Structures, processes and outcomes in preschool units with high and low proportions of second language learners of Swedish : A Comparative Study

(1)

Structures, processes and outcomes in preschool

units with high and low proportions of second

language learners of Swedish

A Comparative Study

Eleftheria Beteinaki

L2XS28:

Two-year Master Degree Project in Child Studies

Program: Interventions in Childhood

Specialization: Welfare and Social Sciences

Supervisors: Mats Granlund and Madeleine Sjöman

Examiner: Will Farr

(2)

2

SCHOOL OF EDUCATION AND COMMUNICATION (HLK) Jönköping University

Master Thesis 15 credits Interventions in Childhood Spring Semester 2020

ABSTRACT

Author: Eleftheria Beteinaki

Main title: Structures, processes and outcomes in preschool units with high and low proportions of second language learners of Swedish

Subtitle: A Comparative Study

Pages: 34

Second language learners (SLLs) are students at risk for negative outcomes in preschool. Theoretical statements from systems theory and the ecological model of inclusive education indicate that a main focus of interest when supporting children should be the preschool environment that surrounds them. The purpose of this study is to describe and compare structures and processes in preschool settings that are presupposed by theory to have an impact on children’s functioning, along with children’s outcomes in preschool units with different proportions of SLLs. A combination of questionnaire data and data from observations were used in order to describe and compare the structure of child-teacher ratio, the process-related variables of teachers’ tone, children’s proximity and verbal interaction and lastly the outcomes of involvement and social interactions in different play types between the units. Data were analyzed with independent sample t-tests and the results revealed that units with different proportions of SLLs differ regarding processes and outcomes but not the structure of child-teacher ratio. Teachers’ tone in units with low proportion of SLLs was more positive and children were more verbal to the teacher while children in units with high proportions of SLL, had more self-talk/ private speech, less social-interactions and associative play, and lower levels of involvement. The differentiations in processes and children outcomes, along with the non-differences in child-teacher ratios between units with high and low proportions of SLLs highlight the need for further research in the field in order to investigate which environmental factors affect children outcomes in attempts to intervene and safeguard inclusiveness.

Keywords: Second language learners, Sweden, preschool, child-teacher ratio, proximity, verbal interaction, teacher’s tone, children outcomes, involvement, social interaction, types of play, questionnaires, observations.

Postal address Högskolan för lärande och kommunikation (HLK) Box 1026 551 11 JÖNKÖPING Street address Gjuterigatan 5 Telephone 036–101000 Fax 0361625 85

(3)

3

1. Introduction

The global agenda guided by the Sustainable Development Goals for 2030 states that in order for the world to become a better place, the provision of lifelong education characterized by inclusiveness, equity and quality for all is crucial (UNESCO, 2018). Inclusiveness and quality in education refer to the process of meeting the needs of all children and especially the ones belonging to the vulnerable groups due to disability, special needs and a culturally or linguistically different background (Bartolo, Björck-Åkesson, Giné, & Kyriazopoulou, 2016; Kyriazopoulou, Bartolo, Björck-Åkesson, Giné, & Bellour, 2017; UNESCO, 2017).

Second language learners (SLLs) are considered a vulnerable group of students, with high risk of poor academic, social and psychological outcomes, because of family-related factors along with the fact that their native language and potentially their culture, are different from the major language and culture in their school environment (Halle, Hair, Wandner, McNamara, & Chien, 2012; Winsler et al., 2014; Jerome, 2009; OECD, 2016). These reports of poor outcomes are in contrast with the vision of equity that highlights that every child’s personality, talents and abilities must be developed to the fullest in education settings (UNCRC, 1989; Kyriazopoulou et al., 2017). For this reason and even though most countries make efforts to develop and implement inclusive policies, achieving full inclusion in education and ensuring equal opportunities for all learners, still remains a challenge (UNESCO, 2017).

Many studies attempts to find causes and prevent negative outcomes in students’ functioning and development by exploring and targeting individual factors on the child level such as gender, age, special needs or the different mother tongue (Gustafson et al., 2017; Gustafson et al., 2018; Engvik et al., 2014). Another branch of research also focuses on how factors in the environment affect child outcomes (Sandberg et al., 2010; Mashburn et al., 2008; Brodin & Renblad, 2015; Coelho, Cadima, & Pinto, 2019; Prykanowski, Martinez, Reichow, Conroy, & Huang, 2018; Sjöman, Granlund, & Almqvist, 2016; Åström, Björck-Åkesson, Sjöman, & Granlund, 2020; Sjöman, 2019a). However, a main value in documents developed to enhance inclusive policies and quality education is to recognize and consider that aspects of the educational system, including processes and structures (within and out of the preschool setting) are interrelated with child characteristics and have major effects on children’s outcomes (Kyriazopoulou et al., 2017; UNESCO, 2017). In other words, there is a need for research which examines the interrelation of individual factors with the environmental factors surrounding the context for specific subgroups of children, such as the SLLs (Domínguez, Vitiello, Fuccillo, Greenfield, & Bulotsky-Shearer, 2011) in all education levels starting from

(6)

6

the preschool sector which has major effects in children’s learning and holistic development (Kyriazopoulou et al., 2017; UNESCO, 2017) .

2. Background

2.1 SLLs in Sweden

The national curriculum for Swedish preschools recognizes that the variations in children’s conditions and characteristics create different requirements in the preschool environments in order to support those children’s functioning optimally. These differences need to be taken into account and thus, preschool education in Sweden presents variations in its structure dependent on the children served (Swedish National Agency for Education, 2019). Since the number of children in North Europe who live and grow up in ethnically diverse contexts is increasing (Aukrust & Rydland, 2009), children being SLLs is one of those circumstances that create special conditions in Swedish preschools (Swedish National Agency for Education, 2019). Reports highlight that 25% of the children under the age of five in Sweden are SLLs and preschool is the place where they will eventually learn and develop the Swedish language (Puskás, 2017). The educational system in Sweden highlights the importance of providing support for the development of every child in need and children being SLLs of Swedish are considered children in need of special support (Eurydice, 2018).

In Sweden SLLs can be divided in two groups. The first group consists of students from officially recognized language minorities in Sweden which includes Sami, Tornedal Finnish, Finnish, Romany and Yiddish (Noorani & Baïdak, 2019; Eurydice, 2018a). The second group is the one of the newly arrived students who do not speak Swedish. This is a more heterogeneous group as families might have migrated to Sweden from several countries and for several reasons and are categorized in subgroups according to a legal framework (Nilsson & Bunar, 2016; Eurydice, 2018a).

