Introduction

There is growing interest in how play contributes to children’s well-being, and how it may support the development of skills needed to meet the demands of the social world1. For example, playful interactions provide children the opportunity to develop peer relationships in which they learn to solve problems and negotiate conflicts1,2. Playfulness is essential to play: it is what makes it imaginative and symbolic3. Playfulness has been defined in a number of different ways. Some definitions emphasize playfulness as a state of mind3,4,5, while others emphasize playfulness as a trait or disposition to play6,7,8. Indeed, the interactionist approach focuses on the interaction between the trait of playfulness, playfulness as a state of mind, and psychologically meaningful situations9. In addition to an overall score, several dimensions of playfulness have been described in the literature8. For the purposes of this study, we focus on playfulness as a trait, adopting Barnett’s definition of playfulness as the tendency to shape (or reshape) a situation in such a way as to entertain and/or amuse oneself and possibly others10. To assess young children’s playfulness as a trait, it is important to ask young children about their perceptions of themselves as playful beings, as playfulness is not always directly observable11. The current study takes this into account and examines the role of children’s own perceptions of their playfulness as a concurrent and longitudinal predictor of both teacher-reported social maturity with peers and peers’ own perceptions of their classmates as desirable playmates (i.e., social preference). Empirical evidence suggests that children’s playfulness is stable over time, using both observations and children’s self-reports12,13. Similarly, cross-sectional studies show that adults’ views of their own playfulness do not differ across different age groups, with older adults seeing themselves just as playful as younger adults14. Although playfulness is present at all ages, it likely takes different forms at different ages. For example, adults may be more reluctant to seek pleasure than children15. To understand the role of playfulness in children’s social-emotional development, it is good to know that playfulness is present in both and appears to be stable in children. The first, preliminary, aim of the current study is to assess the stability of children’s self-perceptions of their playfulness during the first three years of formal schooling, between ages 5 and 7.

Constructs closely related to playfulness, such as humor, suggest that there may be differences across boys’ and girls’ playfulness. For example, teachers consistently report that boys engage in more physical play and are more active and exuberant than girls16,17. Parents and teachers of 2–6 year olds have reported that girls show more cognitive and social spontaneity and less physical spontaneity than boys, although notably there was no difference in overall playfulness in this study18. In another study, girls showed less manifest joy and sense of humor than boys17. Notably, no differences were found between boys’ and girls’ playfulness when assessed using observational methods19, suggesting potential parent and teacher-biases in their perceptions of children’s playfulness and highlighting the importance of assessing children’s own perceptions of their playfulness. There is some evidence that sex differences may still emerge when using a self-reported approach to assessing playfulness. For example, girls aged 6 appear to perceive themselves as more playful than boys12, although this difference may disappear with age20. Taken together, there is a mixed picture within the extant literature regarding sex differences in children’s playfulness which underscores the importance of considering sex when examining playfulness in childhood. The second, preliminary, aim is therefore to examine sex differences in playfulness over time.

The core aim of the current study is to explore how playfulness, social maturity and social preference are related over time. Children’s social maturity has been conceptualised in multiple ways, with definitions emphasising social skills21 and/or adaptive interpersonal relationships22. The unifying element across these approaches is effectiveness in social interaction23. Importantly, by primary school, more than 30% of children’s social interactions are with their peers24. Children who are socially mature are often liked by their peers 25,26. Peer liking is widely assessed using sociometric status, or social preference, which captures peer insights into children’s social standing in the classroom27.

As children develop, social interactions become more sophisticated and they are required to use their social skills across different contexts effectively, making social skills more important and broader in scope28. This is particularly true as children enter formal schooling, which Bronfenbrenner describes as an ecological transition29. The desire to be socially accepted plays an important role in motivating children to adapt to new situations, such as the transition to school. This drive encourages them to develop a range of skills necessary to navigate their new social environment and form positive relationships with their new peers30,31.

