Data were prepared and analysed in R Studio version 1.3.1056 (18) and R version 4.0.2. The analysis script and raw data are available here: https://osf.io/637rd/?view_only=e11a2e1accd843c59cecb3a54fc7767e .
Descriptives based on Time 1 data
Initially descriptive data are presented using Time 1 data for time spent playing across place and adventurous play rating across place. Figure 1 shows the mean hours spent playing in each place, as estimated based on mother and father report. Children were reported to spend more time playing inside at home or other people’s homes than in any other place. This was followed by playing outside at home and at other people’s homes. Beyond play at home, children spent most time playing at playgrounds and in green spaces. The least number of hours were spent playing at indoor play centres and near water. Mother’s and Father’s report was very similar although mothers estimated that children spent more time playing in green spaces and on the street relative to other places, than fathers. Full statistical analysis of differences across place are shown in supplementary file 2.
Figure 2 shows the mean adventure level for play in each place. This indicates that play was least adventurous at home and on the street near home. The adventurousness of play was slightly greater for play outdoors at home or other people’s homes and again, slightly more adventurous near water and on playgrounds. Play was most adventurous at indoor play centres and in green spaces, although Mothers and Fathers disagreed on which of these was the most adventurous place. Full statistical analysis of differences in adventurous play ratings across place are shown in supplementary material as Additional File 2.
Stability from time 1 to time 2
Table 1 shows the means and standard deviations for each variable at time 1 and time 2 for mothers and fathers. Paired sample t-tests indicated that there were no significant differences between mothers and fathers on any of the variables (all p’s > .05). Furthermore, paired t-tests showed that most of the scores from time 1 to time 2 were stable and did not differ significantly. The only exception was ratings of average level of adventurous play, which were higher at time 2.
Table 1. Mother and Father means and standard deviations for each metric at Time 1 and Time 2.
|
T1
|
T2
|
|
Mothers
|
Fathers
|
Mothers
|
Fathers
|
1. Total hours spent playing
|
1470.36 (864.57)
|
1296.30 (615.39)
|
1346.99 (720.52)
|
1273.45 (733.17)
|
2. Hours spent playing outdoors
|
799.27 (610.43)
|
631.96 (356.90)
|
717.58 (496.23)
|
659.82 (472.40)
|
3. Hours spent playing in nature
|
175.58 (215.45)
|
127.24 (171.87)
|
173.36 (214.88)
|
136.66 (196.76)
|
4. Hours spent playing in adventurous places
|
364.39
(359.05)
|
304.02
(253.43)
|
345.19
(345.01)
|
306.27
(313.95)
|
5. Hours spent playing adventurously
|
1176.9 (916.33)
|
1101.62 (676.51)
|
1227.11 (758.9)
|
1173.15 (783.05)
|
6. Average level of adventurous play
|
2.53
(0.76)
|
2.62
(0.72)
|
2.74**
(0.61)
|
2.72*
(0.56)
|
Note. Time 1 to Time 2 change is significant at * p < .05; ** p < .001.
Reliability
Concordance Correlation Coefficients (CCCs) were computed using the CCC function in epiR to assess the agreement between caregivers of the same child on each of the metrics of interest at time 1 and also to assess test-retest reliability for mothers and fathers separately across time 1 and time 2. CCCs were used over Intraclass Correlation Coefficients because many of the variables were not normally distributed.
Cross-informant agreement
As shown in Table 2, for the majority of metrics, there was poor to moderate agreement between caregivers based on time 1 data.
Table 2. Concordance Correlation Coefficients (CCC) for cross-informant and test-retest reliability for each metric.
|
Cross-informant agreement CCC
[upper and lower bounds]
|
Test retest reliability CCC Mothers
[upper and lower bounds]
|
Test retest reliability CCC Fathers
[upper and lower bounds]
|
1. Total hours spent playing
|
0.51
[0.34 – 0.65]
|
0.73
[0.62 – 0.80]
|
0.47
[0.26 – 0.64]
|
2. Hours spent playing outdoors
|
0.49
[0.32 – 0.63]
|
0.68
[0.56 – 0.77]
|
0.49
[0.29 – 0.65]
|
3. Hours spent playing in nature
|
0.36
[0.17 – 0.52]
|
0.73
[0.63 – 0.80]
|
0.61
[0.45 – 0.74]
|
4. Hours spent playing in adventurous places
|
0.44
[0.26 – 0.59]
|
0.76
[0.67 – 0.83]
|
0.63
[0.47 – 0.76]
|
5. Hours spent playing adventurously
|
0.37
[0.18 – 0.54]
|
0.67
[0.55 – 0.76]
|
0.39
[0.16 – 0.57]
|
6. Average level of adventurous play
|
0.41
[0.22 – 0.56]
|
0.49
[0.35 – 0.61]
|
0.42
[0.22 – 0.58]
|
Test-rest reliability
For the majority of metrics, the reliability fell in the moderate range, although there are notable exceptions. Father report of total hours spent playing, hours spent playing outdoors and hours spent playing adventurously did not reach the threshold for moderate reliability and were substantially lower than the reliability of the same metrics for mothers. Further, the reliability was notably lower for both parents for the average level of adventurous play variable, relative to the other metrics. This may be due to the timing of the survey in relation to the Covid-19 pandemic and adjustments in risk perception, which is consistent with the Time 1 to Time 2 change, shown in Table 1.
Internal consistency
For each of the time variables, the estimated time spent playing in each place is summed to create the total. Given this, we would not expect there to be strong internal consistency across items; the more time a child spends playing in one place decreases the available time for them to play in another place. Consistent with this, the Cronbach’s alpha, using time 1 data, ranged from 0.32 to 0.66 for these variables. In contrast, the mean level of adventurous play rating is calculated based on parent’s ratings of how adventurously their child plays across places. Whilst we would expect higher levels of adventurous play in some places relative to others, we might also expect some consistency within each child. This is reflected in the internal consistency for this variable which, based on time 1 data, was good (Mothers α= 0.89; Fathers α = 0.89).
‘Other’ responses
15% of participants reported that there was another place that their child played in an adventurous way. Only 17 participants (<4%) stated that their child may play in other places but they do so unsupervised. Table 3 shows the frequency with which other places were identified as places where children might play adventurously.
Table 3. Categories of places listed under ‘other places your child plays adventurously’ and number (proportion) of participants listing each place.
Place
|
Number (percentage) of parents listing ‘other’ place
|
School
|
10 (4.1%)
|
Theme parks/Adventure parks
|
9 (3.7%)
|
Cycling/Mountain biking
|
7 (2.9%)
|
Climbing Walls
|
5 (2.0%)
|
Farms/allotments
|
4 (1.6%)
|
Go Ape rope courses
|
4 (1.6%)
|
Beavers and Cub Scouts
|
3 (1.2%)
|
Others that were only mentioned once were: ski slope, skate park, cricket club, camping, forest school, gymnastics, DIY. Thirteen participants listed places that had been included in the CPS (at home, swimming, woods, other people’s gardens etc.).