Introducing the Play in Education, Development and Learning (PEDAL) Research Centre

ABSTRACT

This paper provides an overview of the work being conducted at the Play in Education, Development and Learning (PEDAL) Research Centre, based in the Faculty of Education, University of Cambridge, UK. PEDAL has three main aims, (1) To conduct world-class research, (2) To build capacity in play research and (3) To influence policy and practice. The present paper provides an overview of the history and rationale for PEDAL, followed by some detailed examples of the foci of its work. Exemplar research is presented on the theme of ‘Measuring Play’; an endeavour that is common to many of the different research projects led by PEDAL investigators and Ph.D. students.

KEYWORDS:

• Play
• measurement
• child development
• research centre

What is the PEDAL research centre?

The Play in Education, Development and Learning (PEDAL) Research Centre was established in 2015 following a donation from the LEGO Foundation to both establish a research centre and to endow a new Professorship in this area of research at the University of Cambridge. It was initially established and led by Dr David Whitebread and then Professor Anna Vignoles as interim director. Since 2018, it has been directed by Professor Paul Ramchandani, the LEGO Professor of Play, in conjunction with Dr Jenny Gibson and Dr Sara Baker, at the Faculty of Education at the University of Cambridge.

In establishing a multi-disciplinary research centre to investigate the role of play in child development and learning, the team have had three main goals from the outset:

1. To conduct world-class research. With the aim of producing independent, peer-reviewed rigorous research about play and to break new ground in terms of play research methodology.

2. To build capacity in play research. PEDAL offers to supervise and support Ph.D. studies to students from around the world. We also mentor postdoctoral research associates and research assistants as they develop their academic careers. The PEDAL team teaches undergraduate and postgraduate students in the Faculty of Education. PEDAL also hosts international seminars and conferences to share current thinking and contribute to scientific debates about play.

3. To influence policy and practice. PEDAL’s research aims to shape opinion, policy and practice. We communicate our findings to a variety of non-academic stakeholders, helping them to make informed choices about the role of play at home, in school and in wider society. This is challenging work, but we work with a wide variety of stakeholders to reach those in practice in education and child health and development and those who make policy decisions about children’s lives.

Why establish a research centre on play?

Play is an integral part of a happy childhood. All children and most adults engage in playful activities, including physical play, play with objects, symbolic play, pretence play and games with rules. Whilst much is known about play and it is integrated into education and child development practices in many places, its role and importance remain contested (Nathan, Pellegrini, & Burghardt, 2010). Many educationalists and child development specialists make the case that play enhances learning, has a clear role in education, and supports emotional well-being, good mental and physical health, creativity and social competence. It follows from this that play may have the potential to contribute hugely to social, cultural and economic development, as well as to the fulfilment of the individual potential of children. However, whilst these areas remain contested, there is a need for high-quality, robust research to more fully understand the place of play in children’s lives. Underpinning this is the aim of giving children and young people the very best start in life.

What does the PEDAL centre do?

Now, 4 years on from its establishment, the centre has 15 staff and 9 doctoral research students, who, along with masters and undergraduate students undertake a wide range of research studies in the field of play. The team work in three broad areas of research:

• Social Play: Social Lives, led by Jenny Gibson, studies children’s peer play and interaction and how it relates to wider social development in middle childhood.

• Play in Early Life, led by Paul Ramchandani, studies the development of play and infant and child development from the start of life, including early parent–child relationships.

• Stepping Stones in Science: Guiding Playful Learning, led by Sara Baker, studies the role of child-led learning in primary school, with a particular focus on science education.

As well as continuing to undertake and publish research studies, over the past year we have increased the international focus of our work, our interactions with policy-makers and practitioners and have launched a new website and repository (PEDAL Hub – www.pedalhub.net), which combines a searchable library of play research with up to date blogs and articles on play and play-related research and practice. At present most of the content is produced by the PEDAL team, but we hope to encourage others to contribute and to use the resources and links featured on the site.