Measures are being taken in preschools in order to support children who are SLLs. A significant contribution is the opportunity for SLLs to develop their mother tongue along with the Swedish language. However, this opportunity is not always provided for the group of newly arrived students when there is a shortage of equivalent teachers or not enough students in the same classroom from the specific language group (Eurydice, 2018a). In a more general sense, the way that the Swedish educational system responds to the needs of SLLs, who are not coming from the officially recognized language minorities, displays variations as there are no

(7)

7

national regulations on how to integrate them and how to organize and conduct their education (Nilsson & Bunar, 2016). In addition, in Sweden especially the subcategory of SLLs that come from parents who are immigrant workers has not been academically researched thoroughly and therefore schools do not receive additional governmental support for them as it happens for instance in schools with refugee children or asylum seekers (Nilsson & Bunar, 2016).

In order to investigate the development and functioning of children who are SLLs, a multilevel framework is needed (Domínguez et al., 2011). Such a framework can generate the knowledge that will help modify educational environments to enhance the academic and socio-emotional outcomes for these children.

2.2 Theoretical approach to the issue of SLLs

The theoretical presuppositions in this study arise from the Systems Theory (Wachs, 2000) and the Ecological model of human development (Bronfenbrenner & Morris, 2006) that highlight the effect of preschool environmental factors on SLLs outcomes. Those theories support that basic complex phenomena such as children’s behavior, learning and functioning consist of elements from different domains around the child that are interconnected and interrelated. In attempts to understand such complex phenomena it is necessary not only to identify their elements but also investigate their interactions (Wachs, 2000).

2.2.1 School environmental factors and SLLs functioning

The environment of every school is rather complex and consists of physical, social and policy-related aspects that can eventually affect the participation and functioning of individual children (Castro & Palikara, 2018). Therefore, to achieve the desired outcomes, instead of focusing solely on interventions at the child level, additional steps need to be taken in order to change and improve the surrounding environment and its effect.

As mentioned earlier, SLLs are exposed to multiple risk factors that can have a cumulative effect and are much likely to lead to dysfunction in childhood (Carneiro, Dias, & Soares, 2016). Dysfunction or elsewise lack of functioning could be perceived as non-positive “aspects of the interaction between an individual and that individual’s contextual factors” (WHO, 2007, p.228).

Contemporary literature supports the aforementioned statement of dysfunction of SLLs as they are often reported to face academic and behavioral problems in the mainstream schooling system (Hoover & Patton, 2005). Previous studies focused on the differences of native speakers

(8)

8

and SLLs in school, reported that SLLs demonstrate more shyness and withdrawal in contexts where they are required to speak another language than their native language (Ash, Rice, & Redmond, 2014). Castro-Olivo, Preciado, Sanford and Perry (2011) support this view and moreover report a positive relationship between socio-emotional resiliency and academic outcomes in SLLs. Additionally, a study by Halle et al. (2012) highlights that developmental trajectories in SLLs are dependent not only on cognitive factors but also on social and behavioral factors.

2.2.2 Preschool as a biofurcation point

An important aspect in the process of human development according to systems theory and more specifically to dynamic systems theory are the bifurcation points, which represent the transitions from one stage of functioning to another. One example of a major life change, in the microsystem level of the individual is to start preschool (Wachs, 2000). Preschool is the first experience of schooling and can be perceived as a transition point from the home environment to the context of the school. Moreover, for a child who is SLL the impact of this contextual shift is enhanced by the exposure to a different language and potentially a different culture (Wachs, 2000). All new influences and the presence of multiple risk factors are affecting the balance in the child’s system. The resulting instability can be seen as a time of potential behavior variability of children and depending on the resources they have and the demands of the new context it can lead to competence or dysfunction.

2.2.3 Preschool unit characteristics and provision of support for SLLs

Regarding the demands and resources, in the Swedish preschool context, as mentiond before the opportunity for additional support is not always guaranteed for the group of SLLs. If they do not belong to the official language minority groups and there is a shortage of equivalent teachers or not enough students from a specific language group, support might not be provided (Eurydice, 2018a). Since national regulations are absent, the inclusion methods for SLLs of each preschool unit in Sweden can vary (Eurydice, 2018a). Similarly, the demands of different units concerning SLLs vary i.e. some units expect SLLs to adapt to the new language and context without additional support, while other units perceive the presence of SLL as situations that demand adaptations and special arrangements, such as the decrease of the number of children in the unit (Nilsson & Bunar, 2016).

(9)

9

Specific unit characteristics of the preschool such as the proportion of SLLs affect the resources provided (Eurydice, 2018a) and potentially the teachers’ way of interacting with the children as well. In a study by Sandberg et al. (2010) it is stated that when there is a high proportion of children with special needs in the class, teachers will adapt their methods accordingly so that they meet their needs, something that is not as likely to happen when those children are a minority. Lastly, Granlund et al. (as ci in Sjoman, 2018b) showed that the type of activities in the preschool units was related to the proportion of SLL. Therefore, the proportion of SLLs in the preschool unit is a factor that needs to be investigated even more in relation to children’s outcomes and the way it can affect the overall environment of the classroom.

2.2.4 Proximal processes and preschool

One major aspect that leads to changes and can explain individual behavioral and developmental variability (Bronfenbrenner & Evans, 2000; Wachs, 2000; Bronfenbrenner & Morris, 2006) are the proximal processes. Proximal processes are multiple influences and interactions between the person and its immediate external environment. Accordingly for children who are SLLs, those complex interactions and interrelations with the proximal systems, one of which is the preschool, are influencing their development and functioning (Kyriazopoulou et al., 2017; UNESCO, 2017). Research shows that interaction processes with teacher and peers in the preschool setting “have the greatest impact on the quality of children’s experiences and outcomes” in school (Kyriazopoulou et al., 2017, p.18).

The proximal processes can vary for several reasons, for example the child’s characteristics, the environment where the processes take place, the nature of the expected outcomes and the changes that take place over time (Bronfenbrenner & Morris, 2006). Regarding SLLs in preschool, the fact that they do not speak the language of the teacher and the class can affect their interactions with the teachers and peers especially if they do not receive any type of support such as mother tongue support and activity adaptations.

2.2.5 Specific aspects/ processes in preschool context

In order to specify the aforementioned theoretical presuppositions and in an attempt to provide a connection between theory and practice regarding specific aspects and processes in the environment and their effect on inclusiveness and children outcomes in school, the European Agency for Special needs and Inclusive education contributed by introducing the ecological model of inclusive early childhood education (Kyriazopoulou et al., 2017). According to this

(10)

10

model there are specific processes that are indicators of proximal processes in preschool with involvement in learning and social activities to be considered one among the most important ones (Sjöman, 2018). A more detailed description of the preschool unit processes according to the ecological model of inclusive early childhood education includes:

1) positive interactions with adults and peers, 2) involvement in play and other daily activities, 3) child-centered learning,

4) personalized assessment for learning and 5) accommodations, adaptations and support.

