playgrounds influences interactions between children.

METHODS:

In a 2-year cluster-randomized controlled trial, 8 control schools were asked to not change their play environment, whereas 8 intervention schools increased opportunities for risk and challenge (eg, rough-and-tumble play), reduced rules, and added loose parts (eg, tires). Children (n = 840), parents (n = 635), and teachers (n = 90) completed bullying questionnaires at baseline, 1 (postintervention), and 2 (follow-up) years.

RESULTS:

Intervention children reported higher odds of being happy at school (at 2 years, odds ratio [OR]: 1.64; 95% confidence interval [CI]: 1.20–2.25) and playing with more children (at 1 year, OR: 1.66; 95% CI: 1.29–2.15) than control children. Although intervention children indicated they were pushed/shoved more (OR: 1.33; 95% CI: 1.03–1.71), they were less likely to tell a teacher (OR: 0.69; 95% CI: 0.52–0.92) at 2 years. No significant group differences were observed in parents reporting whether children had “ever” been bullied at school (1 year: P = .23; 2 years: P = .07). Intervention school teachers noticed more bullying in break time at 1 year (difference in scores: 0.20; 95% CI: 0.06–0.34; P = .009), with no corresponding increase in children reporting bullying to teachers (both time points, P ≥ .26).

CONCLUSIONS:

Few negative outcomes were reported by children or parents, except for greater pushing/shoving in intervention schools. Whether this indicates increased resilience as indicated by lower reporting of bullying to teachers may be an unanticipated benefit.

What’s Known on This Subject:

Qualitative research suggests that keeping children busy and active during playtime can reduce bullying. However, whether increasing risk and challenge in play, which is critical for child development, affects how children interact with one another is unknown.

Children from intervention schools with more opportunities for risk and challenge were happier at school and played with more children. Although they reported being pushed and shoved more, this translates into lower reporting of bullying to adults.

Play positively affects children’s cognitive, social, and emotional development,1,3 encouraging curiosity, imagination, gross motor activity, decision-making, and creative thinking, particularly playing outdoors in an unrestricted environment.4,9 Healthy play should also involve some risk, such as trying something new, being close to losing control, and overcoming fear.10,11 From a young age, children seek risk when they play and use their judgment to assess risk and potential injury.12,13 Children want risk; playgrounds are said to be “boring” if no risk is involved.14

However, little research has examined the pros and cons of greater risk and challenge in play.8,15,18 Qualitative studies show that keeping children busy and active during playtime can increase collaborative play and reduce bullying,19,21 but how risk and challenge itself influence bullying is understudied.21 Examining this issue seems timely given that bullying is common22,23 and negatively affects children’s health.24,25 Making playgrounds more interesting by meeting the need for greater risk and challenge could conceivably make children less likely to seek out other challenges in the form of unwanted behaviors.26 Although bullying has been defined as intentional, repeated, and negative behavior directed at a person who has difficulty defending him- or herself either because of an actual or perceived power imbalance,27 varying definitions exist, with some preferring “aggression.”28 There is also concern that the term “bullying” is overused and often includes behaviors that do not constitute bullying.29

We recently undertook a 2-year cluster-randomized controlled trial to determine whether increasing opportunities for risk and challenge within the school playground influenced physical activity and body weight in children.30 This study was powered on the basis of physical activity but provided a unique opportunity to explore whether the intervention affected children’s interactions with one another, particularly negative interactions like bullying.

The PLAY study was undertaken in 16 New Zealand primary schools (years 1–8 inclusive) that were randomly assigned to the intervention (given NZ$15 000 to assist with altering school play spaces) or the control (asked to not change anything in their school play spaces and given NZ$5000 toward playground redevelopment at study end) conditions for 2 years (1-year intervention, 1-year follow-up). Ethical approval was obtained from the University of Otago (reference 10/077) and the Auckland University of Technology (reference 10/95) ethics committees.

Main methods are outlined elsewhere.30 In brief, low- to middle-decile schools (greater proportion of students from low socioeconomic areas) with at least 150 pupils in the Otago and Auckland regions of New Zealand were approached (n = 21) and recruitment stopped once 16 schools provided signed consent. The research was discussed with all eligible children in class and signed consent was obtained from parents. No incentives were provided for children or parents. Pairs of schools were created by matching for region, school roll, and decile ranking and were randomly assigned to the intervention or control conditions by tossing a coin. Baseline data were collected from March to December 2011.

At baseline and 1 year, the play environments of all 16 schools were evaluated for 7 play opportunities: (1) risk and challenge, (2) engagement with natural elements, (3) ability to actively manipulate and change the play environment (eg, loose parts), (4) wheeled play (eg, bicycles, skateboards), (5) ball games, (6) opportunity for children to socialize, and (7) quality of independent access. Each item was scored from 1 (very poor) to 5 (excellent), and an overall score was determined for the whole school play environment (maximum of 35). After baseline evaluations, intervention schools were provided with tailored suggestions for improvements including the following: reducing the number of school rules (allowing tree climbing or rough-and-tumble play), allowing children to play outside when raining, using bicycles and skateboards at school (currently not typically allowed), and encouraging free play (use of loose parts such as tires and tree stumps to encourage imaginative play). Figure 1 shows one of the larger environmental changes (building mounds in a previously flat space). Because only low- to middle-socioeconomic schools were recruited, most recommendations involved no to little additional cost (eg, letting the grass grow long to encourage imaginative play, repurposing real estate signs for sledding down hills, plastic piping for water play). All children in intervention schools were exposed to the intervention; however, for pragmatic reasons, only those in years 2 (aged 6–7) and 4 (aged 8–9) were asked to participate in outcome assessments.

