Although we know that the complexity of pretend play increases as a function of age (Ariel 2002; Szabó 2014), there are limited studies reporting the developmental trends of pretend play behaviours of typically developing pre-schoolers (Lillard et al. 2013). The knowledge of pretend play in typically developing pre-schoolers would be beneficial in the early identification of deviations in pre-schoolers who have or are at risk of developing developmental disabilities. Children with developmental disabilities exhibit delay in pretend play behaviours compared to their typically developing peers (Barton & Wolery 2008). A delay in the occurrence of pretend play could be a potential diagnostic marker in identifying several child developmental disorders such as autism spectrum disorders, attention deficit hyperactivity disorder, and hearing impairment, thus indicating the pressing need to assess the pretend play behaviours exhibited by these children. In order to compare and contrast deviant pretend play behaviours, assessing the occurrence of pretend play behaviours in typically developing pre-schoolers is essential. Typically developing pre-schoolers spend a substantial portion of their time engaging in object play in their schools. This study could help in understanding the developmental trend of pretend play and also help in the formulation of Individualised Education Programme objectives for children with developmental disabilities involving play. This would enable children with disability experience normalised, contextually pertinent experiences like their peers.

Profiling pretend play behaviours exhibited by pre-schoolers in school setting will contribute to the emerging concept of play-based assessment and intervention which helps to gather information about the developmental deviations of a child in developmental domains such as communication, gross and fine motor skills in a real and functional context. To address the aforementioned limitations the present study aimed at profiling the developmental trends of pretend play in pre-schoolers and to understand the influence of play scenarios (free and structured toy play) on pretend play behaviours.

Forty-eight typically developing children aged between 3.5 and 5.6 years of age were divided into four groups (Group I: 3.5–4.0; Group II- 4.1–4.6; Group III- 4.7–5.0; Group IV- 5.1–5.6 years) with 12 children in each group. The children were recruited from three pre-schools following the English medium of instruction. All schools were located in the Dakshina Kannada District of the state of Karnataka, South India. Dakshina Kannada District is a district in Karnataka with a total population of 2 089 649, and seven taluks (Chandramouli 2011).

The participants’ selection for the current study followed stringent criteria. The Modified Kuppuswamy Classification of Socio-economic Status Scale for the financial year 2019 (Mohd Saleem 2019) was used to determine the children’s socio-economic status. This scale estimates the socio-economic status by considering the family’s education, occupation, and income. As Socio Economic Status (SES) influences and predicts several aspects of a child’s development with its influence being pervasive in nature, and across all domains (Hoff, Laursen & Tardif 2002), the current study included children only belonging to the upper middle class SES to avoid the influence of SES on the pretend play behaviours in pre-schoolers. The ‘Assessment of Language Development’ (Lakkanna, Venkatesh & Bhat 2008) was used to determine children’s receptive and expressive language ability. This tool is a norm-referenced performance-based tool used to assess children’s receptive and expressive language abilities from birth to 7 years and 11 months. Children with age-appropriate receptive and expressive language skills were included in the study. The 10 questions of the World Health Organization (WHO) Disability Checklist (Singhi et al. 2007) were used to rule out any disabilities. This checklist is a questionnaire to screen for disabilities such as physical, mental, speech, hearing, and visual impairments, and epilepsy in children between 2 and 9 years. Children who passed the screening test were recruited for the study. The pre-schools that were selected followed state board curriculum with English as a medium of instruction. All the teachers who were employed in these schools had completed a Basic Training Certificate (BTC) and a Diploma in Education.

The investigators visited each of the three schools before the study’s commencement to select suitable recording location. An appropriate classroom was selected which had limited auditory and visual distractions to ensure high quality pretend play. The classroom had adequate space, lighting, ventilation, and play mats to sit comfortably and play. This environment was chosen for recording as it promotes more natural peer interactions (Abbas, Othman & Megat Abdul Rahman 2016). The selected classroom was rearranged by moving the furniture away from the centre of the room in order to provide space for the child to play. The selected room for observation was devoid of any play materials or toys. The recording of the play skills of children was conducted using Sony-HDRCX405 camcorder. The Camcorder was placed in a suitable position on a tripod stand to cover the entire play area to avoid personal handling by the investigator during the recording session. Video-based analysis was preferred over other forms of observational analysis, as it is repeatable and practical for offline analysis by multiple observers (Sommer, Hachul & Roßbach 2016). The investigator remained as a passive observer throughout the complete duration of each of the recordings.