One of the most exciting aspects of PEDAL is that the study of play forms a focal point around which ideas from different disciplinary perspectives coalesce. The current PEDAL team has members with backgrounds in teaching, educational assessment, psychology, psychiatry, applied linguistics, public health, international development, and theatre studies, to name a few. Further, we are delighted that our PEDAL team is drawn from across the globe including members from St. Lucia, UK, France, China, USA, Kenya, India, Ghana, Republic of Ireland, and beyond. This intersection of diverse backgrounds and disciplines enables each of us to think about play in new ways and to explore new research methods and questions. For example, recent projects have drawn on methods from neuroscience, philosophy, computer science and architecture, none of which are ‘core’ background disciplines in our team. We have found it necessary to expand our research methods toolkit as the questions we are asking about play become more nuanced and complex.

As a consequence of this broadening perspective, we have established many connections with external collaborators. A recent initiative concerning ‘Play and Health’ has led to a collaboration with researchers, clinicians and other stakeholders in three different European countries and in West Africa. Across the Atlantic, the Stepping Stones in Science project has made connections with an international coalition on ‘Measurement of Play’ and one of our postdoctoral fellows, Dr Zhen Rao, is heading to Harvard next year for a lab visit with Professor Paul Harris to develop her work relating to children’s emotional development and pretend play (Rao & Gibson, 2019). Closer to home, we have active collaborations with colleagues at UCL, University of York and Imperial College London.

Given the constraints of space in the present article, we cannot give a detailed account of the many exciting projects underway in the PEDAL centre. Instead, to give a flavour of our activities, we here present an overview of our research relating to a single topic: measurement of play. We first provide a rationale for focusing on this topic before going on to illustrate how measurement development is integral to our research using examples drawn from across the PEDAL portfolio of projects.

Attempting valid and reliable measurement of play

Notwithstanding the wide disciplinary scope represented in the PEDAL team, it is a reasonable characterisation that most of our research falls within a (post)positivist research paradigm, drawing on elements of the scientific method. This includes observation and (quasi)experimentation in the service of generation, testing and refinement of theory (Ellefson, Baker, & Gibson, 2019). That is to say, our studies tend to assume that there is some objective reality about play and we try to adopt methods that reduce bias in our understanding of playful phenomena. A precise and valid measurement of play is a crucial part of this endeavour and given the notorious ambiguity of play as a construct, we have spent a large proportion of our time working on various aspects of the conceptualisation and implementation of different approaches to play and its measurement.

As for any approach to research, we are aware that there are limitations associated with our predominant approach. Some may view a focus on understanding the role of play in developmental or educational outcomes associated with play, alongside attempts at measurement, as promoting a limiting form of instrumentalism. This risks endorsing a view where play is only valued for its outcomes rather than for its intrinsic worth and its status as a recognised human right for all children (UN General Assembly, 1989). Further, there is also concern, particularly with respect to applications in educational research and ‘evidence-based practice’, that scientific methods can sometimes be undemocratic, ignoring the socio-political realities of research contexts and ‘subjects’ (Biesta 2010). To mitigate these potential risks, we aim to foster a culture of reflexivity in PEDAL whereby we reflect on our own perspectives, biases and assumptions and provide spaces where we can constructively challenge each other in our thinking about play. We hold a fortnightly reading group where we read and discuss work from across the breadth of research perspectives on play and early childhood, and we also make efforts to engage stakeholders (including parents, educators, and, children and young people) in the shaping of our research agenda and practice.

Returning to the topic of measurement, our quest for improving measurement of constructs relating to play stems from a desire to gain a deeper scientific understanding of this natural phenomenon – why do we play? What forms can play take? How does play function as a developmental mechanism? Is play in school really necessary? How does play change as children grow and develop? – and so on. In our view, measurement is an excellent first step towards answering these fundamental questions (Pellegrini, Symmons, & Hoch, 2004). Firstly, developing a good measurement tool requires researchers to clearly define the constructs to be measured. This is important given that some claims about play are so broad as to be meaningless. To paraphrase Peter K Smith (Smith, 2009, p. 197), some play researchers have had a tendency to wear ‘play goggles’ – i.e. having such a positive view of play that it is treated as a panacea and negative or null results are summarily reinterpreted or dismissed. Improving definition and precision helps to differentiate the putative effects of play from those of other closely related developmental processes such as social cognition or attachment. A clear specification of the constructs being measured also allows other researchers to critically evaluate, replicate and build on published work, for example, it would be difficult to accurately appraise the implications of a study using a measure of pretend play that does not report whether the measure used captured non-verbal as well as verbal communication during play.