FIGURE 1

Photos showing an example of a play environment change (adding mounds) within 1 school and how this developed over the 2-year PLAY study. A, Baseline (2011); B, 1 year (May 2012); C, 1 year (November 2012); D, 2 years (2013).

FIGURE 1

Photos showing an example of a play environment change (adding mounds) within 1 school and how this developed over the 2-year PLAY study. A, Baseline (2011); B, 1 year (May 2012); C, 1 year (November 2012); D, 2 years (2013).

A peer-relations assessment package31 was used to assess bullying at all time points with separate questionnaires for children, parents, and teachers. The Peer Relations Assessment Questionnaires–Revised (PRAQ-R) survey multiple stakeholders in the school community, so that repeated questionnaires can assess intervention effectiveness. Because of copyright issues, permission was not granted to reproduce these questionnaires in full; only those questions relevant to our specific groups were included in analyses.

The PRAQ-R for Junior Students was completed during school time, with the children in small groups and separated so they could not see each other’s answers. Ten questions asked the children (1) how they felt at school, (2) whether they liked playtime, (3) how they played during playtime, and (4) whether they had ever been a bully or a victim at school (Supplemental Table 5). Cartoons illustrated some concepts (eg, Fig 2 shows physical bullying). Other questions were presented with smiley faces ranging from happy to sad. The questionnaire did not mention “bullying,” and no discussion preceded the questionnaire other than to explain to children that they were not being examined and that their answers would remain private. The questions were read out loud to the children by the same researcher at each measurement, and children recorded their answers on the forms.

FIGURE 2

One question in the PRAQ-R for Junior Students showing that some questions use cartoons to illustrate concepts. This figure shows physical contact.

FIGURE 2

One question in the PRAQ-R for Junior Students showing that some questions use cartoons to illustrate concepts. This figure shows physical contact.

The PRAQ-R for Parents (3 questions) was completed at all 3 time points, included a definition of bullying (Supplemental Table 6), and assessed whether bullying had “ever” occurred. Parents were asked about (1) their child’s relationships with other children in their class, (2) whether their child had been bullied at school, and (3) how this had affected their child.

Only the teachers of children in years 2 and 4 at baseline (eg, 4–11 teachers per school) and all 16 school principals were asked to complete the PRAQ-R for Teachers each time. The questionnaire included a definition of bullying (Supplemental Table 7) and 8 questions asking about (1) different types of bullying, (2) where it occurred, (3) whether it was reported to them, and (4) whether the teachers believed the school was a safe place for students.

The data were analyzed according to modified intention-to-treat following the CONSORT (Consolidated Standards of Reporting Trials) statement extended for cluster-randomized trials.32,33 The sample size of 331 per group was based on detecting important differences in the main outcomes (BMI z score, minutes of moderate to vigorous physical activity) rather than for bullying, a secondary outcome of PLAY.30 Generalized estimating equations with the use of a logit link, with an exchangeable correlation matrix and robust SEs, were used to analyze the data, which were recoded to have binary outcomes before analysis (see Supplemental Tables 5 and 6). The results, which compared the 2 groups, are presented as population-averaged estimates of the odds ratios (ORs). The models accounted for clustering by school and included terms for sex, age, measurement period, and dummy variables for the pairs of schools formed as part of the randomization procedure. Exploratory models considered sex × group interactions to determine if the results differed for boys and girls; where these were significant, separate estimates for boys and girls were derived from the model. Because the number of teachers was relatively small, regression analysis with the use of robust SEs was used to analyze these questionnaires for each time period. The results are therefore presented as differences in scores (Supplemental Table 7). No adjustment was made for multiple testing.

A total of 840 children were measured at baseline, with 630 remaining at 2 years (75% retention). Only 4 children (0.5%) withdrew from the main study, the remainder being lost to follow-up because of moving out of the area, a level of usual “transience”34,35 (Fig 3). Intervention and control children were broadly comparable at baseline (Table 1).

FIGURE 3

Flow of participants in the PLAY study by group. aOtago (n = 11), Auckland (n = 10); bOtago (n = 8), Auckland (n = 8); cOtago (n = 4), Auckland (n = 4); dOtago (n = 4), Auckland (n = 4). eRandomly excluded from consents received because of a lack of accelerometers.

FIGURE 3

Flow of participants in the PLAY study by group. aOtago (n = 11), Auckland (n = 10); bOtago (n = 8), Auckland (n = 8); cOtago (n = 4), Auckland (n = 4); dOtago (n = 4), Auckland (n = 4). eRandomly excluded from consents received because of a lack of accelerometers.

TABLE 1

Characteristics of the Study Population at Baseline

ControlIntervention
nValuenValue
Total N 422  418
Sex, % girls 235 52.9 216 47.2
Age, y 422 7.9 (1.1) 418 8.0 (1.2)
Height, cm 420 128.2 (8.6) 418 129.0 (8.5)
Weight, kg 422 29.0 (7.4) 418 29.4 (7.5)
BMI 420 17.4 (2.7) 418 17.4 (2.8)
Waist circumference, cm 421 61.0 (8.2) 417 61.0 (8.7)
WHtR 419 0.48 (0.05) 417 0.47 (0.05)
BMI z scorea 420 0.73 (1.07) 418 0.69 (1.11)
Ethnicity, %
Māori 64 14.4 88 19.2
Pacific 50 11.3 54 11.8
Asian 40 9.0 33 7.2
NZEO 231 52.0 211 46.1
Unknownb 59 13.3 72 15.7
ControlIntervention
nValuenValue
Total N 422  418
Sex, % girls 235 52.9 216 47.2
Age, y 422 7.9 (1.1) 418 8.0 (1.2)
Height, cm 420 128.2 (8.6) 418 129.0 (8.5)
Weight, kg 422 29.0 (7.4) 418 29.4 (7.5)
BMI 420 17.4 (2.7) 418 17.4 (2.8)
Waist circumference, cm 421 61.0 (8.2) 417 61.0 (8.7)
WHtR 419 0.48 (0.05) 417 0.47 (0.05)
BMI z scorea 420 0.73 (1.07) 418 0.69 (1.11)
Ethnicity, %
Māori 64 14.4 88 19.2
Pacific 50 11.3 54 11.8
Asian 40 9.0 33 7.2
NZEO 231 52.0 211 46.1
Unknownb 59 13.3 72 15.7

Data are presented as mean (SD) unless otherwise indicated. NZEO, New Zealand European and Other; WHtR, waist-to-height ratio.

a

World Health Organization reference data.36

b

Ethnicity unknown because of 1 of the following: not stated, refused, questionnaire not returned, response unidentifiable, absent at baseline, and no ethnicity available at subsequent time points.