(Kyriazopoulou et al., 2017, p.18)

Those processes in the preschool sector will be elaborated further in order for specific factors of the preschool environment to be highlighted when discussing inclusiveness in cases of SLLs in Sweden.

2.2.5.1 Positive interactions : verbal interactions, proximity and teacher’s tone

Firstly, positive interactions with adults and peers in school is a desired outcome of preschool itself but in cases of SLLs these type of interactions are also reported to facilitate children’s language acquisition and therefore lead to positive results in other domains as well (Puskás, 2017). Those positive interactions are indicated by teachers’ and peers’ responsiveness towards the child, the proximity, the verbal interactions and teachers’ tone while interacting with children (Mashburn et al., 2008).

Proximity is the degree of closeness and can indicate a positive social interaction between the children and the other persons with whom the child is in proximity, whether those are peers or teachers (Mashburn et al., 2008). Research reports are divergent since proximity to the teacher is reported to lead to less positive peer interaction, however, the teacher’s presence was

associated with positive changes in children’s interactions with their peers (Acar, Hong, & Wu, 2017).

Verbal interactions are crucial for children’s socio-emotional and cognitive development in general and specifically the verbal interactions between peers improve children’s interactive play behaviors (Stanton-Chapman, Walker, & Jamison; 2014). Second language acquisition takes place through social interactions of the language learners with native speakers, and according to Chapman (2000) a positive verbal interaction with the teacher and peers who are native speakers can lead to language development for SLLs. Thus, it is likely that quality

(11)

11

language interactions between teacher and students can result in improved vocabulary both for SLLs and native speakers.

Additionally, emotion is a mediator in the process of children’s learning and acquiring knowledge and skills (Park, 2014). The variable of teacher’s tone, which indicates the degree of their positive interactions and affect (Spivak & Farran, 2016), is an indicator of teachers’ sincere attempts to respond to children’s communication (Puskás, 2017;Mashburn et al., 2008). Research has shown that teachers’ positive emotional tone and approval responses can predict some children’s outcomes such as positive social interactions and can help in gaining cognitive self-regulation skills (Fuhs, Farran, & Nesbitt, 2013; Spivak & Farran, 2016). When teachers are responsive and the interactions with peers are positive, children’s engagement is positively influenced (Sjöman, Granlund, & Almqvist, 2016). Moreover, when preschool teachers sincerely try to support and respond to children’s communication and needs, they have a significant impact on children’s belongingness, engagement and learning in the preschool class (Kyriazopoulou et al., 2017; Puskás, 2017).

2.2.5.2 Involvementand social interactions through play

As mentioned earlier proximal processes are crucial for child development, learning and participation. Since participation is defined as involvement in real life situations (WHO, 2007), involvement is a construct that reflects those proximal processes (Imms et al., 2017; Bronfenbrenner & Morris, 2006). Child-centered approaches, inclusion and equity in preschool education are principles adopted by the Swedish curriculum from preschool (Swedish National Agency for Education, 2019) but no outcome indicators for children are specified. Implicitly, in the absence of systematic assessments, learning outcomes are evaluated by the state of involvement of students in preschool classes (Swedish National Agency for Education, 2019; European Agency for Special Needs and Inclusive Education, 2020; Sjöman, 2018).

This focus on involvement concerns all daily activities, even the ones that might not be considered as academic or social activities but are necessary for children’s’ development and classroom’s functioning (Farran, Kang, & Plummer, 2003). Involvement in daily activities also includes involvement in social interactions with adults and peers, which requires that opportunities for social types of play are taken into consideration (Kyriazopoulou et al., 2017). Parten (1932) in order to describe children’s level of social participation, categorized play in different types, and each type indicates a different type of social interaction, where the more social the type of play, the more beneficial for children’s outcomes. Regarding SLLs, since not all types of play require social interaction between children, it is assumed that not all types of

(12)

12

play are optimal for learning the second language (Piker, 2013). Associative and cooperative play are the types of play that involve interactions with other peers and thus enhance the higher thinking as well as language (Farran & Son-Yarbrough, 2001).

2.2.5.3. Accomodations, adaptations and support

As mentioned before, research has repeatedly shown the effect of environmental factors on children’s outcomes. Regarding environmental accommodations, adaptations and support for SLLs, classroom characteristics such as small class sizes and small child-teacher ratios can create ecological conditions that impact social and instructional interactions and generate positive effects on children’s development and functioning (Mashburn et al., 2008). The variable of child-teacher ratio is an example of what is perceived as a good structural indicator of the inclusion methods in preschools. In Sweden there are reports that indicate that there is a reduction of child-teacher ratios in preschools where there are high demands of special support (Sandberg et al., 2010; Mashburn et al., 2008). Even though child-teacher ratios is one factor that is taken into consideration regarding classroom quality (Brodin & Renblad, 2015), reports about their effect on child outcomes such as peer interactions are inconsistent in research (Iluz, Adi-Japha, & Klein, 2016). There are reports showing that the ratios are associated with students’ social interaction skills (Meng, 2019; Iluz et al., 2016; Brodin & Renblad, 2015). However the findings are contradictory on whether high or low child-teacher ratios promote better peer interactions (Iluz et al., 2016).

The conclusions of the presuppositions arised from the theortical frameworks used in this study are illustrated in Figure 1. The specific processes and factors of the preschool context in Sweden and their effects are presented according to the ecological model of inclusive education. The indicators of those processes and factors are displayed as well as they will be the main variables of interest of this study. More specifically indicator of structures is the child-teacher ratio. Indicators of processes are proximity and verbal interactios of the students along with the teacher’s emotional tone. Lastly the level of involvement of students is an indicators of the preschool outcomes along with the social interactions that are indicated by social talk and different types of play.

(13)

13 Figure 1.

Specific processes and aspects in preschool context and their effects along with indicators that will be used as variables of interest in this study.

2.4 Rationale

Previous research as well as theoretical presuppositions indicate that children who are SLLs constitute a group at risk for negative academic, social and other outcomes in preschool. A plethora of processes and factors in the preschool environment are highlighted by theory and models of inclusiveness to be interrelated and have an effect on SLLs outcomes. Earlier research on children with SLL refers to student populations in primary school and above (OECD, 2016). However, since preschool has a major influence on children’s later academic and social success (Watson, 2012), it is reasonable to be investigated further. Even though it is reported that children can display difficulties that affect their functioning in preschools in Sweden (Almqvist et al., 2018), to date there is a lack of research that compares Swedish preschool units with different proportions of SLLs in terms of structural- and process factors and outcomes in an attempt to evaluate their inclusiveness. As suggested by the Swedish preschool curriculum (Swedish National Agency for Education, 2019) and the guidelines of inclusive and equitable early childhood education (Kyriazopoulou et al., 2017), such factors

(14)

14

may have an impact on children’s outcomes in preschool (OECD, 2016); therefore they need to be investigated especially for the risk groups.