Furthermore, having measures that have not only been assessed for validity but also calibrated for reliability is a strategy for reducing some sources of bias and unwanted variability (i.e. measurement error, Rust & Golombok, 2014). For instance, in a study attempting differentiation between physical aggression and rough and tumble play it would be important that observational results were not the result of one individual’s interpretation. In our experience, such inter-observer calibration can take a number of weeks, several discussions and lots of practice to achieve. Another important consideration is the temporal stability of measurement – if a PEDAL researcher measured your ‘playfulness’ on a Monday morning would they get the same result as measuring it on a Saturday night out? We have been grappling with these types of questions and more in our endeavours!

Having set out our general approach to research and measurement, we now turn to a discussion of examples of play measurement development in various studies being carried out in the PEDAL centre. You will see that the ways in which this fundamental topic is addressed vary according to the purpose of each study – from understanding playful pedagogy, through measuring intervention outcomes, to understanding more about how play works as a developmental mechanism.

Measuring children’s playfulness and play with their peers and friends

The Social Play, Social Lives research stream aims to learn more about the role of play in supporting children’s social development, particularly in relation to communication skills, peer relations and friendships in the primary school years (ages 4;0–11;00 years). Finding ways to capture the complexity of play in middle childhood has formed a large part of this research programme. At the core of this programme is the Children’s Relationships with Peers through Play (ChiRPP) study (Gibson & Fink, 2019). The study was conducted in schools in the Cambridge area, was subject to ethical review and all participants had informed consent by the person with parental responsibility, additionally, each child assented to each activity in which they took part.

ChiRPP is a longitudinal study exploring play both as a concept in its own right and as a scenario for understanding the associations between children’s language skills, socio-cognitive understandings and social competencies. The study follows 150 children over three data collection timepoints; from Reception class (Time One), through Year 1 (Time Two), to Year 2 (Time Three) of UK primary school education. An additional 90 children were recruited at Time One to support the development and testing of new measures.

The study adopts a multi-method and multi-informant approach. That is, for each one of the main areas of child development of interest to the study (i.e. verbal and cognitive ability, social cognition, social competence and play), more than one type of research method is used and more than one informant is involved. For instance, play is measured by teachers’ ratings of children’s play, a peer-reported playfulness scale, a self-reported playfulness scale and dyadic play observations. This multifaceted approach is a deliberate attempt to reflect different ways in which play may be conceptualised – from a special context for bringing together and refining social skills, through to ideas of intrinsic ‘playfulness’ akin to a personality or temperamental trait.

Related to the latter, the child self-reported playfulness scale (CSRP) is a newly developed measure created as part of the ChiRPP study. It is based on the self-concept interview method by Eder (1990) and Harter (1982, 1985) via ‘puppet-show’ and aims to explore children’s own perceptions of themselves as having a playful disposition. The puppet show method involves asking children to point to one of two puppets who make opposing statements, for instance, ‘I tell funny stories’ and ‘I don’t tell funny stories’, based on the puppet they believe they are most like. This method has previously been used to support children to report on complex constructs like self-esteem. The CSRP has been tested and retested on a sample of 98 children aged between 5 and 7 years old within a 6-week time span, and on a sample of 244 children over a 1-year span. Interestingly, children’s self-reports of their own playfulness were found to have good internal consistency, temporal stability and to demonstrates no differential item functioning across gender, age and language ability. Further, the scores correlated positively with a teacher rated measure of child play skills, but not a measure of general social maturity. This suggests that at least some aspects of playfulness can be considered distinct from general social competence (Fink, Mareva, & Gibson, 2019). Thus, the CSRP has promising initial indications of reliability and validity as a child self-reported measure of play.