Five questions asked the children about “happiness” at school and during playtime and some inconsistencies were apparent (Table 2). For example, intervention children were more likely to play with a lot of children at 1 year (OR: 1.66; 95% confidence interval [CI]: 1.29–2.15) and reported significantly higher odds of being happy at school at 2 years (question 5; OR: 1.64; 95% CI: 1.2–2.25). However, they were also less likely to “like the children in their class” (OR: 0.53; 95% CI: 0.41–0.70) at 2 years. In terms of bullying, no significant differences were observed between groups in verbal (question 9a: shouting) or relational (question 9b: exclusion) bullying at any time point. Intervention children were more likely to say they had been pushed/shoved (question 9c: physical bullying) during playtime, but this finding was only statistically significant at 2 years (OR: 1.33; 95% CI: 1.03–1.71). This increased reporting of pushing/shoving was reflected by lower (not higher) reports of bullying to teachers at 2 years (OR: 0.69; 95% CI: 0.52–0.92). Further analysis of “liking playtime” (question 7) showed there was a significant interaction between group and sex (P = .002), with boys being 1.52 times (95% CI: 1.01–2.28 at 1 year) more likely and girls half as likely (0.46; 95% CI: 0.23–0.92 at 2 years) to like playtime. By contrast, there were no significant sex × group interactions for the 3 questions regarding type of bullying (verbal, relational, or physical).

TABLE 2

Results of the PRAQ-R for Junior Students at Each Time Point and ORs for Differences Between Intervention and Control Children at 1 and 2 Years for Those With Baseline Measures and at Least 1 Repeat Measure

ControlaInterventionaORb (95% CI)P
Itemn%n%
Q5. Happy at school
Baseline 296 80.4 286 76.7
1 year 259 73.0 259 72.4 1.03 (0.71–1.49) .88
2 years 199 65.3 230 72.3 1.64 (1.20–2.25) .002
Q6. Likes the children in their class
Baseline 206 55.7 215 58.0
1 year 158 44.3 158 44.3 0.83 (0.64–1.07) .15
2 years 133 43.3 101 32.0 0.53 (0.41–0.70) <.001
Q7. Likes playtime
Baseline 301 82.5 304 82.6
1 year 272 77.3 268 75.7 1.06 (0.75–1.50) .76
2 years 223 73.6 241 77.2 1.39 (0.85–2.28) .19
Q8. Plays with a lot of children
Baseline 108 29.4 111 29.8
1 year 88 24.8 123 34.3 1.66 (1.29–2.15) <.001
2 years 94 30.6 101 31.8 1.00 (0.59–1.70) .99
Q18. Likes going to school
Baseline 215 58.3 214 57.2
1 year 177 49.7 166 46.1 0.76 (0.56–1.03) .07
2 years 140 45.6 117 36.7 0.65 (0.41–1.03) .07
Q9a. Is shouted at during playtime
Baseline 217 59.5 225 61.1
1 year 233 66.2 229 64.9 0.83 (0.60–1.16) .29
2 years 183 60.2 189 60.2 0.97 (0.73–1.29) .82
Q9b. Is excluded during playtime
Baseline 75 20.7 62 16.8
1 year 47 13.5 55 15.5 1.11 (0.72–1.70) .64
2 years 29 9.6 34 10.8 1.14 (0.69–1.90) .61
Baseline 164 45.2 179 47.9
1 year 154 44.0 169 47.2 1.29 (0.96–1.73) .09
2 years 114 38.0 134 42.1 1.33 (1.03–1.71) .03
Q13. Tells a teacher
Baseline 208 56.5 229 61.2
1 year 205 58.1 219 60.8 1.07 (0.76–1.50) .70
2 years 200 65.6 184 57.7 0.69 (0.52–0.92) .012
Q17. Have been told off by a teacher
Baseline 68 18.4 72 19.4
1 year 76 21.3 96 26.8 1.26 (0.87–1.83) .22
2 years 78 25.4 99 31.2 1.18 (0.80–1.75) .40
ControlaInterventionaORb (95% CI)P
Itemn%n%
Q5. Happy at school
Baseline 296 80.4 286 76.7
1 year 259 73.0 259 72.4 1.03 (0.71–1.49) .88
2 years 199 65.3 230 72.3 1.64 (1.20–2.25) .002
Q6. Likes the children in their class
Baseline 206 55.7 215 58.0
1 year 158 44.3 158 44.3 0.83 (0.64–1.07) .15
2 years 133 43.3 101 32.0 0.53 (0.41–0.70) <.001
Q7. Likes playtime
Baseline 301 82.5 304 82.6
1 year 272 77.3 268 75.7 1.06 (0.75–1.50) .76
2 years 223 73.6 241 77.2 1.39 (0.85–2.28) .19
Q8. Plays with a lot of children
Baseline 108 29.4 111 29.8
1 year 88 24.8 123 34.3 1.66 (1.29–2.15) <.001
2 years 94 30.6 101 31.8 1.00 (0.59–1.70) .99
Q18. Likes going to school
Baseline 215 58.3 214 57.2
1 year 177 49.7 166 46.1 0.76 (0.56–1.03) .07
2 years 140 45.6 117 36.7 0.65 (0.41–1.03) .07
Q9a. Is shouted at during playtime
Baseline 217 59.5 225 61.1
1 year 233 66.2 229 64.9 0.83 (0.60–1.16) .29
2 years 183 60.2 189 60.2 0.97 (0.73–1.29) .82
Q9b. Is excluded during playtime
Baseline 75 20.7 62 16.8
1 year 47 13.5 55 15.5 1.11 (0.72–1.70) .64
2 years 29 9.6 34 10.8 1.14 (0.69–1.90) .61
Baseline 164 45.2 179 47.9
1 year 154 44.0 169 47.2 1.29 (0.96–1.73) .09
2 years 114 38.0 134 42.1 1.33 (1.03–1.71) .03
Q13. Tells a teacher
Baseline 208 56.5 229 61.2
1 year 205 58.1 219 60.8 1.07 (0.76–1.50) .70
2 years 200 65.6 184 57.7 0.69 (0.52–0.92) .012
Q17. Have been told off by a teacher
Baseline 68 18.4 72 19.4
1 year 76 21.3 96 26.8 1.26 (0.87–1.83) .22
2 years 78 25.4 99 31.2 1.18 (0.80–1.75) .40