Children’s early playfulness is associated with greater social skills32 and is a positive predictor of their later prosocial skills33. It is plausible that greater playfulness leads children to more play opportunities that, in turn, promote social skills and peer popularity34,35. This is supported by empirical research, for example, a study of 4-year-olds found that children with higher levels of social spontaneity (a component of playfulness6;), engage in more social play and have lower levels of withdrawn behavior than children with lower levels of social spontaneity3. As socially competent children are often liked by their peers25,26, it is likely that children’s playfulness contributes to their social preference.

However, playfulness does not always manifest as positive social behavior. In fact, it has been suggested that an impulsive expression of playfulness can be perceived as antagonistic and even destructive20. In children, playfulness and humor, both of which can be expressed as disruptive clowning behavior in class, can negatively impact a child’s relationship with both their teachers and peers32,36. It is possible that teachers’ own perceptions that more playful boys are less socially competent may, over time, influence peer preferences, which in turn may have a negative impact on social preference37. For example, teachers perceive playful boys (but not girls) as less socially competent36, and by age 8 there appears to be a shift in peer perceptions as well, such that the clowning behavior of boys (but, again, not girls) is perceived as disruptive by their peers36. In sum, playfulness, when expressed in a friendly way, can contribute to social maturity and social preference, but playfulness can also have negative social repercussions.

The current study

The current study first examines whether; (i) self-reported playfulness is stable over time, and (ii) differs as a function of sex, before exploring the longitudinal links between children’s self-reported playfulness, teacher ratings of social maturity and peer ratings of social preference across ages 5, 6 and 7 years, corresponding to the first three years of formal schooling in England, UK. We examined playfulness as a unidimensional construct, which has been shown to be a useful approach in young children12,38. As verbal ability has been associated with social competence39,40,41 and social play42, and as verbal ability is needed to communicate features of one’s self-concept43, we ensured that verbal ability at school entry was controlled for in our models.

Method

Participants

Data were derived from the Children’s Relationships with Peers through Play (ChiRPP) study44, an England-based longitudinal study which explores children’s social play and peer relationships during the first three years of primary school. Trained researchers collected data from children using a protocol receiving ethical approval from the University of Cambridge Psychology Research Ethics Committee.

The full sample comprised 244 children (131 boys, 54%) first seen at Reception (the first year of formal schooling in England). The current study reports on the longitudinal ChiRPP sample who were followed up in Year 1 and Year 2 recruited with parental consent from 5 schools across Cambridgeshire, UK. All Reception classrooms in each school were invited to participate. The longitudinal sample comprised 172 children (99 boys, 58%, M = 63 months, SD = 3.86 months), with 84% of participants having at least one parent with a university-level degree, and 7% of children qualifying for pupil premium (additional funding for disadvantaged students of all ability levels). Participants were predominantly White (79.5%), with 12.9% Asian, and 7% of mixed ethnic background. Data were collected between April to July 2016 (Reception; Time 1), January - May 2017 (Year 1; Time 2), and October 2017 - February 2018 (Year 2; Time 3). There were on average 294 days (SD = 74 days) between Time 1 and Time 2, and 268 days (SD = 18 days) between Time 2 and Time 3.

At Time 2 (M age = 73 months, SD = 4.53, 58% boys), 161 were retained in the sample. Finally, at Time 3, the sample comprised 152 children (M age = 81 months, SD = 4.69, 59% boys). Children dropped out because they moved schools, no child dropped out because of revoked parental consent. There was no significant difference in age, playfulness, social maturity or social preference between those children who were included compared to those lost from the sample between Time 1 and Time 2, ts ≤ 1.52, p ≥.064, nor was there a difference between the number of boys and girls lost from the sample, 2 = 0.71, p =.401. However those lost from the sample between T1 and T2 had significantly lower verbal ability scores compared to those who were retained, t(170) = 2.58, p =.005. There was no significant difference in age, playfulness or teacher-reported social maturity between those children who were included compared to those lost from the sample between Time 2 and Time 3, ts ≤ 1.27, p ≥.104, nor was there a difference between the number of boys and girls lost from the sample, 2 = 0.53, p =.467. However those lost from the sample between Time 2 and Time 3 had significantly lower social preference at Time 2 compared to those who were retained, t(159) = 2.10, p =.019.