3. Aim of the study

In an attempt to provide a picture of the state of inclusiveness in the preschools, the aim of this study is to investigate whether aspects of classroom structures, observed processes and children’s outcomes vary between preschool units with high and low proportions of SLLs. Previous evaluations in preschools tended to focus on macro elements and general structural elements instead of outcomes (Kyriazopoulou et al., 2017, p.17). This study aims to compare two groups of preschool units categorized based on the structural feature of the proportion of children with SLL. The two groups will be compared on one specific structural element of the preschool setting (child-teacher ratio), and three specific process aspects: children’s proximity to others, children’s verbal interaction, and teacher’s emotional tone. These structural and process factors are in previous studies, international guidelines and reports, described to affect classroom quality (Kyriazopoulou et al., 2017; Mashburn et al., 2008; Rodriguez & Garza, 2014) and will be presented along with the comparisons of children’s average level of involvement, and the average frequency of social interactions and attendance in different types of play in units with high and low proportions of SLLs.

Therefore the research question is:

1. How do preschools with high and low proportions of SLLs differ concerning structures (child-teacher ratio), processes (children’s proximity, verbal interactions, and teacher’s emotional tone) and children outcomes (average level of involvement and average

frequency of social interactions and attendance in different types of play)?

4. Methodological framework

4.1 Design

The nature of the aim and research questions indicate a comparative research of the different preschool units, where the variables of interest are only observed and not manipulated, therefore it is a passive-observational study design (Kazdin, 2003).

For this study the variables of interest are structural and process-related aspects of the preschool units, along with variables on children’s outcomes in the unit level. The focus is on

(15)

15

studying preschool settings with high and low proportions of SLLs that vary in the above characteristics which are considered to be manipulated by “nature” and not the researcher (Kazdin, 2003).

Additionally, since the way that the preschools are organized and function may depend upon the proportion of SLLs (Nilsson & Bunar, 2016; Sandberg et al., 2010), it is crucial for the design to focus on results on unit level, thus, the preschool units need to be divided in groups based on their proportion of SLLs.

The analyses will be conducted cross-sectionally, the research focus will be the preschool units; the exposure will be differences in structures and processes of the preschool units and the outcome will be unit averages on levels of involvement and average frequency of social interactions and different types of play.

4.2 Procedure

Preschool units’ data from Early detection-Early intervention project (TUTI) and Participation and Engagement in Preschool International project (PEPI) of CHILD group in Jönköping University of Sweden were merged together and were the initial units of analysis for this study. Details about TUTI and PEPI projects, are described elsewhere (Jönköping University, 2019; Sjöman, 2018; Granlund, Almqvist, Gustafsson, Gustafsson, Golsäter, Proczkowska & Sjöman, 2015; Åström et al., 2020). Prerequisites for including preschool units from the aforementioned projects were that there was available information from all the instruments included in this study and that at least 25% of the students of the preschool unit were observed, in a way that the behavioral sample from children was representative of the unit.

4.3 Participants

A total of 68 Preschool units from TUTI and PEPI projects matched the inclusion criteria and thus were chosen for further analyses in this study.The descriptive statistics of the total number of preschool units are presented in Table 1.

(16)

16 Table 1

Descriptive Statistics of preschool units.

N Min Max M SD

Number of children in the unit 68 10.00 42.00 19.36 5.11 Children’s mean age in months 68 16.00 67.82 50.90 9.43 Percentage of girls in the unit 68 5% 79% 31.88% 16.56% Percentage of boys in the unit 68 21% 95% 68.12% 16.56% SLL in need of mother tongue support 68 .00 24.00 3.66 5.63 Child-teacher ratio 68 2.67 7.67 5.38 1.13 Percentage of SLL in the unit 68 0% 100% 19.25% 27.54% Valid N (listwise) 68

Since the aim of this study was to describe units with low and high proportions of SLLs the included units needed to be divided in the respective two groups. The cut-off point for inclusion of units in the aforementioned groups of SLL, was set at 25%. In a study by Sandberg and colleagues (2010) it was highlighted that when there is a relatively high proportion of children with special needs in the class, teachers will adapt their methods accordingly so that they meet their needs, something that is not as likely to happen when those children are a minority. This statement was taken into consideration along with the statistical report showing that in the preschool sector in Sweden, 25% of the children under the age of five in Sweden are SLLs (Puskás, 2017). Thus, for the present study, the percentage of 25% was used as representable of a high proportion of SLLs in the preschool unit that has potential to affect structures and teachers’ methods.

From the total number of included units, 72,1 % of them (n = 49) was classified as units with low proportion of SLLs which means that the percentage of SLL in the units was less than 25%. The remaining 27,9 % of units (n = 19) was classified as units with high proportion as the percentage of SLL was 25 % or more. The units with high and low proportion of SLLs did not have statistically significant differences regarding the average number of children in the unit, children’s mean age and the average number of boys. However, units with high proportions of SLLs had on average more girls (M = 6.79, SD = 1.55) compared to the units with low proportions of SLLs (M = 5.33, SD = 2.56); t(66) = -2.324, p = .023 (see Table 2).

(17)

17 Table 2

Differences between SLL groups in preschool unit characteristics.

Unit Character. SLL Group N M SD SEM T df p Number of children Low 49 19.55 5.61 0.80 .472 66 .639 High 19 18,89 3.67 0.84 Child age (months) Low 49 50.67 9.55 1.36 -.334 66 .739 High 19 51.52 9.37 2.15 Girls Low 49 5.33 2.56 0.37 -2.324 66 .023 High 19 6.79 1.55 0.36 Boys Low 49 14.22 6.20 0.89 1.343 66 .184 High 19 12.11 4.75 1.09 4.4 Instruments 4.4.1 Principal’s Questionnaire

For the descriptive characteristics of the preschool units, some of which form the preschool’s structural aspects of interest in this study (child-teacher ratio), a questionnaire was designed as part of TUTI project and was given to the principals of the schools. Those questionnaires also provided relevant information regarding characteristics of each unit, such as number of children in the unit, the mean age of students and the number of SLLs who are entitled to mother tongue support.

The number of SLLs who are entitled to mother tongue support in each unit was used to create the variable of high and low proportion of SLLs in the unit. This variable was used as an independent variable of high interest as previous reports highlighted that it affects the resources provided (i.e. mother tongue support) (Eurydice, 2018a) and it has a major role on whether the teachers will adapt their methods accordingly to meet the children’s needs or not (Sandberg, 2010).