Int-Con, intervention-control; Q, question.

a

Although 840 children were in the study at baseline, these numbers and percentages refer to those selecting the positive or option 1 of categorical coding (see Supplemental Table 5 for more information about individual items and grouping of items). The denominator varied depending on the item because it may not have been answered within the questionnaire.

b

Int-Con ORs (95% CIs) adjusted for baseline value, age, and sex and accounted for clustering by school.

Of the 840 children with baseline measures, 635 parent questionnaires were received at baseline, 542 at 1 year, and 517 at 2 years. Parents indicated that intervention children were significantly more likely to have happy relations with other students at school at 1 year (OR: 1.96; 95% CI: 1.18–3.26), although this pattern was reversed at follow-up (OR: 0.49; 95% CI: 0.26–1.00) (Table 3). There were no significant differences in the number of parents reporting whether their child had ever been bullied at school (P = .23 at 1 year, P = .07 at 2 years), nor whether they had been bothered by this bullying (both P ≥ .29).

TABLE 3

Results of the PRAQ-R for Parents at Each Time Point and ORs for Differences Between Intervention and Control Children at 1 and 2 Years for Those With Baseline Measures and at Least 1 Repeat Measure

ControlaInterventionaORb (95% CI)P
Itemn%n%
Q2. Child has happy relations with other students
Baseline 296 92.8 289 91.5
1 year 249 88.3 238 91.9 1.96 (1.18–3.26) .009
2 years 242 95.3 235 89.4 0.51 (0.26–1.00) .049
Q3. Child has been bullied at the school
Baseline 159 50.2 162 51.8
1 year 156 55.5 153 58.9 1.28 (0.86–1.90) .23
2 years 128 50.6 159 61.2 1.60 (0.97–2.65) .07
Q6. Child has been bothered/upset by bullying at school
Baseline 158 51.3 169 54.9
1 year 152 55.1 153 60.5 1.09 (0.81–1.47) .58
2 years 129 52.4 155 61.0 1.27 (0.82–1.97) .29
ControlaInterventionaORb (95% CI)P
Itemn%n%
Q2. Child has happy relations with other students
Baseline 296 92.8 289 91.5
1 year 249 88.3 238 91.9 1.96 (1.18–3.26) .009
2 years 242 95.3 235 89.4 0.51 (0.26–1.00) .049
Q3. Child has been bullied at the school
Baseline 159 50.2 162 51.8
1 year 156 55.5 153 58.9 1.28 (0.86–1.90) .23
2 years 128 50.6 159 61.2 1.60 (0.97–2.65) .07
Q6. Child has been bothered/upset by bullying at school
Baseline 158 51.3 169 54.9
1 year 152 55.1 153 60.5 1.09 (0.81–1.47) .58
2 years 129 52.4 155 61.0 1.27 (0.82–1.97) .29

Int-Con, intervention-control; Q, question.

a

These numbers and percentages refer to those selecting the positive or option 1 of categorical coding (see Supplemental Table 6 for more information about individual items and their coding).

b

Int-Con ORs (95% CIs) adjusted for age and sex and accounted for clustering by school. Baseline values were not included in the model because there was not enough difference for them to be included in the model (lack of convergence).

A total of 104 teachers completed the baseline questionnaire, with 90 (87%) completing at least 1 subsequent questionnaire (others had left the schools). As was observed with children and parents, few differences were apparent between the intervention and control schools (Table 4). No significant differences were observed regarding the frequency of name-calling or cruel teasing. However, teachers in the intervention schools reported a higher frequency of physical bullying (difference in scores: 0.23; 95% CI: 0.07–0.39) and a tendency for more deliberate exclusion (0.20; 95% CI: 0.00–0.40) at 2 years. Differences in these behaviors were only observed at break time, not in the classroom, and were only significant at 1 year (0.20; 95% CI: 0.06–0.034). However, there was no evidence of a corresponding increase in reported bullying to teachers from children (P ≥ .26 at both time points), nor did it affect how the teachers viewed the safety of their school environment for children (P ≥ .19 for both years).