The study attempted to analyse pretend play in two play scenarios: Free Play scenario (FP) and Structured Toy Play scenario (SP). In the FP scenario, the investigator instructed the children to play anything, with any material available in the classroom. Here, the investigator did not provide any specific toy kit, but they were allowed to play with any other materials that were available in the classroom (chalk, duster, board, charts, etc.). However, in the SP scenario, the investigator provided children with toy kits which were placed in the centre of the classroom. The toys provided to all the children in the present study included a combination of gender-neutral (lego blocks, ball, miniature doctor’s kit) and gender-specific toys (Male: superhero toy, Female: doll) (Cherney & London 2006) that was aimed to elicit a variety of pretend play acts. Each child was provided with a toy kit to ensure that every child had the opportunity to play with all the toys in the toy kit.

The investigators visited the schools once a week for four consecutive weeks and scheduled a video recording session which lasted an entire day. The investigator recorded the play scenarios twice a day (morning and afternoon) to avoid fatigue or lack of interest. During pilot study, the investigator observed that SP scenario influenced the FP scenario. Hence, the recording sequence was maintained for all the children, with morning session targeting FP and afternoon session targeting SP. A 20 min recording of each scenario for every child was obtained (20 min recording, 10 min coding) with an average of 5 min recording per session per week. Children were divided into a group of four and the members of the group remained constant for both scenarios every week.

The coding scheme used for the current study was adopted from Play in Early Childhood Evaluation System (PIECES) (Kelly-Vance & Ryalls 2005) as this scheme elaborately describes pretend play. According to this coding scheme, pretend play was divided into simple and complex pretend play (CPP). Simple Pretend Play (SPP) consisted of play acts such as Approximate Pretend Play (AP), Self-Directed Play-Act (SD), Object Directed Play-Act (ObD), Repetitive Combination (RC), Agentive Play-Act (AGP), Object Substitution Play-Act (OSub), and Self Substitution Play-Act (SSub). Complex pretend play consisted of play acts such as 2 step Complex Pretend Play (2CPP), 3 step Complex Pretend Play (3CPP), 4+ Complex Pretend Play (4CPP), Pre-Planned Complex Pretend Play (PCPP), Complex Pretend Play with Substitution (SCPP), and Multi-Theme Complex Pretend Play (MCPP). The complete coding scheme is provided in Appendix 1. Coding was carried out in both FP and SP scenarios. The combined frequencies of AP, SD, ObD, RC, AGP, OSub, and SSub formed the SPP frequency. Similarly, combined frequencies of 2CPP, 3CPP, 4CPP, PCPP, SCPP, and MCPP formed the CPP frequency.

The interrater reliability was ascertained by two investigators who were Speech-Language Pathologists with more than 3 years of experience in the field. Both investigators were blinded to the child’s age, and therefore independently coded the entire data as per the coding scheme based on the frequency of pretend play occurrence. The investigators were trained to code pretend play behaviours in practice trial sessions conducted over a week and substantial interrater agreement (κ = 0.642, p < 0.01) was obtained. Each investigator took an average of 40 min to complete the coding of both FP and SP scenarios for each child.

The statistical analysis was performed using SPSS software version 25. The current study considered groups (Group I, Group II, Group III, and Group IV) as independent variables and pretend play behaviours (SPP and CPP) as dependent variables. The obtained data were subjected to descriptive statistical analysis (frequency and range). The Kolmogorov-Smirnov and Shapiro-Wilk test was performed to ascertain the data distribution, which warranted non-parametric tests. Kruskal Wallis H test was conducted to analyse the difference between SPP and CPP across the different age groups. Post hoc test was performed using Mann Whitney U Test to analyse pairwise comparison between the different groups for SPP and CPP in both SP and FP scenarios. Wilcoxon Signed Rank Test was conducted to analyse the influence of FP and SP scenarios on SPP and CPP occurrence across the four age groups. Spearman’s Rank Order Correlation was used to check the relation between age and occurrence of SPP and CPP in both the FP and SP scenarios.