(18)

18 4.4.2 COP and TOP

For the variables of preschool processes (proximity, verbal interactions, teacher’s tone) and children’s outcomes (involevement, social interactions and types of play) the Child

Observation in Preschool [COP] (Farran et al., 2003) and Teacher Observation in Preschool

[TOP] (Bilbrey, Vorhaus, Farran, & Shufelt, 2010) observation tools were used by independent observers for data to be extracted.

COP and TOP are structured observation systems for children’s and teachers’ behavior. The COP/TOP instruments give the opportunity to observe children’s and teachers’ behaviors, and at the same time provide information of how, where and with whom, children are spending their time in the preschool (Lillvist, 2010). Since the aim of this study focuses on structural aspects (child-teacher ratio), processes (proximity, verbal interaction, teacher’s tone) and outcomes (involvement and social interaction through different types of play), COP and TOP serve as ideal instruments for this study.

Both instruments are based on a series of three-second observations of behavior (sweeps) across a day-long visit, resulting in approximately 20-30 sweeps for each participant of this study. Momentary time sampling, like the three-second observations of COP and TOP, is proven to result in closer estimates of children’s engagement and problem behavior compared to other techniques (Prykanowski et al., 2018). The observers are processing and coding the observation data in a structured platform where information is provided for eleven dimensions for COP and ten dimensions for TOP (Farran et al., 2003; Bilbrey et al., 2010; Sjöman, 2019b).

Moreover, COP and TOP are standardized and adapted to the Swedish preschool context (Lillvist, 2010; Sjöman, 2019b). For the Swedish samples a specific adapted version of the observation tool was created (Sjöman, 2018a) and for this study a specific number of items (n = 4 from COP and n = 1 from TOP) was chosen to be further analyzed according to the study’s aim. Those items are presented below:

 COP:

1. Verbal/To Whom: captures whether the child is talking or not, and to whom.

This variable could be coded as being non-verbal and being verbal overall and the following codes existed depended on whom the child was being verbal to: to child, to teacher, to self,

(19)

19

2. Proximity: captures who is within a one meter radius of the target child indoor and three meter radius outdoors. The codes for this variable were the same as the codes concerning to whom the child was verbal to.

The variables of proximity and being verbal and to whom were considered important to be investigated as the degree of “closeness” and the verbal interactions are indicating the teachers’ and peers’ responsiveness towards the child (Mashburn et al., 2008).

3. Interaction: captures the social part of learning experience or in other words the type of engagement in play. There are nine different types of interaction coded in this instrument.

Unoccupied: not engaged even though there is an activity going on, Onlooker: just observing,

Alone: may be in proximity to others but not interacting with others in any way, Parallel: when children are using similar materials or engaging in a similar learning

activity, but each is working independently,

Associative: when children are sharing materials and interacting, but there is no

distinguishable goal for the group,

Cooperative: when children working together with some sort of shared goal, rules,

and/or organization,

Social (talk): when the child is interacting with peers or the teacher without this being

a part of a learning activity (i.e. in transition times, routines or centers),

Non-academic: when children are in transition times between activities or are engaged

in routines of the class but are not interacting with others,

Time out: when the child is isolated as a demand from the teacher.

The variable of interaction in this instrument is highly related to the different types of play, and it indicates different types of social interactions (Parten, 1932) which are the main focus in quality and inclusive early childhood education especially for the Swedish preschool (Swedish National Agency for Education, 2019; Kyriazopoulou et al., 2017). However, since the main focus of this study was the type of social interactions and not the specific types of play, the above categories were merged together to create two play groups named social interactions and non-social interactions. Associative, cooperative and social (talk) codes all together involved the category of social interactions thus, an index was created out of them. Similarly, parallel, alone and onlooker made up the category and the index of non-social

(20)

20

interactions. Time out, non-academic and unoccupied were not taken into account in this study as they were not coding forms of social interaction.

4. Involvement: captures how involved the children are in learning and social activities, i.e. engagement.

Codes for this variable vary from very low to very high level of involvement in a five graded rating scale but to obtain a high degree of inter-rater reliability they were merged to create only three levels, low/medium-low, medium, medium-high/high.

This variable was chosen because children’s level of engagement in various activities is seen as an indicator of preschool learning outcomes since in Swedish preschools learning outcomes are not formally assessed in any other way (Swedish National Agency for Education, 2019).

 TOP

5. Tone: captures teacher’s tone which indicates the general vibe of the classroom. It can be coded as vibrant when the teacher is interacting in a strongly positive way with the children by laughing, smiling and being excited; pleasant when the teacher is not verbal but displaying interest towards the child; flat when the teacher is neutral and indicates nothing; negative when the teacher’s expressions and actions display that she is not pleased or she is annoyed and extreme negative when teacher’s actions and words are aggressive and/or insulting.

The variable of teacher’s tone was chosen to be investigated as it is an indicator of teachers’ sincere attempts to respond to children’s communication (Puskás, 2017; Mashburn et al., 2008).

4.5 Data Analysis

The data from the instruments described above were already collected, saved in a computerized format and were statistically coded as a part of the TUTI and PEPI projects of the CHILD group in Jönköping University (2019). The nature of the aim of this study along with the research questions and the way the data were collected indicated a quantitative research approach which is instrument-based (Creswell, 2014). In order to answer the established research question, collected data were analyzed using IBM SPSS 21.

(21)

21

Two of the first and most crucial steps of this study were to merge data from the two projects and to aggregate individual data of the children from COP/TOP across multiple observation cycles and multiple children in order to provide information about the preschool unit.

The goal of the data analysis was to report potential differences in child-teacher ratios, proximity, verbal interactions and child outcomes regarding frequency of involvement, social interactions and attendance in different play types in preschool units with high and low proportion of SLLs. Descriptive statistics were computed and between-group differences were analyzed by using independent sample t-tests in order to highlight how much of the variance of the variables of interest could be attributable to the different proportions of SLLs in the unit. 4.6 Reliability and Validity

In an attempt to avoid internal attrition because of lack of parental consent, all participants in this study were observed anonymously, without registering personal data. Information about the units mean age, proportion of gender and proportion of SLL was provided by schools’ principals. Since there is a high probability that not all students of the unit were present in the day of the observations, the percentages of SLLs might differ. This threat to validity was attempted to be solved by having a prerequisite of at least 25% of the students of each preschool unit to be observed, so that sample is representative of the unit.

Moving to construct validity, the COP/ TOP instruments are standardized and adapted to the Swedish preschool context (Lillvist, 2010). However, even if the instruments are standardized and validated, a possibility of potential cofounding variables that could interfere with the results still exists. Thus, the statements about causal relations should be treated carefully (Shadish, Cook, & Campbell, 2002). In other words, potential differences in structures, processes and outcomes in different preschool units cannot be completely attributable to the proportion of SLLs as other factors that were not observed nor manipulated might have had an effect.