TABLE 4

Results of the PRAQ-R for Teachers at Each Time Point and Adjusted Differences Between Intervention and Control Children at 1 and 2 Years for Teachers With Baseline Measures and at Least 1 Repeat Measure

ItemnMean (SD)nMean (SD)
Q1a. How often does physical bullying occur?
Baseline 47 1.70 (0.55) 43 1.81 (0.59)
1 year 44 1.66 (0.53) 43 1.74 (0.58) 0.04 (−0.11 to 0.19) .54
2 years 43 1.67 (0.47) 34 1.88 (0.54) 0.23 (0.07 to 0.39) .007
Q1b. How often does name-calling occur?
Baseline 47 2.32 (0.52) 43 2.49 (0.70)
1 year 44 2.23 (0.48) 43 2.47 (0.63) 0.11 (−0.04 to 0.25) .13
2 years 43 2.21 (0.48) 34 2.41 (0.50) 0.08 (−0.13 to 0.29) .43
Q1c. How often does deliberate exclusion occur?
Baseline 47 1.77 (0.52) 43 1.86 (0.68)
1 year 44 1.91 (0.52) 43 1.86 (0.68) −0.04 (−0.21 to 0.13) .66
2 years 43 1.58 (0.63) 34 1.94 (0.65) 0.20 (−0.00 to 0.40) .05
Q1d. How often does cruel teasing occur?
Baseline 47 1.66 (0.56) 43 1.77 (0.72)
1 year 44 1.59 (0.54) 43 1.79 (0.60) 0.11 (−0.08 to 0.30) .24
2 years 43 1.65 (0.48) 34 1.79 (0.59) 0.10 (−0.08 to 0.27) .26
Q3a. Have you personally noticed bullying occurring in the classroom?
Baseline 47 1.62 (0.53) 42 1.69 (0.47)
1 year 44 1.58 (0.50) 43 1.74 (0.49) 0.14 (−0.08 to 0.36) .19
2 years 43 1.65 (0.48) 34 1.71 (0.46) 0.04 (−0.14 to 0.21) .68
Q3b. Have you personally noticed bullying occurring at recess or lunch?
Baseline 47 1.96 (0.42) 43 2.02 (0.41)
1 year 44 1.84 (0.43) 43 2.05 (0.30) 0.20 (0.06 to 0.34) .009
2 years 43 1.88 (0.39) 34 1.97 (0.30) 0.07 (−0.12 to 0.26) .43
Q4. The school is a safe place
Baseline 47 1.32 (0.47) 43 1.37 (0.54)
1 year 43 1.28 (0.45) 43 1.40 (0.49) 0.05 (−0.11 to 0.21) .54
2 years 43 1.33 (0.47) 34 1.47 (0.51) 0.12 (−0.07 to 0.30) .19
Q11. How often do students tell you they have been bullied at school?
Baseline 47 1.89 (0.67) 43 1.95 (0.58)
1 year 43 1.88 (0.54) 43 2.05 (0.58) 0.12 (−0.10 to 0.35) .26
2 years 43 1.93 (0.40) 34 1.91 (0.51) 0.01 (−0.15 to 0.16) .93
ItemnMean (SD)nMean (SD)
Q1a. How often does physical bullying occur?
Baseline 47 1.70 (0.55) 43 1.81 (0.59)
1 year 44 1.66 (0.53) 43 1.74 (0.58) 0.04 (−0.11 to 0.19) .54
2 years 43 1.67 (0.47) 34 1.88 (0.54) 0.23 (0.07 to 0.39) .007
Q1b. How often does name-calling occur?
Baseline 47 2.32 (0.52) 43 2.49 (0.70)
1 year 44 2.23 (0.48) 43 2.47 (0.63) 0.11 (−0.04 to 0.25) .13
2 years 43 2.21 (0.48) 34 2.41 (0.50) 0.08 (−0.13 to 0.29) .43
Q1c. How often does deliberate exclusion occur?
Baseline 47 1.77 (0.52) 43 1.86 (0.68)
1 year 44 1.91 (0.52) 43 1.86 (0.68) −0.04 (−0.21 to 0.13) .66
2 years 43 1.58 (0.63) 34 1.94 (0.65) 0.20 (−0.00 to 0.40) .05
Q1d. How often does cruel teasing occur?
Baseline 47 1.66 (0.56) 43 1.77 (0.72)
1 year 44 1.59 (0.54) 43 1.79 (0.60) 0.11 (−0.08 to 0.30) .24
2 years 43 1.65 (0.48) 34 1.79 (0.59) 0.10 (−0.08 to 0.27) .26
Q3a. Have you personally noticed bullying occurring in the classroom?
Baseline 47 1.62 (0.53) 42 1.69 (0.47)
1 year 44 1.58 (0.50) 43 1.74 (0.49) 0.14 (−0.08 to 0.36) .19
2 years 43 1.65 (0.48) 34 1.71 (0.46) 0.04 (−0.14 to 0.21) .68
Q3b. Have you personally noticed bullying occurring at recess or lunch?
Baseline 47 1.96 (0.42) 43 2.02 (0.41)
1 year 44 1.84 (0.43) 43 2.05 (0.30) 0.20 (0.06 to 0.34) .009
2 years 43 1.88 (0.39) 34 1.97 (0.30) 0.07 (−0.12 to 0.26) .43
Q4. The school is a safe place
Baseline 47 1.32 (0.47) 43 1.37 (0.54)
1 year 43 1.28 (0.45) 43 1.40 (0.49) 0.05 (−0.11 to 0.21) .54
2 years 43 1.33 (0.47) 34 1.47 (0.51) 0.12 (−0.07 to 0.30) .19
Q11. How often do students tell you they have been bullied at school?
Baseline 47 1.89 (0.67) 43 1.95 (0.58)
1 year 43 1.88 (0.54) 43 2.05 (0.58) 0.12 (−0.10 to 0.35) .26
2 years 43 1.93 (0.40) 34 1.91 (0.51) 0.01 (−0.15 to 0.16) .93

Data are presented as means (SDs) of the scores for each question (see Supplemental Table 7 for more information about individual items and their coding). Int-Con, intervention-control; Q, question.

a

Data were analyzed by regression analysis with the use of robust SEs for each time period, accounting for clustering by school.