Measures

Playfulness

The Child Self Report Playfulness scale (CSRP;12) measures children’s perceptions of their own playfulness. The CSRP items were presented to each child on a tablet. They watched a video of two gender-neutral puppets introducing themselves and the rules of the task. Using a neutral intonation, the puppets described themselves with one of two opposite statements. The child was asked to point to the puppet whose statement was most similar to their own. After two practice trials, which were unrelated to the content of the study, the child was presented with 11 playfulness items (e.g., “I make up new games to play”). The experimenter was seated next to the child during this task to record responses and provide explanations or prompts to the child if necessary. Based on previous findings12 only nine items were included in the total CSRP score which therefore could range from 0 to 9. In the current study, reliability of the 9-item scale ranged from acceptable to good, alphas (using the omega function based on tetrachoric correlations given binary nature of items) at Time 1 = 0.80, Time 2 = 0.73, and Time 3 = 0.75. To further confirm that the 9-items all loaded onto a single factor, a confirmatory factor analysis was conducted using mean- and variance-adjusted weighted least squares (WLSMV) estimation given the binary nature of the items. The model fit the data well at all three time-points, Time 1: 2 (27) = 36.18, p =.111, RMSEA = 0.05, CFI = 0.94, TLI = 0.92, with standardised factor loadings ranging from 0.38 to 0.74; Time 2: 2 (27) = 38.40, p =.072, RMSEA = 0.05, CFI = 0.81, TLI = 0.75, with standardised factor loadings ranging from 0.22 to 0.58; Time 3: 2 (27) = 37.48, p =.086, RMSEA = 0.05, CFI = 0.87, TLI = 0.83, with standardised factor loadings ranging from 0.25 to 0.83.

Social maturity

The Peer Social Maturity scale (PSMAT;45) is a teacher-report rating of children’s social skills during peer interactions. The PSMAT consists of 7 items. Teachers were asked to rate individual children on six dimensions of socially mature behavior based on what the teacher knows to be typical or average for children of that age (e.g., “The maturity of the child’s everyday modes of playing sociably with peers”). For each item, the child was rated on a scale from very much below average for their age (a score of 1), to about average for their age (a score of 4), to very much above average for their age (a score of 7). The total PSMAT score was calculated by adding the scores for each item and could range from 7 to 49. Internal consistency of the PSMAT is excellent with alphas ranging from 0.95 to 0.98 across classroom samples46. Individual children’s PSMAT scores are stable over time with longitudinal correlations ranging from 0.48 to 0.5246. In the current study, internal consistency was also excellent, with Time 1 Cronbach’s alpha = 0.95, Time 2 Cronbach’s alpha = 0.98 and Time 3 Cronbach’s alpha = 0.97.

Social preference

Social preference was assessed using the peer nomination sociometric interview technique27. Children were asked, using photographs of their peers as prompts, to nominate three children in their class with whom they “most like to play” and three children with whom they “least like to play”. Cross-sex nominations were allowed47. Each child’s individual “like most” and “like least” scores were standardised within the classroom to account for classrooms of different sizes. Social preference was calculated by subtracting the classroom standardised total of “like least” from the standardised total of “like most” nominations.

Verbal ability

The CELF Preschool 2 UK48 measures expressive and receptive language skills in young children. Two subtests were used in the current study at T1 and T2: (i) the Sentence Structure subtest, and (ii) the Sentence Recall subtest. In the Sentence Structure subtest, which assesses understanding of grammatical rules at sentence level and consists of 22 items, the experimenter read a sentence aloud and the child had to choose one of four pictures presented that best matched the sentence. For the Sentence Recall subtest, which assesses oral language expression, children had to repeat sentences of increasing complexity that were read to them by the experimenter. Given the age of children at T3, the CELF Preschool 2 was no longer an appropriate measure of language ability. At this time-point children completed the 31-item Sentence Comprehension subscale as an index of receptive language, and the 25-item Naming subscale as an index of expressive language of the Assessment of Comprehension and Expression49.