For the statistical conclusion to be valid, there is always a chance of type I error (α) and low statistical power, especially in this study where many individual tests were carried out (Shadish, Cook, & Campbell, 2002). On the other hand, since this project is mainly a descriptive and comparative study and it does not focus on measuring effects, or report relations the results can still be discussed.

The COP/TOP observations are based on one day assessments in random time-points. These time measurements could be a threat to external validity as there is a possibility that the results would vary if the observations and ratings took place in different time points of the day or the

(22)

22

academic year (e.g. in the beginning of the school year versus middle or end) (Cook & Campbell, 1979). However, information regarding the time points of the observations were not available in this study therefore it is unknown if systematic differences between the groups regarding the time they were observed exist.

Lastly, this study used data based on observations to define aspects regarding processes and outcomes such as social interaction and level of involvement. Those aspects are difficult to capture with complete accuracy especially when the children cannot confirm the interpretations afterwards (Nilsson et al., 2015). In order to enhance the quality of this study it would be useful to include children’s perspective especially for the interpretation of data. If the age of children does not allow that, then it might be useful for triangulation to use teacher’s interpretations as well.

4.7 Ethical considerations

Regarding the ethical considerations, this study will use data that are already collected from previous projects, therefore approval from the Regional Ethical Review Board in Linköping was already obtained (Reference No 2012/199-31 and 2014/479-31). Similarly obtained was the informed consent of parents of the children who participated in the study. The use of the consent was not extended beyond the original projects. Additionally, this study was based on naturalistic observations, it did not cause harm and no unauthorized person will be able to identify participants. The researchers invaded as little as possible on participants’ space following the guidelines of the American Psychological Association (as ci by Kazdin, 2003). For this reason and since only unit level analyses were made, even though some participants did not provide active consent (in the TUTI project), they were observed without their personal information being linked to observations. This procedure safeguarded personal integrity and confidentiality and at the same time internal attrition was avoided as there were not too many drop-out cases (Kazdin, 2003).

5. Results

A total number of 68 preschool units was the sample of analysis for this study. From the total number of included units, n = 49 (72,1%) were classified as units with low proportion of SLLs which means that the percentage of SLL in the units was less than 25%. The remaining 19 units (27,9%) were classified as units with high proportion as the percentage of SLL was 25% or more.

(23)

23

5.1 Structures: child-teacher ratio

The results from the independent samples t-test in preschool units did not show any statistically significant differences regarding the child-teacher ratios between the units with low proportion of SLLs (M = 5.26, SD = 1.13) and the units with high proportion of SLLs (M = 5.7, SD = 1.1). In both groups the child-teacher ratio was approximately 5:1.

5.2 Processes

5.2.1 Teachers’ tone

Teachers’ tone differed significantly between units with low (M = 3.31, SD = 0.14) and high proportion of SLLs (M = 3.16, SD = 0.1); t(66) = 4.294, p<.001 More specifically, teachers in units with low proportions of SLLs were observed to have a significantly more positive tone compared to the teacher in units with high proportions of SLLs1_.

5.2.2 Children’s verbal interactions

Even though the overall observed verbal interaction of children did not differ significantly, as participants in both units were mostly nonverbal, there were some significant differences regarding whom the children were talking to (see Table 3). The most frequent verbal behavior in both units was being verbal to another child. However, in units with low proportions of SLLs children were verbal to the teacher at the same frequency as they were verbal to another child, a behavior that was not shared by the children in the units with high proportion of SLLs in a statistically significant way (p = .029). At the same time children in units with high proportions of SLLs were significantly more frequently observed talking to themselves compared to the units with low proportions of SLLs (p = .002).In a more descriptive way, it can be said that in both units children were most frequently observed being non-verbal. When they were verbal though, they were observed to most frequently talk: to another child, to teacher, to themselves, to a small group, to the whole group with the teacher, to small group with the teacher and to the whole group without the teacher; in that order. Significant differences exist in the fact that: children in units with low proportions of SLLs talk with the same frequency to the teacher as

1_{The labels for the variable of teacher’s tone were: 1= extreme negative, 2= negative, 3= flat, 4= positive, 5=}

very positive. Thus a mean value of above 3 is reflecting a more positive tone. The higher the mean number the more positive the average teacher’s tone.

(24)

24

they talk to another child when children in units with high SLL proportions are not talking to the teacher that frequently and they talk to themselves significantly more.

Table 3

Differences between SLL groups concerning whom they are verbal to.

Verbal to whom SLL Group N M SD T df p C Low 49 .18 .07 .614 66 .542 High 19 ,17 .06 T Low 49 .18 .06 2.231 66 .029 High 19 .14 .05 Self Low 49 .07 .03 -3,298 66 .002 High 19 .10 .03 SG Low 49 .02 .02 -.018 66 .986 High 19 .02 .01 WG+T Low 49 .02 .03 .402 66 .689 High 19 .02 .03 SG+T Low 49 .01 .01 -.805 66 .424 High 19 .01 .01 WG Low 49 .00 .00 .596 66 .553 High 19 .00 .00

Note. The abbreviations of whom the children were verbal to refer to; C: to another child, T: to

teacher, Self: to themselves, SG: to a small group of children, WG+ T: to the whole group of children and teacher, SG+T: to a small group of children and teacher and WG: to the whole group of children.

5.2.3 Proximity

Concerning proximity, in both units with high and low proportion of SLLs, most frequently observed was the proximity to a small group with the teacher, followed by the proximity to a small group (without teacher). Non-significant between groups differences were revealed for any of the types of proximity as seen in Table 4.

(25)

25 Table 4

Differences between SLL groups in Proximity.

Proximity SLL Group N M SD t df p SG+T Low 49 .34 .11 -.557 66 .579 High 19 .36 .09 SG Low 49 .22 .08 1.074 66 .287 High 19 .20 .07 Self Low 49 .14 .08 -1.310 66 .195 High 19 .17 .09 C Low 49 .13 .05 -.166 66 .869 High 19 .13 .004 WG+T Low 49 .13 .01 1.603 19.926 .114 High 19 .09 .01 T Low 49 .03 .03 -.817 66 .417 High 19 .04 .03 WG Low 49 .00 .01 -.762 66 .455 High 19 .01 .01 5.3 Outcomes 5.3.1 Involvement

As indicated by the results of the independent samples t-test, the proportion of SLLs in the unit did not have a significant effect on the low/ medium-low and medium levels of involvement of children between preschool units with high and low proportion of SLLs as in both groups those were the most frequently observed levels of involvement. However, when it comes to medium-high/high levels of involvement, which is the preferable level of involvement as it is connected to better learning outcomes, the results suggested that the proportion of SLLs had a significant effect. Students in preschool units with low proportion of SLLs were more frequently observed in medium-high/high levels of involvement, compared to the units with high proportions of SLLs (p = 0.047) (see results in Table 5). In Figure 2 the distribution of different levels of involvement is presented for in the units with high and low proportion of SLLs.