The developmental trends of pretend play during FP scenario was ascertained using Kruskal Wallis H test that revealed a significant difference (p < 0.05) between the age groups in both CPP and SPP. To further analyse which of the age groups exhibited significant difference, a post hoc test of pairwise comparison was performed using Mann Whitney U test with Bonferroni correction for multiple comparisons (p < 0.008). The Mann Whitney U test findings in the FP scenario are described in Table 1 and Table 2.

To ascertain the correlation between age and pretend play behaviours, Spearman’s Rank Correlation test was performed. The results of Spearman’s Rank Correlation test revealed that in CPP, there was a strong positive correlation (r_s = 0.746, p = 0.000) and SPP, a moderate positive correlation (r_s = 0.430, p = 0.002) indicating an increase in pretend play complexity across the groups for both SPP and CPP.

Kruskal Wallis H test revealed a significant difference between the groups in CPP (p < 0.05) and not SPP. This indicates that although a developmental trend was observed in both SPP and CPP, only CPP demonstrated a significant difference. A pairwise comparison was performed using Mann Whitney U test with Bonferroni correction for multiple comparisons (p < 0.008). A statistically significant difference (p = 0.006) was observed only for group I and IV comparison.

The results of Spearman’s rank correlation revealed that in SPP (r_s = 0.304, p = 0.036) and in CPP (r_s = 0.371, p = 0.009), there was a significant but weak positive correlation between age and pretend play behaviours. A weak positive correlation indicates that both SPP and CPP demonstrated an increase in occurrence with age.

Wilcoxon Signed rank test results indicated a significant effect of play scenarios on the frequency of SPP occurrence (Z = –3.991; p = 0.000) and CPP occurrence (Z = 3.136; p = 0.002). The finding revealed that the play scenario has a significant influence on the pretend play skills demonstrated by pre-schoolers.

The study findings revealed that pretend play (SPP and CPP) was observed to have the highest frequency of occurrence in 5.1 to 5.6-year old pre-schoolers and lowest in 3.5 to 4 year old pre-schoolers. This indicates a developmental trend with higher occurrences of complex forms of pretend play in older pre-schoolers (Melzer & Palermo 2016). One possible explanation for the developmental trend of pretend play could be the bidirectional influence of executive functions on pretend play. The development of executive functions could have influenced the higher occurrence of pretend play in older children, while children’s engagement in pretend play could have also facilitated the development of executive function (Thibodeau et al. 2016). Moreover, the development of core component of executive functions (cognitive flexibility, inhibition, and working memory) increases the occurrence of pretend play in older children (4.7–5.6 years). The development of core components like cognitive flexibility and working memory is critical for switching back and forth between the make-believe world and to remember the rules of pretence during the switch between scenarios successfully (Thibodeau et al. 2016), which could have had a role in the improvement of pretend play in older pre-schoolers.

Another possible reason could be the influence of creativity on pretend play (Bergen & Mauer 2000; Holmes et al. 2015). The increased occurrences of complex forms of pretend play in the older groups (4.7–5.6 years) in the present study could be attributed to the progressive development of creative skills, which is also in agreement with the study conducted by Bergen and Mauer (2000). Creativity is essential in pretend play as it helps in the symbolic transformation of objects and action, which enhances the ability to involve in complex forms of pretend play (Holmes et al. 2015). As creativity increases with age, it could have influenced the pretend play behaviours in older pre-schoolers. Although developmental markers of simple forms of pretend play were present among the younger groups (3.5–4.6 years), the frequency of complex pretend play was considerably less.

Furthermore, the nature of the FP scenario could have contributed to an increased occurrence of pretend play. As the FP scenario did not have any definite set of toys, children’s imagination and creativity were not constrained. Hence a wide repertoire of pretend play forms was observed. The finding aligns with Wooley and Tullos’s (Woolley & Tullos 2008) observation who reported that children’s pretend play tends to be more abstract during pre-school years and not dependent on toys and objects. Studies have also reported that children develop the ability to make up their own fantasy play and modify the script (Goldstein 2012) without depending on toys or play objects during the pre-school period.