Observational methods also form part of the ChiRPP study repertoire. Pretend play with a peer was an activity of particular interest to the research team, given the claims that have been made for its potential role in children’s development. Motivated by the desire to use a rigorous, longitudinal method that would enable the study of the changes in these play behaviours and how they relate to social skills over time, the team conducted observations of pairs of children engaged in pretend play in all three timepoints of the study.

Briefly, children were paired with a preferred classmate and left in a quiet space to play with a Playmobil zoo or castle (toy selection was randomised across study timepoints). These play objects have been used in previous research as a means to stimulate play and social interactions (Ensor & Hughes, 2005; Kuhnert, Begeer, Fink, & de Rosnay, 2017). Video recordings were made of the interactions (around 7–8 min long) and these were subsequently annotated by researchers to ‘code’ for four observable dimensions of children’s peer social pretend play, based on the existing literature on this topic:

• Calls for attention is coded when there are explicit verbal and non-verbal cues of one child trying to attract the attention of their play partner. For example, pointing to an object or saying ‘look!’, ‘here!’, ‘see?’

• Joint proposals is coded when there is reference to a specific action that facilitates the development of pretend play plot. For example, ‘you have to stay in my arms’ and ‘let’s go under the umbrella in case someone squirts us’.

• Role assignment is coded when children verbally assign a pretend role to themselves and their play partner, e.g. ‘I am the baby and you be mummy’, ‘I am the person who bought this castle’.

• Pretend play enactment is coded based on the actions, emotions, physical gestures and tone of voice during pretence. For example, ‘fire! I am shooting the big lion’ while making shooting noises.

A 15% sample of all codes was independently coded by trained researchers, reliability was found to be high (Cohen’s κ min = .79, max = .83).

The next step was to use these codes in planned analyses to illuminate the nature of relations between social pretend play behaviours and other aspects of development. To date, the team has focused on analysing data from the Time One data collection wave. The most recent analysis shows that variance in children’s peer pretend play behaviours is primarily influenced by dyadic effects (i.e. the particular pairing of two children), rather than individual child characteristics such as gender and language ability (Gibson, Fink, Torres, Browne, & Mareva, 2019). To the best of our knowledge, no previous studies have accounted for dyadic level variability in peer social pretend play behaviours (although interestingly such methods are common for parent–child play), and we suggest that failure to account for such effects may have contributed to lack of consistent findings in the field. The next steps planned for this measure will be to investigate its concurrent and longitudinal associations with independent measures of social development, friendship and social maturity.

To conclude this section, as the first longitudinal study designed specifically to measure social play and its relation to 4–6 year-old children’s social development using multi-method, multi-informant approaches and based on a replicable measurement strategy, we believe that ChiRPP has the potential to make a significant contribution to play research.

The long-term aim is that this study informs future research on the importance of considering play dynamics and interactional contexts. We now turn to discuss some examples of measurement development relating to our work on parent–child play.

Measuring parental playfulness and parent–child play

Parent–child relationships are key during the early years of an individual’s life. In the early years, parents are typically the most significant influencers of many aspects of their children’s development, including cognitive (e.g. Ryan, Martin, & Brooks-Gunn, 2006; Sethna et al., 2017; Shannon, Tamis-LeMonda, London, & Cabrera, 2002), social (e.g. MacDonald & Parke, 1984), emotional (e.g. Hubbs-Tait et al., 1996), and language development (e.g. Tamis-LeMonda, Shannon, Cabrera, & Lamb, 2004). Play is an essential and often core part of these interactions right from the start of life. When children are very young it can often be that parents lead much of the play, though even in a child’s earliest days a parent’s ability to follow their child’s lead and respond sensitively is a key part of successful parenting. As children grow and develop they can take the lead in play activities more, but parents can continue to influence children’s play through shared activities (Göncü, Mistry, & Mosier, 2000; LaForett & Mendez, 2017; Muhonen, von Suchodoletz, Doering, & Kärtner, 2019). These shared play experiences mean that parents can structure and scaffold their child’s exploration and learning, but also continue to develop and foster their relationship (Ginsburg, 2007).