(26)

26 Table 5

Differences between SLL groups in levels of Involvement.

Note. The numbers displayed in M and SD columns represent percentages.For example, the

reported mean value M = .44 in the first row indicates that the units with low proportion of SLLs were observed in Low/Medium-Low levels of engagement in 44% of observations, on average.

Figure 2

Distribution of different levels of involvement.

Note. The percentages

on the bars display the frequency of the different levels of involvement for each group. The blue colored bars represent the units with low proportion of SLLs and the green colored bars the units with high proportions of SLLs. Level of Involvement SLL Group N M SD t df p Low/Medium Low Low 49 .44 .08 -1.54 66 .127 High 19 .48 .10 Medium Low 49 .29 .11 -.915 52.992 .364 High 19 .32 .07 Medium High/High Low 49 .25 .15 2.027 61.927 .047 High 19 .19 .07

(27)

27 5.3.2 Social Interactions

Lastly, regarding the outcome of social interactions and types of play, in groups with both high and low proportion of SLLs, the majority of interactions was of the non-social type as children were most frequently observed to be engaged in parallel play or were observed being alone (see Table 6 below). However, the results of the independent samples t-test highlighted, that preschools with low proportions of SLLs have significantly more social interactions compared to preschools with high proportions of SLLs students (p = .045). More specifically, when socially interacting, associative play was more frequently observed in both groups compared to cooperative play or social talk. However, associative play in units with low proportions of SLLs was significantly more frequent compared to units with high proportions, p = .031, when the other types of social interaction (cooperative play and social talk) did not show any significant differences between the groups.

Table 6

Differences between SLL groups in types of Interaction. Interaction type SLL Group N M SD SEM t Df P Non-Social Low 49 .35 .09 .01 -.309 66 .758 High 19 .36 .09 .02 Social Low 49 .26 .11 .02 2.055 51.521 .045 High 19 .21 .07 .02 Parallel play Low 49 .26 .06 .01 1.455 66 .150 High 19 .24 .07 .02 Associative play Low 49 .21 .12 .02 2.224 51.557 .031 High 19 .15 .07 .02 Alone Low 49 .08 .05 .01 -1.707 66 .093 High 19 .11 .08 .02 Cooperative play Low 49 .02 .02 .00 -.965 66 .338 High 19 .03 .03 .01

(28)

28

Social talk Low 49 .03 .02 .00 .144 26.123 .887 High 19 .03 .02 .01

Onlooker Low 49 .01 .02 .00 -.566 66 .573

High 19 .01 .02 .00

Note. Social interaction type is the sum of Associative, Cooperative play and Social talk, while

non-social interaction type is the sum of Parallel play, Alone and Onlooker behavior.

6. Discussion

6.1 Exploration of results

This study was an attempt to describe inclusiveness by comparing structures, processes and outcomes in preschool units with high and low proportion of SLLs in Sweden. The results highlighted lack of differences in the structural factor of child-teacher ratios and the process aspect of proximity. However, significant differences were revealed in process aspects of teacher’s tone, children’s verbal interactions, and the desired children’s outcomes of involvement and social interactions and types of play when the units with different proportions of SLLs were compared. Those results illustrated in Figure 3 will be discussed in relation to previous literature regarding inclusiveness in early education and support of SLLs.

Figure 3

(29)

29 6.1.1 Outcome of involvement

First, the results show that in units with high proportions of SLLs children are observed less frequently in higher levels of involvement compared to the units with low proportions of SLLs. The lower frequencies of medium high and high levels of involvement that was reported for units with a high proportion of SLL students in this study support and may justify previous reports for poor academic, social, psychological and other outcomes of SLLs in preschool (Jerome, 2009; Halle et al., 2012; Winsler et al., 2014). Since involvement is a facilitator of children’s learning and holistic development (Åström et al., 2020), and higher levels of involvement are the desired outcome and indicator of inclusiveness, further research should be implemented in order to identify the barriers and facilitators of involvement (Piker, 2013), especially in vulnerable groups of students such as SLLs.

According to the ecological model of inclusive early childhood education (Kyriazopoulou et al., 2017) there is a plethora of factors that can either work as a barrier or facilitator of inclusiveness where the desired outcomes are child engagement (involvement), belongingness and learning. Some of them are the curriculum and pedagogical approaches, the infrastructure and teachers’ skills and attitudes (UNESCO, 2017).

Previous research has shown the effect of environmental factors on children’s involvement in different types of activities in preschool (Sandberg et al., 2010; Mashburn et al, 2008; Brodin & Renblad, 2015; Coelho et al., 2019; Prykanowski at al., 2018; Sjöman at al., 2016; Åström et al., 2020). The type of activity, the accessibility and adequacy of the materials, the quality of social interactions and the quality and style of teacher-child interactions are proven to affect involvement (Coelho et al., 2019; Prykanowski et al., 2018; Sjöman et al., 2016). In this study the type of activity and the factors regarding materials were not investigated nor described. However, observations took place in order to provide a picture of the quality of social interactions and the quality and style of teacher child interactions.

6.1.2 Processes and involvement

6.1.2.1 Teacher’s emotional tone

Starting with the style of teacher-child interaction, a significant result was found regarding the teacher’s tone. Teachers were observed to have a more positive tone in units with low proportions of SLLs compared to teachers in units with high proportions of SLLs. Emotion and learning are connected (Park, 2014) and research has previously shown that teachers’ warmth,

(30)

30

expressed joy, positive emotional tone and approval responses in classrooms are associated with and can predict children’s outcomes such as positive social interactions and the development of cognitive self-regulation skills (Fuhs et al., 2013; Spivak & Farran, 2016). Additionally positive tone is one of the strategies teachers can use to support the needs of SLLs (Ng et al., 2018). In this study teachers in units with high proportion of SLL were observed to have a flat tone on average. This lack of teachers’ positive tone could be linked to the outcomes of lower levels of involvement and less social interactions in line with the systems theory and the proximal processes’ concept of bioecological model which highlight the bidirectional impact of environment and child outcomes (Imms et al., 2017; Bronfenbrenner & Morris, 2006; Wachs, 2000). On the other hand, one could say that since the majority of time in Swedish preschool is spent in free play activities and teachers’ main task is managing (Åström et al, 2020), it could be reasonable that teachers tend to have a flat tone.