Our study showed that increasing risk and challenge in the school playground to encourage more imaginative and risky play had relatively little effect on how children interacted with one another. Generally speaking, intervention children appeared happier at school, although some inconsistency in their response to different questions was apparent. Although both children and teachers reported more pushing/shoving or physical contact at intervention schools, this situation did not seem to be an issue for children, teachers, or parents. Interestingly, intervention children were significantly less likely to inform the teacher that someone had tried to hurt them, and the small difference noted in physical bullying by teachers was only significant at 1 year, and not serious enough to affect their view of whether the school was a “safe place” for children. Nor was this additional physical contact “going home” to parents, with parents from both groups reporting similar levels of bullying. However, the child’s sex may be important; boys appeared to enjoy playtime more than girls, perhaps reflecting this increased physical contact.29

It is difficult to compare our data with the literature because of the scarcity of studies investigating how risk and challenge in play influences interactions between children. Teachers believe that playground interventions encourage more cooperative play and that busy, active children “stay out of trouble.”19,21 Introducing loose parts (materials not typically considered as play equipment, eg, wood offcuts, tires) has been thought to lead to fewer incidents of aggressive behavior.19 School principals have also attributed a reduction in playground bullying to introducing an activity coordinator into schools to encourage activity during breaks. Whether this reduction was because children were more active, or simply because an additional adult was present on the playground is uncertain,20 but earlier work indicates that adult supervision is important for reducing bullying.37,39 However, it is also time-consuming, costly, and not always sustainable.37 Our PLAY intervention aimed to provide more opportunities for self-motivated play and engagement between children to encourage collaborative play and to increase risk and challenge,30 which, in turn, appears to have increased the amount of physical contact. However, it could be argued that the greater degree of interaction between children and the increased pushing are indicators of “success” in that the suggestions were so well accepted, they were noticeable in the playground. This is not to underplay the seriousness of bullying, which can be pervasive. Our results indicate that 50% to 61% of parents reported their child had “ever” been bullied, figures similar to those observed previously.40 Although serious bullying is the single reason why students feel unsafe at school,22 it seems unlikely that our children were being seriously bullied given that they reported being happy at school.

Results from all 3 participating groups indicated that despite this additional pushing and shoving in intervention schools, children in these schools were no more likely (and indeed, less likely at 2 years) to tell a teacher they had been bullied than children in control schools. There are several potential reasons for this apparent discrepancy. It is possible that children accepted the pushing and shoving as a normal part of the more dynamic play environment. We speculate that this, in turn, is helping children learn how to resolve minor conflicts among themselves without seeking the help of teachers. This observation was raised by teachers as a possible mechanism through separate exit interviews undertaken with school staff.41 Alternatively, it is possible that the children felt they could not approach the teachers about the pushing and shoving because it was part of the altered play environment. This possibility seems unlikely, however, because related qualitative work with these young children showed that children in intervention schools were often unaware the intervention even existed, and certainly could not describe how the playground had changed over time.29 Although teachers in intervention schools reported more bullying during breaks than in control schools, exit interviews with staff involved directly with the intervention did not reflect these views. On the occasions bullying was mentioned, it was to emphasize that bullying and undesirable behaviors had declined since the intervention.41 Finally, the measurement tool may be an issue; teachers found the PRAQ-R to be difficult to answer because they simply did not see bullying as a big issue in their schools. Teachers often felt the need to explain their “middle of the road” answers and were concerned that the middle answers (eg, occasional bullying) may indicate a problem at the school where there was none; however, they did so because they were reluctant to answer that there was never a problem.

The main strengths of our study were the large sample size, 2-year duration, and the assessment of children’s interactions from multiple viewpoints, which enabled a more comprehensive view of the school environment. Intervention fidelity was achieved with intervention schools providing more opportunities than control schools for risk and challenge and wheeled play in particular, whereas control schools did not make any substantial changes to their play environment.30 The main limitation of our study is that bullying is difficult to assess.42 When we commenced PLAY, few options were available for quantitatively assessing children’s relationships, and we chose the short PRAQ-R questionnaires to reduce participant burden. However, its brevity meant that we could not gather information about the wider context or intent of bullying, so it may be insufficient for assessing such a complex behavior. There was also no time frame (eg, “last month/fortnight/week/term”) in any of the questionnaires, and convergence problems when analyzing the data meant that we dichotomized variables rather than treating them as multinomial, thus losing information and statistical power.43 The questions asked of each group also may appear somewhat disconnected, which might complicate comparison and interpretation. However, a strength of the questionnaire is that it assesses “bullying” from multiple perspectives; thus, different questions are required. Our observation that boys liked playtime more than girls was surprising given that thoughtfully designed play spaces deliver more choice and opportunity to expand physical skills for both boys and girls whether they possess low or high physical ability.44 However, caution should be applied when interpreting these findings, given that our study was not powered to detect differences between boys and girls as shown by the wide CIs. Finally, although we had a large sample of children and parents, we had a smaller sample of teachers, which led to wide CIs for many of the estimates. In retrospect, we should have asked all teachers at each school to complete the teacher questionnaire, because it referred to the wider school environment rather than to the children specifically being measured in PLAY.

Our study shows that increasing risk and challenge in the school playground produced positive effects, with children reporting being happier in intervention schools, a view that was supported by their parents. Although findings from teachers initially appeared to be more negative, these were attributed to the questionnaire and were in direct contrast to strong supportive views expressed by other school staff more directly involved in the intervention. Although intervention children also reported greater pushing and shoving at school, the teaching staff felt that this increased contact translated into greater resilience in the children, as indicated by significantly lower reporting of bullying to them. Such a positive effect suggests an interesting and unanticipated benefit of introducing activities that encourage children to explore greater degrees of risk and challenge in the school environment and warrants further examination.