Play provides both children and parents with a unique and individualised medium to express positive affect. The contribution of playfulness in these interactions, over and above other important aspects of interaction, such as sensitivity and warmth, has not been clearly delineated in research. It has been demonstrated that playful interactions allow parents to engage with their children in a meaningful way, and with fewer constraints (Ginsburg, 2007) and that parental playfulness is linked with positive outcomes for children, with higher levels of playfulness linked to lower child negativity (Menashe-Grinberg & Atzaba-Poria, 2017). However, accurately assessing parental playfulness and assessing its relative importance in children’s lives, remains a significant research challenge.

One key component of this challenge is that of measurement. Many of the key elements of measurement outlined earlier in this article apply, such as finding valid and reliable measurements and considering the different perspectives of different actors in the play. In addition, there are definitional difficulties. In investigating parent–child play with very young children, the child is not always leading the play (an important component of some definitions of play) and distinguishing playful activity from a playful countenance or affect can be a significant difficulty. Fisher, Hirsh-Pasek, Golinkoff, and Gryfe (2008) summarise some of the approaches that theorists have taken in their attempts to define play: while some identify play by the characteristics of the activity such as the lack of rules, the need for a structure, or involving social interaction, others define it by the personal attributes such as intrinsic motivation and positive affect.

Breaking down the construct of play can make it easier to focus on specific elements of play. As Cabrera, Karberg, Malin, and Aldoney (2017) point out, much of the current literature on play and its role in child development has been based on broad measures of play that do not provide clarity on what type of play or affect is most important for a particular aspect of children’s development (Elias & Berk, 2002; Lillard, 2012). Themes that recur in some of the definitions of parent–child play include the need for the activity or behaviour to have a sense of fun, enjoyment, engagement, and spontaneity. Göncü et al. (2000) suggest that ‘fun’ can be identified by looking at the play partner’s face and focusing on smiles or laughter. Hirsh-Pasek and Golinkoff (2008) include concepts of engagement or being engrossed. These facets also need to be considered in the specific context of assessment to ensure that it is play or playfulness that is being measured, but also that elements of play are not excluded by the definition or form of measurement.

At PEDAL, the Play in Early Life team have been working to understand how play develops in the early years, and how it is related to other aspects of child development. The first steps in this have been a study led by Dr Marisol Basilio (ESRC Fellow). The aim of this study was to create and assess the validity and reliability a new measure of parental playfulness. Using data from the New FAM study, directed by Professor Claire Hughes, this project involved approximately 200 first-time parents. Parental playfulness with their children was assessed at three timepoints early in the children’s lives, at age 4, 14, and 24 months. This was based on independent coding of videos of parent–child play at each time point. Both mothers and fathers were included.

A new scale for parental playfulness was developed (Basilio, Laverty, & Whitebread, 2018), which focuses on multiple aspects of the play interaction. It first measures the type of behaviour being observed; this is done by breaking the interaction down into segments of one minute each. The parent’s behaviour is then systematically assigned to a specific category of play (e.g. physical or pretend play; based on types of play devised by Whitebread, Basilio, Kuvalja, and Verma 2012). All the possible different behaviours are identified based on the type of activity the dyad is engaging in during the course of the observation. It also captures some aspects of the quality of the interaction such as the intensity of play, rigidity, and overall child affect.

Following this detailed coding, an overall global score is assigned for the level of playfulness exhibited by the parent during the whole interaction. This rating is based on the parent’s use of cues to the child, the variety of play categories, and the energy level, among other things. It is a 9-point scale where a score of 0 suggests behaviour that is ‘not playful at all’ and a score of 9 indicating behaviour that is ‘highly playful’.

This scale is, therefore, capturing several aspects of the play interaction – the nature of the activity, the level of engagement, as well as energy and complexity of play. It provides a detailed account of the play interaction, allowing for a zoomed-in record of the behaviour, as well as a more high-level indication of the overall nature of playfulness in that interaction.