Procedure

At each time-point, children were seen by trained graduate-level researchers individually in a quiet room at their school. Children completed the language assessment as per the manualised instructions, in addition to the sociometric interview and child self-reported playfulness scale. Other measures were also collected that are not the focus of the current study44. Teachers completed a paper version of the Peer Social Maturity Scale after at least one term of teaching and knew the children well.

Analytic strategy

For all measures used in the study, a z-test was used to test for normality using skewness and kurtosis. Due to the large sample size (50 < n < 300), the null hypothesis was rejected if the skewness and/or kurtosis had an absolute z-value greater than 3.2950. The scores for all measures were normally distributed except for social preference at Time 2 (z = 4.66), but the kurtosis was between − 2 and 2 and the skewness was between − 1 and 1, indicating that the distribution was not too peaked or skewed51.

Using SPSS v28 a mixed ANOVA with repeated measures was performed to examine the effect of time and sex and a possible interaction effect on playfulness scores. Mauchly’s test was used to test the sphericity assumption. Pearson’s correlations were calculated to measure the linear association between playfulness, social maturity and social preference.

A cross-lagged panel model was then constructed in MPlus v852, including playfulness, social preference and social maturity at all three time-points, to model within-construct stability as well as cross-construct relations within and across time-points. Verbal ability was included as a covariate and regressed onto the Time 1 variables. The MLR estimator was used as it is robust to non-normality and accounts for missing data. Given the exploratory nature of the analysis all paths across time and constructs were included. To examine sex differences in the model paths, two multigroup models were constructed. In the constrained multigroup cross-lagged panel model, paths for boys and girls were constrained to be equal. In the unconstrained cross-lagged panel model, paths were permitted to freely vary53. In order to compare model fit for the constrained and unconstrained model when using MLR as an estimator, a scaling correction was applied to the chi-square difference test54.

Results

Descriptive statistics

Descriptive statistics are displayed in Table 1. Note that the number of participants differed between measures and timepoints.

Table 1 Descriptive statistics for study variables.
Full size table

Preliminary analysis

Stability of playfulness and sex differences

A mixed ANOVA with repeated measures was performed to examine the effect of time and sex and a possible interaction effect on playfulness scores in 150 participants. The sphericity assumption was tested using Mauchly’s test, which was significant, χ2(2) = 0.95, p =.031. Therefore, the degrees of freedom were corrected using the Greenhouse-Geisser method, ε = 0.96. There was an overall significant difference between the playfulness scores at different time points, F(1.912, 282.980) = 4.25, p =.016, η2 = 0.03. Post hoc tests using Bonferroni correction revealed that playfulness scores were significantly lower at time 1 (M = 6.78, SE = 1.92) than at time 3 (M = 7.30, SE = 1.62), mean difference = − 0.48, 95% CI [−0.93, − 0.03], p 0.032. There was no significant difference in playfulness scores between time 1 and 2, mean difference = − 0.11, 95% CI [−0.54, 0.32], p = 1.00 and between time 2 and 3, mean difference − 0.37, 95% CI [−0.00, 0.74], p =.052. Although there were changes in the overall level of playfulness, the ranking of children’s playfulness scores was quite stable, with correlation coefficients ranging from 0.21 to 0.27 (see Table 2).

There was also a significant main effect of sex, such that boys (M = 6.8, SD = 1.4) had significantly lower playfulness scores than girls (M = 7.2, SD = 1.3), F(1, 148) = 5.02, p 0.027 However, there was no significant interaction effect between time and sex on playfulness scores, F(1.912, 282.980) = 1.95, p =.147, η2 = 0.01. This means that the effect of time on playfulness scores was similar for boys and girls and that the sex difference was similar across time points.

