Method

Participants

Survey respondents were 427 parents/caregivers (93% female) with a child aged 5 to 11 years. Participants were self-selecting, recruited via social media and newsletters sent from a range of organizations. Their children (45% female) were aged 5 to 11 years (M = 8.02, SD = 1.98). All participants were living in Northern Ireland and the majority identified as White British or White Irish, consistent with the population of Northern Ireland, where > 98% are white. There was diversity in other demographic factors, see Table S1 in Supplementary Material for full demographic characteristics. To be included, participants had to correctly answer two questions that were designed to screen out bots and participants not paying attention (see Section 2 of Supplementary Material for details).

Measures

The full survey was designed to mirror the British Children’s Play Survey (BCPS; 1), which did not include participants in Northern Ireland. The focus in the present paper is on associations between children’s play and mental health. For completeness we analysed the data addressing the same research questions as the published BCPS paper and present these results in Supplementary Material Section 4.

Children’s Play Scale (CPS)

The CPS [25] includes questions about play in seven places. Full details are provided in Supplementary Materials Section3. Briefly, respondents report the frequency and length of time their child plays in each place in Spring/Summer and Autumn/Winter. These responses are then used to give an estimate of total time spent playing in each place within a year. The adventurous play supplement asked parents to rate how adventurously their child plays at each place. This was rated on a five-point Likert scale ranging from 1 (very low levels of adventure) to 5 (maximum levels of adventure). For this study we extracted three variables: total time spent playing outdoors, time spent playing adventurously and time spent playing unadventurously. To calculate time spent playing adventurously, we calculated the total time spent playing using only places where parents rated that their child played with at least a mild level of adventure (2 on the Likert scale). To calculate time spent playing unadventurously, we calculated the total time spent playing using only places where parents rated that their child played with very low level of adventure (1 on the Likert scale). We conducted sensitivity analyses with the cut off at 2 for unadventurous and 3 for adventurous and the pattern of results and conclusions remained the same. The test re-test and cross-informant reliability of the CPS and adventurous play supplement have been evaluated in Dodd, Nesbit [25].

SDQ

The SDQ is a 25-item screening questionnaire that asks about positive and negative attributes. Items can be combined to create five subscales: emotional symptoms, conduct problems, hyperactivity/inattention, peer relationship problems and prosocial behaviour. An internalising problems score can be calculated by summing emotional problems and peer relationship problems, and an externalising score can be calculated by summing conduct problems and hyperactivity/inattention. The SDQ is widely used and has strong psychometrics (see www.sdqinfo.org [26]).

PANAS

The 10-item Positive and Negative Affect Schedule for Children-P [27] asks parents to read each of 10 items which describe an emotion and select to what extent their child has felt this way during the past few weeks. Five items are positive emotions, five are negative. Responses are summed to give scores for positive affect and negative affect. The scale has good psychometric properties [27].

Kessler-6 (K6)

The Kessler-6 [28] is a 6-item scale, designed to capture distress, based on symptoms of anxiety and depression, in adults. The original scale asks respondents to consider the previous 4-week period and to rate, using a five-point Likert scale, how often they felt each of the six emotions listed. To gain a measure of longer-term parent mental health, which we use as a control variable, we asked respondents to describe how often they had felt each emotion over the past year. The K6 has good psychometric properties [29].

Procedure

Participants were recruited via social media and newsletters sent from a range of organizations. Interested parents were invited to take part in the study via advertisements that directed them to an information sheet and consent form. After consent the survey opened in SurveyMonkey. Recruitment began on 3rd April 2020, when the first responses to the survey were received. The survey was closed on 26th April 2020. The UK-wide lockdown due to Covid-19 began on March 26th meaning the survey was always completed within one month of the start of the lockdown. Respondents completed the survey on one occasion and were clearly instructed to answer the CPS and SDQ questions thinking about life before it was affected by the Covid-19 pandemic. For the PANAS they were asked to respond with the previous few weeks in mind. Participants were given a £10 voucher.

Missing Data, Distributions and Outlier Checks

There was some missing data, in particular on three of the time questions on the CPS. Missing data were imputed using the mice package in R [30] and pooled results are reported. Where variables were skewed they were transformed and outliers were Winsorized. See supplementary materials Sect. 2 for full details.

Data Analysis Plan

Given that analyses included three variables extracted from the CPS we examined correlations between these measures. Time spent playing adventurously was positively associated with time spent playing outside (r = 0.59, p < 0.001) and negatively associated with unadventurous play (r = − 0.63, p < 0.001). Time spent playing outside and unadventurous play were not significantly correlated (r = 0.07, p = 0.168).

To examine our hypotheses, Pearson r correlations, pooled from the multiply imputed datasets, were used to evaluate associations between the three play variables and four child mental health variables (SDQ internalising, SDQ externalising, PANAS positive affect, PANAS negative affect). A Bonferroni-corrected alpha of 0.0125 (0.05/4) was used to correct for multiple comparisons. We then examined whether these correlations remained significant after controlling for parent mental health and child sex, child age, child disability or mental health problem, child birth-order, parent employment status, parent marital status, parent age and parent level of education. Again, a Bonferroni corrected alpha level of 0.0125 was used.

Results

All correlations can be seen in Table 1. These results show that time spent playing adventurously was significantly associated with both children’s internalising scores on the SDQ and positive affect as measured using the PANAS, with children who spent more time playing adventurously having fewer internalising symptoms and more positive affect. No significant correlations were found between adventurous play and either externalising symptoms or negative affect. In contrast, hours spent playing unadventurously was not related to positive affect and was positively related to internalising scores on the SDQ. Children who spent more time playing unadventurously had more internalising symptoms. For hours spent playing outdoors, no significant correlations with internalising or externalising scales nor for PANAS positive or negative affect scales were found.

Table 1 Mean and standard deviation for each child mental health measure and Pooled Pearson correlation coefficients (r) and 95% confidence intervals for relationships between play variables and children's mental health Full size table

All three significant correlations were robust to controlling for demographic variables. The full results of these regressions can be seen in the results file here: https://beta.ukdataservice.ac.uk/datacatalogue/doi/?id=8793#!#0. Although it was not our primary focus, in addition to the play variables, the following demographic factors were significant predictors of child mental health. For both regressions predicting internalising problems (one with adventurous play as a predictor and one with unadventurous play as a predictor), the demographic variables that were significant predictors were child disability status, marital status and parent mental health (p < 0.05). Children had fewer internalising symptoms when they were not described as having a learning disability, mental health problem or physical disability, when their parents were married or living as married and when the reporting parent had lower levels of distress as reported on the K6. For the model predicting positive affect, the demographic variables that were significant predictors were child sex, child age, child disability status and parent mental health (p < 0.05). Children had more positive affect if they were female, younger, did not have a disability, and if their parent reported lower distress on the K6.

Brief Discussion

It was hypothesised that children who spend more time playing adventurously would have fewer internalizing symptoms and that children who typically spent more time playing adventurously would have less negative affect and more positive affect during the first weeks of the Covid-19 lockdown. Results support these hypotheses with the exception that there was no evidence of an association between adventurous play and negative affect.

The results provide some indication of specificity because no significant association was found between adventurous play and externalising problems. Furthermore, there was a positive association between time spent playing unadventurously and internalising problems, and no significant correlation between mental health and time spent playing outdoors, which further suggests specificity.

A limitation of this research is that the sample, although relatively diverse, was not recruited to be representative of the general population. Given the small effect sizes and confidence intervals close to zero, replication of these findings is important.