To expand on the assessment of parental playfulness the Play in Early Life team is now using this measure in other developmental cohorts including the Healthy Start, Happy Start study in London and the Oxford Fathers Project. These are being used to both investigate and compare this measure of parental playfulness with others (such as the Playfulness scale developed by Cabrera et al., 2017), but also to expand the study of playfulness from assessments of individual parents with children, to settings which include two parents or carers in interaction with their child. We are also developing comparative studies to assess the contribution of playfulness compared to other better studied and understood aspects of the parent–child relationship, such as warmth and engagement. In this way we can better understand the place and contribution of parental playfulness to key aspects of children’s development and the development of parent–child relationships.

The Healthy Start, Happy Start (HSHS) study is a randomised controlled trial, which is funded by the National Institute of Health Research (NIHR) to test the effectiveness and cost-effectiveness of a brief psychological intervention for parents and children (Ryan, O’Farrelly, & Ramchandani, 2017), called Video Feedback Intervention to Promote Positive Parenting and Sensitive Discipline (VIPP-SD; Juffer, Bakermans-Kranenburg, & Van Ijzendoorn, 2017). The intervention is being tested to see if can prevent enduring behavioural problems in young children (aged 1 and 2 years old) (Ramchandani et al., 2017). At Three hundred families are participating, with half the families randomly assigned to the control group (treatment as usual), while the other half received the VIPP-SD intervention.

VIPP-SD draws on attachment theory in its promotion of sensitive parenting to strengthen the relationship between child and caregiver and also social learning/coercion theory to provide parents with consistent ways of responding to negative behaviour in their children. It uses play-based sessions with parents and children, which are then video-recorded and used in subsequent sessions to help parents understand their children’s behaviour and communication and to consider different ways of responding within play-based and other settings. This study gives us an opportunity to consider parent–child play over a 2 year period and as children are recruited when aged 1 or 2 years old, effectively gives a developmental window between ages 1 and 5 years to study the development of parent–child play, in mothers, fathers and in play settings including both parents or carers with their child together. As other robust measures of child behaviour and executive functioning and of family and parental well-being are also included this will enable us to better understand the developing role and importance of parent–child play early in children’s lives.

In one further note on future work, we are also building on the team’s expertise in early intervention to develop the use of video-feedback approaches with teachers in reception school settings. This work is in its earliest stages, but we plan to incorporate learning about early child play from video-feedback, alongside learning from the guided play field, which combines sensitive teacher instruction with respect for children’s autonomy and exploration (Weisberg, Hirsh-Pasek, Golinkoff, Kittredge, & Klahr, 2016) to optimise children’s opportunities for learning and socio-emotional development at the start of school.

Conclusion

We hope that you have enjoyed this whistle-stop tour of the PEDAL research centre and found something of interest in our work on play measurement. We plan in future reports to provide more detail on different aspects of our work, including the studies of playful pedagogy for Primary School science and our burgeoning research theme on play and health. We are delighted to join the diverse research community represented by the contributors to and readership of the International Journal of Play, and we look forward to engaging discussions and collaborations with this community in the near future.

Acknowledgements

The PEDAL Centre receives funding from the LEGO Foundation through a donation and a separate grant related to autism and play. It also receives funding from the National Institute for Health Research (NIHR) Health Technology Assessment (HTA) programme for leading and other involvement in randomised controlled trials. The ESRC has funded three postdoctoral fellows (Dr Marisol Basilio, Dr Dave Neale and Dr Zhen Rao) who are part of the PEDAL team.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes on contributors

Vicky Yiran Zhao is Ph.D. a student at the PEDAL Research Centre in Cambridge.

Krishna Kulkarni is a Ph.D. student at the PEDAL Research Centre in Cambridge.

Jenny Gibson is a Senior Lecturer in Psychology and Education at the Faculty of Education, University of Cambridge and leads research on children's social play and social development at the PEDAL Research Centre.

Sara Baker is a Senior Lecturer in Psychology and Education at the Faculty of Education, University of Cambridge and leads research on cognitive science and children's learning at the PEDAL Research Centre.

Paul G. Ramchandani is LEGO Professor of Play in Education Development and Learning at the Faculty of Education, University of Cambridge and is Director of the PEDAL Research Centre.