• CI

confidence interval

•
• OR

odds ratio

•
• PRAQ-R

Peer Relations Assessment Questionnaires–Revised

Dr Farmer assisted with the study design, was project manager and PhD student on the PLAY study, assisted with statistical analyses, and wrote the first and subsequent drafts of the manuscript; Dr Williams was a co-investigator, assisted with the study design, was responsible for all statistical analyses, and commented on the manuscript; Dr Mann was a co-investigator, assisted with the study design, and commented on the manuscript; Dr Schofield was a co-investigator, assisted with the study design, and oversaw data collection in Auckland; Ms McPhee was the project coordinator in Auckland; Dr Taylor conceived the idea for the study, was the principal investigator of the project, was responsible for the overall study design and monitoring of data collection, and redrafted the manuscript with Dr Farmer; and all authors read and approved the final manuscript.

This trial has been registered with the Australia New Zealand Clinical Trial registry (identifier ACTRN12612000675820).

FUNDING: The PLAY study was funded by the Health Research Council of New Zealand, and the Otago Diabetes Research Trust. Dr Farmer was in receipt of a Medicine Award and subsequently a Lottery Health Research New Zealand PhD Scholarship during her PhD study. Dr Taylor is partially funded by a fellowship from the Karitane Products Society Limited. The funders had no role in the design of the study; the collection, analysis, and interpretation of the data; the writing of the manuscript; or the decision to submit the article for publication.

We recognize and thank the many children, families, schools, and teachers who participated in this research and the research assistants who helped with data collection.