In addition to these findings, the investigators also observed that most children adopted pre-play behaviours such as gestural (pointing, eye contact, facial expression) and verbal behaviours to attract their play partner’s attention. This was observed prior to the commencement of play, as well as throughout the FP duration. The presence of pre-play behaviours indicates the children’s interest in involving others in play and establishing common grounds for play acts (Gibson et al. 2020) which is the foundation for pretend play. The FP scenario’s flexibility because of the absence of toys has increased the opportunities for verbal interaction in children, which is also in accord with the study by Weisberg et al. (2013).

In the SP scenario, a developmental trend was observed for both CPP and SPP. Analysis of CPP occurrences revealed a higher frequency in older children (5.1–5.6 years) than younger children (3.5–5 years). This is further supported by Lillard’s (2017) findings, who reported that older pre-school children start pretending with objects more than younger children. The most compelling explanations for the major observations of the SP scenario are described below.

Firstly, the higher frequencies of pretend play (both SPP and CPP) in children provide evidence of cognitive skills development as a function of age. This is in accord with Mathew’s (Matthews, Stienstra & Djajadiningrat 2008) findings, who has reported that with the development of cognitive abilities, children also develop better attention shifting ability, which is a requirement for complex forms of pretend play. Besides, children’s pretend play mainly consists of activities that are comprehendible and controllable by them (Cohen 2018); this could also be the reason for the finding of the present study where younger children (3.5–4.6 years) exhibited fewer occurrences of CPP with higher occurrences of SPP.

Secondly, in the SP scenario, children were provided with a toy kit that included gender-neutral and gender-specific toys that varied in their structure (Low structure and High structure). High structure toys are more realistic (e.g. Doctor’s kit, Doll) than low structure toys (e.g. Legos) (Cho 2000). As most of the toys included in the toy kit were of high structure (Doctor’s Kit, Doll, and Super Hero Toy), the types of pretend play demonstrated by the children could be limited. This restriction is a result of the challenges that the child faces in violating the rules of toy usage (Cho 2000). Another possible reason could be attributed to the high structure toys’ realistic nature, which could have restricted children from engaging in object-directed pretend play (Cho 2000). The present study findings support the evidence that children exhibit a limited type of pretend play when a restrictive set of toys are given.

Another interesting finding of this study is the reduction of gestural and verbal behaviours exhibited by children in SP compared to the FP scenario. This could be attributed to the usage of toys, which could have reduced social pretend play occurrences and increased the occurrences of object-based pretend play (Weisberg 2015). The finding also confirms the reduction of interactive forms of pretend play among children in the presence of toys (Sosa 2016). As children are paired with the same play partners in both SP and FP scenarios, good rapport was already established with their peers in the FP scenario. The involvement of verbal and gestural behaviours during pre-play and during play situations was to set up mutual understanding between play partners. During the SP scenario, as the children were already familiar with their play partners, this scenario witnessed reduced verbal and gestural behaviours.

The comparison of SP and FP scenarios revealed a significant difference in SPP and CPP occurrence, indicating that the play scenario influences pretend play in pre-schoolers. In the present study, the FP scenario elicited higher frequencies of CPP than the SP scenario. The factor that could have influenced the difference in SPP and CPP occurrence could be the presence of a toy kit. Toys trigger symbolic expression in children, and the presence of various toys stimulates the child to prolong the play in a particular way (Goldstein 2012). When children are provided with low structure toys, they exhibit higher frequencies of pretend play (Cho 2000). However, most of the toys included in the toy kit were high structure in the present study. Therefore, this scenario elicited higher frequencies of object-directed pretend play and non-interactive type of play than socio-dramatic forms of pretend play.

On the other hand, the FP scenario required children to utilise their creativity and cognitive abilities to play, which provided a vital experience for children to learn social, cognitive, and communicative skills (Santer et al. 2007). This does not mean that SP hinders the development of a child; rather, it facilitates the development of self-regulation, turn-taking skills, and planning of play acts (Healey & Healey 2019). As children grow older, they become more sensitive to the context of the play, such as the play scenarios and play partners (Heimann, Laberg & Nordøen 2006). This could have led children to respond differently to variations in the social contexts and modify their behaviour to meet the contextual requirements. Although the study had several contributing findings, there were certain limitations like, reduced sample size, restricted toy combinations and the sample being limited to upper-middle SES.