Correlations between the study variables

Table 2 shows the bivariate correlations between the study variables (above the diagonal). Partial correlations were also calculated to determine the relationships between playfulness, social maturity and social preference while controlling for verbal ability (Table 2, below diagonal).

With and without controlling for verbal ability, social maturity and social preference were significantly and positively related to each other at all time points. Only at Time 3 playfulness was significantly and positively related to social maturity and social preference. When controlling for verbal ability, playfulness was also positively related to social preference at Time 2.

Table 2 Summary of bivariate correlations between study variables (above the diagonal), also while controlling for Language skills at time point 1 (below the diagonal).
Full size table

Longitudinal associations between playfulness, social maturity and social preferences

To examine the associations concurrently and longitudinally across all three core study variables; playfulness, social maturity and social preference a cross-lagged path model was constructed, with verbal ability regressed onto Time 1 variables. The path model showed good fit to the data, 2(15) = 19.77, p =.181, RMSEA = 0.043, CFI = 0.986. All three variables demonstrated significant stability over time, and verbal ability at Time 1 only significantly predicted Time 1 social maturity, it was not significantly associated with Time 1 playfulness or social preference. Time 1 social maturity significantly predicted Time 2 social preference. Notably both Time 2 social maturity and Time 2 social preference significantly predicted Time 3 playfulness. Time 2 social preference also predicted Time 3 social maturity (see Fig. 1). Significant indirect effects were also found, such that Time 1 teacher-reported social maturity predicted Time 3 playfulness indirectly, mediated by both Time 2 teacher rated social maturity (standardised coefficient = 0.08, p =.039) and peer-reported social preference at Time 2 (standardised coefficient = 0.07, p =.009). In addition to a bidirectional relationship between social maturity and social preference, both social maturity and social preference were found to predict children’s perceptions of their playfulness.

Fig. 1
figure 1

Cross-lagged panel model showing significant standardised coefficients, with standard errors in parentheses (dashed paths non-significant). Note. * p <.05, ** p <.01.

Full size image

Given possible sex differences in the pattern of associations between social maturity and playfulness, a multi-group path analysis was conducted where an unconstrained model was compared to a model where paths were constrained to equality for boys and girls. There was no significant difference between the two models, Satorra-Bentler 2 difference (TRd) = 27.05, Δdf = 30, p =.621, suggesting that there is no difference in the pattern of associations for boys and girls.

Discussion

The current study examined young children’s perceptions of their own playfulness and its concurrent and longitudinal associations with teacher-reported social maturity and peer-reported social preference across the first three years of formal schooling. Individual differences in self-reported playfulness appeared to be reasonably stable over time, suggesting that a child who reported being playful at age 5 would also report being playful at ages 6 and 7. This finding is consistent with other longitudinal research on features of children’s self-concept55,56. In addition to the stability of individual differences over time, the results of the current study also suggest a consistent increase in children’s perceptions of their own playfulness between the ages of 5 and 7 years. Physical play is thought to begin early in infancy, peak in childhood and then decline57. Given that playfulness can be seen as a disposition to play8, the increase in children’s perceptions of their playfulness between the ages of 5 and 7 may reflect the peak of play in childhood. Another suggestion is that children’s increasing perceived playfulness reflects how their environment is responding to them. For example, research has shown parental beliefs about play are positively associated with their children’s playfulness32.This is consistent with the idea that the overt expression of personality traits can change as a child matures and learns55. A recommendation is made under the ‘Future Research’ section to include family environment in studies on playfulness.