1
Fisher
E
.
The impact of play on development: a meta-analysis.
Play Cult
.
1992
;
5
(
2
):
159
181
2
Burdette
HL
,
Whitaker
RC
.
Resurrecting free play in young children: looking beyond fitness and fatness to attention, affiliation, and affect.
.
2005
;
159
(
1
):
46
50
[PubMed]
3
Milteer
RM
,
Ginsburg
KR
;
Council on Communications and Media
;
Committee on Psychosocial Aspects of Child and Family Health
.
The importance of play in promoting healthy child development and maintaining strong parent-child bond: focus on children in poverty.
Pediatrics
.
2012
;
129
(
1
). Available at: www.pediatrics.org/cgi/content/full/129/1/e204
[PubMed]
4
Brown
S
.
Play: How It Shapes the Brain, Opens the Imagination, and Invigorates the Soul.
New York, NY
:
;
2010
5
Gill
T
.
No Fear: Growing Up in a Risk Averse Society
.
London, England
:
Calouste Gulbenkian Foundation
;
2007
6
Gray
P
.
Free to Learn: Why Unleashing the Instinct to Play Will Make Our Children Happier, More Self-Reliant and Better Students for Life
.
New York, NY
:
Basic Books
;
2013
7
Little
H
,
Wyver
S
.
Outdoor play: does avoiding the risks reduce the benefits?
Aust J Early Child
.
2008
;
33
(
2
):
33
40
8
Tremblay
MS
,
Gray
C
,
Babcock
S
, et al
.
Position statement on active outdoor play.
Int J Environ Res Public Health
.
2015
;
12
(
6
):
6475
6505
[PubMed]
9
Gleave
J
,
Cole-Hamilton
I
. A World Without Play: A Literature Review on the Effects of a Lack of Play on Children's Lives: Play England. 2012. Available at: http://www.playengland.org.uk/resource/a-world-without-play-literature-review/
10
Stephenson
A
.
Physical risk-taking: dangerous or endangered?
Early Years J Int Res Dev
.
2003
;
23
(
1
):
35
43
11
Sandseter
EB
,
Kennair
LE
.
Children’s risky play from an evolutionary perspective: the anti-phobic effects of thrilling experiences.
Evolut Psychol
2011
;
9
(
2
):
257
284
12
Christensen
P
,
Mikkelsen
MR
.
Jumping off and being careful: children’s strategies of risk management in everyday life.
Sociol Health Illn
.
2008
;
30
(
1
):
112
130
[PubMed]
13
Little
H
,
Wyver
S
.
Individual differences in children’s risk perception and appraisals in outdoor play environments.
Int J Early Years Educ
.
2010
;
18
(
4
):
297
313
14
Hyndman
BP
,
Telford
A
.
Should educators be ‘wrapping school playgrounds in cotton wool’ to encourage physical activity? Exploring primary and secondary students’ voices from the school playground.
Aust J Teach Educ.
2015
;
40
(
6
):
60
84
15
Brussoni
M
,
Gibbons
R
,
Gray
C
, et al
.
What is the relationship between risky outdoor play and health in children? A systematic review.
Int J Environ Res Public Health
.
2015
;
12
(
6
):
6423
6454
[PubMed]
16
Bundy
AC
,
Naughton
G
,
Tranter
P
, et al
.
The Sydney playground project: popping the bubblewrap—unleashing the power of play: a cluster randomized controlled trial of a primary school playground-based intervention aiming to increase children’s physical activity and social skills.
BMC Public Health
.
2011
;
11
:
680
[PubMed]
17
Engelen
L
,
Bundy
AC
,
Naughton
G
, et al
.
Increasing physical activity in young primary school children—it’s child’s play: a cluster randomised controlled trial.
Prev Med
.
2013
;
56
(
5
):
319
325
[PubMed]
18
Niehues
A
,
Bundy
A
,
Broom
A
,
Tranter
P
,
Ragen
J
,
Engelen
L
.
Everyday uncertainties: reframing perceptions of risk in outdoor free play.
.
2013
;
13
(
3
):
223
237
19
Bundy
AC
,
Luckett
T
,
Tranter
PJ
, et al
.
The risk is that there is “no risk”: a simple, innovative intervention to increase children’s activity levels.
Int J Early Years Educ
.
2009
;
17
(
1
):
33
45
20
Taylor
RW
,
McAuley
KA
,
Barbezat
W
,
Strong
A
,
Williams
SM
,
Mann
JI
.
Apple Project: 2-y findings of a community-based obesity prevention program in primary school age children.
Am J Clin Nutr
.
2007
;
86
(
3
):
735
742
[PubMed]
21
Parrish
A-M
,
Yeatman
H
,
Iverson
D
,
Russell
K
.
Using interviews and peer pairs to better understand how school environments affect young children’s playground physical activity levels: a qualitative study.
Health Educ Res
.
2012
;
27
(
2
):
269
280
[PubMed]
22
Cowie
H
,
Oztug
O
.
Pupils’ perceptions of safety at school.
Pastoral Care Educ
.
2008
;
26
(
2
):
59
67
23
Majcherová
K
,
Hajduová
Z
,
Andrejkovič
M
.
The role of the school in handling the problem of bullying.
Aggress Violent Behav
.
2014
;
19
(
5
):
463
465
24
Ttofi
MM
,
Farrington
DP
.
Bullying: short-term and long-term effects, and the importance of defiance theory in explanation and prevention.
Vict Offenders
.
2008
;
3
(
2–3
):
289
312
25
Ttofi
MM
,
Farrington
DP
.
Effectiveness of school-based programs to reduce bullying: a systematic and meta-analytic review.
J Exp Criminol
.
2011
;
7
(
1
):
27
56
26
Walsh
P
.
Fixed equipment: a time for change.
Aust J Early Child
.
1993
;
18
(
2
):
23
29
27
Olweus
D
,
Limber
SP
.
Bullying in school: evaluation and dissemination of the Olweus Bullying Prevention Program.
Am J Orthopsychiatry
.
2010
;
80
(
1
):
124
134
[PubMed]
28
Griffin
RS
,
Gross
AM
.
Childhood bullying: current empirical findings and future directions for research.
Aggress Violent Behav
.
2004
;
9
(
4
):
379
400
29
George
M
,
Fitzgerald
R
,
Farmer
V
,
Taylor
R
.
Changing the Way We Play at School: Hearing the Children’s Voice
.
Dunedin, New Zealand
:
Department of Anthropology and Archaeology, University of Otago
;
2015
30
Farmer
VL
,
Williams
SM
,
Mann
JI
,
Schofield
G
,
McPhee
JC
,
Taylor
RW
.
The effect of increasing risk and challenge in the school playground on physical activity and weight in children: a cluster randomised controlled trial (PLAY) [published online ahead of print March 7, 2017].
Int J Obes (Lond)
.
31
Rigby
K
.
Peer Relations Assessment Questionnaires-Revised (PRAQ-R).
Victoria, Australia
:
ACER Press
;
2010
32
Campbell
MJ
.
Cluster randomized trials
. In:
Ahrens
W
,
Pigeot
I
, eds.
Handbook of Epidemiology.
, 2nd ed.
New York, NY
:
;
2014
:
389
417
33
Campbell
MK
,
Piaggio
G
,
Elbourne
DR
,
Altman
DG
;
CONSORT Group
.
CONSORT 2010 statement: extension to cluster randomised trials.
BMJ
.
2012
;
345
:
e5661
[PubMed]
34
Ministry of Education. Education counts. Indicator report: transient students [online report]. Ministry of Education; 2013. Available at: www.educationcounts.govt.nz. Accessed January 26, 2015
35
Child Poverty Action Group, Inc
. The Revolving Door: Student Mobility in Auckland Schools. Auckland, New Zealand: Poverty Action Group; 2014. Available at: www.cpag.org.nz/news/new-report-student-transience-in-auckland/. Accessed June 15, 2015
36
de Onis
M
,
Onyango
AW
,
Borghi
E
,
Siyam
A
,
Nishida
C
,
Siekmann
J
.
Development of a WHO growth reference for school-aged children and adolescents.
Bull World Health Organ
.
2007
;
85
(
9
):
660
667
[PubMed]
37
Swearer
SM
,
Doll
B
.
Bullying in schools: an ecological framework.
J Emotional Abuse
.
2001
;
2
(
2–3
):
7
23
38
Craig
WM
,
Pepler
D
,
Atlas
R
.
Observations of bullying in the playground and in the classroom.
Sch Psychol Int
.
2000
;
21
(
1
):
22
36
39
Olweus
D
.
Bullying at School: What We Know and What We Can Do
.
Oxford, United Kingdom
:
Blackwell Publishers
;
1993
40
Waasdorp
TE
,
CP
,
Duong
J
.
The link between parents’ perceptions of the school and their responses to school bullying: variation by child characteristics and the forms of victimization.
J Educ Psychol
.
2011
;
103
(
2
):
324
335
41
Farmer
VL
,
Fitzgerald
RP
,
Williams
SM
, et al
.
What did schools experience from participating in a randomised controlled study (PLAY) that prioritised risk and challenge in active play for children while at school? [published online ahead of print February 14, 2017]
.
.
2016
;
42
Evans
CBR
,
Fraser
MW
,
Cotter
KL
.
The effectiveness of school-based bullying prevention programs: a systematic review.
Aggress Violent Behav
.
2014
;
19
(
5
):
532
544
43
Altman
DG
,
Royston
P
.
The cost of dichotomising continuous variables.
BMJ
.
2006
;
332
(
7549
):
1080
[PubMed]
44
Barbour
AC
.
The impact of playground design on the play behaviors of children with differing levels of physical competence.
Early Child Res Q
.
1999
;
14
(
1
):
75
98

## Competing Interests

POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.

FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.