Sex differences in playfulness were observed, such that boys reported lower levels of playfulness than girls. This sex difference was similar across time points. Other studies of playfulness in children of the same age have not found this difference19,36, notably however, these studies have typically assessed playfulness via teacher- or parent-report. As playfulness is not always reflected in overt behavior11,58 and varies from context to context59, children’s perceptions of their own playfulness may differ from the perceptions of other informants60. However, the perceptions of others likely also play a role in sex differences in children’s self-reported playfulness. In young children, physical spontaneity (a component of playfulness6;) seems to be the most important component of playfulness60, and boys generally show more physical spontaneity than girls6. As this component of playfulness seems to be less tolerated by others as children develop61, and feedback from others influences how children see themselves62, this may explain why, in the current study, boys perceive themselves as less playful than girls.

The core aim of the current study was to examine the associations between child-reported playfulness, teacher-reported social maturity and peer-rated social preference. Children’s perceptions of their playfulness at age 7 were predicted directly by both earlier teacher-ratings of social maturity and peer-ratings of social preference. There were also indirect associations, such that children’s perceptions of their playfulness at age 7 were predicted by teacher-rated social maturity at age 5 via both teacher-rated social maturity and social preference at age 6. The results of the current study present a complementary picture to existing concurrent and longitudinal findings showing that playfulness is associated with increased social skills31 and positively predicts children’s later prosocial behaviour32. As other empirical research has shown63 positive social behavior establishes a positive reputation with teachers and peers, and through these social relationships and the feedback children receive on their interactions their social reputation grows and influences how they see themselves62. This study shows that children see themselves as more playful when their teachers are positive about their social maturity and when their classmates prefer to interact with them.

We also found that higher teacher ratings of social maturity predicted peers’ social preference, which in turn predicted higher teacher ratings of social maturity. This suggests a positive feedback loop between teachers’ ratings of social maturity and peers’ social preferences, in line with previous research64,65,66. These findings reflect the evaluation of social behavior by teachers and peers64,65.

Limitations and future research

The current study has a number of strengths, namely its longitudinal design with three time-points over 24 months, multi-informant measurements, and a focus on the child’s own perceptions of their playfulness. However, there were also some limitations. First, the sample was very homogeneous with respect to socioeconomic status and culture, with a large majority of children coming from families with a high level of education and a White British background within a narrow geographical radius of a university town in England. As such, the degree to which these results are replicable across different groups of children is yet unknown. Further research is clearly needed with a broader group of children, especially given differences with respect to the value placed on play and playfulness in different societies and the implications this may have for children’s own perceptions of their playfulness67.

Second, playfulness and its implications for peer relationships is likely to be contextually-bound. That is, in different contexts children are likely to be more or less playful. For example, consider a school in which a playful approach to learning is encouraged compared to one in which a more traditional view of learning is ascribed68. In these different contexts, children with the same tendency to engage in playful behaviours are likely to perceive their playfulness differently, as expressing it will be perceived differently and might have different consequences. In the current study we were unable to take into account school- and teacher-level differences in the value placed on children’s playfulness, we expect this to be a fruitful area of future study. Relatedly, children’s perceptions of their own playfulness are also determined by the family environment, as parents’ own playfulness correlates with children’s self-perceived playfulness69. The interplay between a child’s family perceptions of playfulness and that of their peers would further help to unpack the developmental trajectories of children’s playfulness. Additionally, future research employing self-report measures with distinguishable domains, such as the Children’s Playfulness Scale (CPS)6,60, may contribute to a more nuanced understanding of differences in children’s perceptions of playfulness that captures its multidimensional nature.

In summary, this study builds on previous research by exploring the associations between playfulness, social maturity and social preference over time. This helps us to understand how playfulness plays a role in children’s social-emotional development, which may be particularly important during potentially challenging ecological transitions, such as when a child moves to primary school and needs to establish and maintain new relationships with peers and teachers. The current study highlights that how children feel about their own playfulness is predicted by how significant others, in this case teachers and peers, view and judge their social behavior. It is crucial that teachers understand that their interactions with children profoundly influence the children’s self-perception and position in the group, even in late childhood65. Teachers should therefore be encouraged to acquire this knowledge and be supported in engaging in positive interactions with the children. One way of achieving this would be through play70.