However, the general flat tone of Swedish preschool teachers along with the less positive tone of teacher’s in units with high proportions of SLLs are results in contrast with the recommendations of literature that highlight the importance of warm and responsive teacher attitudes towards students (Fuhs, Farran, & Nesbitt, 2013; Spivak & Farran, 2016; Sjöman, Granlund, & Almqvist, 2016; Kyriazopoulou et al., 2017; Puskás, 2017) and they are opposite of what someone would expect from teacher in preschool contexts considering the preschool children’s emptional needs. It is crucial that teachers interact, communicate and connect with their students and respond to their physical, intellectual and emotional needs, especially the SLLs’, in attempts to enhance school’s cultural responsiveness (Peterson, McIntyre, & Glaés-Coutts, 2018; Henward, Tauaa, & Turituri, 2019). This is a prerequisite to achive children belongingness and participation (Kyriazopoulou et al., 2017).

6.1.2.2 Children’s verbal interactions and proximity

Verbal interactions with the teacher

Another significant point for discussion revealed by the results, which is also related to teacher-child interactions, is that teacher-children in units with low proportions of SLLs are verbal to the teacher with the same frequency as being verbal to another child. At the same time, children in units with high proportion of SLLs, even though equivalently verbal to another child as the aforementioned group, are significantly less verbal to the teacher and more verbal to themselves. In attempts to create a better picture of the verbal interaction behavior, it must be related to observations of the proximity aspects. In this study both groups of units demonstrated

(31)

31

types of proximity in the same observed order and with no significant differences in the frequencies. Combining the result that both groups were most frequently in proximity to small group with the teacher and children in both groups were more frequently verbal to another child, the differences in verbal interactions exist in the fact that while children in the group of low proportion of SLLs more frequently are verbal to the teacher, children in the high proportion group are more verbal to themselves.

In this study only the frequency was observed and not the quality of verbal interactions. As suggested by Meng (2019) it is not just the language interaction that has an effect. Other factors such as teacher’s sensitivity might play an important role in the whole process of language development of young children and especially the language acquisition of SLLs. The lack of positive teacher’s tone along with the less frequent attempts of children to talk to teacher and more likely talking to themselves in the group with high proportion of SLLs rises concerns about teachers’ attitudes and whether they are facilitating inclusiveness (UNESCO, 2017).

On the other hand, self-talk/ private speech is age related phenomenon common in preschool aged children and different theories exist regarding whether this is a sign of confusion, a tool for problem solving activities or a transition point between vocal and inner language (see Frauenglass & Diaz, 1985). The children in the two groups of units did not differ significantly in age, something that could explain the increased private speech in the group with low proportion of SLLs.

Peer verbal interactions

Regarding the verbal interaction with peers, no significant differences were found as both groups were observed being most frequently non-verbal, even though previous literature reported more shyness and internalized behavior for SLLs (Ash et al., 2014). However, there are a few things worth discussing. Interaction with other children may be different dependent on whether the peers speak the same mother tongue with the SLL or not. In this study it was not possible to distinguish whether SLLs were speaking their mother tongue with their peers or not. This distinction would provide more information about the potential feeling of belongingness of children in those groups as it is mentioned in the ecological model of inclusive education that having speakers with the same language background as the child would likely lead to that outcome (Ash et al., 2014).

Previous research reports that the way students interact and the language students speak in class depends upon their language status, proficiency and the language of the instructions

(32)

32

(Sawyer, Atkins-Burnett, Sandilos, Scheffner Hammer, Lopez, & Blair, 2018). Also, according to Vygotsky (1978), language facilitates the internalization of social structures which in turn allows children to learn to interact with others. By enhancing interactions in preschool settings between teachers and students along with the peer interaction, the development and learning of students at risk can be supported (Coelho et al., 2019). So depending on the desired outcome, different types of language need to be used by adults interacting with children. For instance, if the desired outcome is to improve the language acquisition, then there is a need for children to have positive verbal interactions with native speakers of the target language (Chapman, 2000). If the desired outcome is to enhance the feeling of belongingness, then one way is to have speakers with the same language background as the child (Ash et al., 2014). In this study it was not clarified if the SLLs in the units were from the same linguistic background so that they had opportunities to speak to one another in their mother tongue. This can be an interesting aspect to investigate in future research as it could highlight if the verbal interactions and other processes along with the child outcomes in units with SLLs are differentiated depending upon whether the SLLs have the same mother tongue or not.

6.1.3 Outcome of social interaction in play

Continuing with the social interactions in play, children’s different types of play need to be discussed. Play is an important aspect for development and learning recognized by the Swedish curriculum, thus the types of play demonstrated by children in preschool units is also an aspect of interest when it comes to preschool outcomes. Play is a factor that includes and reflects social interactions between children and is essential for learning and development (Swedish National Agency for Education, 2019; Parten, 1932). Thus, the guidelines from the ecosystem model of inclusive early childhood education and care suggest that assessment of children’s outcomes should consider self-initiated and social play as well (Kyriazopoulou et al., 2017).

Previous research reported that SLLs were displaying significantly more onlooker behavior compared to their native peers and were engaged in different type of activities (Sullivan, Hegde, Ballard, & Ticknor, 2015). In this study, in both units with high and low proportions of SLLs, the majority of interactions was of the non-social type and both groups were mainly engaged in parallel play. In previous studies parallel play was considered a social type of play along with associative and cooperative play (see Farran & Son-Yarbrough, 2001), unlike this study. Associative and cooperative play are the only ones requiring interactions with others

(33)

33

and generally more complex verbal interactions, therefore they were the only types of play classified as social play in this study.

The fact that students in this study were primarily observed to be engaged in non-social types of play, raises concerns about how to motivate and support them to engage in other types of play more often. More positive social outcomes and less behavior problems occur when the interactions between peers are positive and more of cooperative type (Spivak & Farran, 2016). Children in units with high proportion of SLLs were also observed to have little cooperative play but also significantly less associative play compared to the units with low proportions of SLLs.

According to Neitzel, Drennan and Fouts (2019) both SLLs and non-SLLs are socially engaged and they all are seeking other peers to play with. However, their interaction type differs in terms of roles, responsiveness, and social construction strategies. Considering that both groups in this study were mainly in proximity to a small group of children with or without the teacher, we could assume that they both have the same opportunities to develop same types of play at least contextually. However, the fact that SLLs develop less associative types of play rises questions on whether they need support for their social play in preschool in order to support daily interactions and to achieve positive socio emotional outcomes (Stanton-Chapman et al., 2014).

6.1.4 Stucture of child-teacher ratio

Continuing with the structural aspects, even though the ecological model of inclusive education emphasizes the importance of adaptations, no differences in the child-teacher ratios were reported between the groups with high and low proportions of SLLs in this study. The lack of structural differences despite the differences in proportions of SLLs, generates questions on whether it is expected of SLLs to adapt to the context and face the challenges without extra support, adaptations and arrangements (Nilsson & Bunar, 2016). On the other hand, previous reports about ratios show contradicting results on whether small or large ratios lead to better outcomes (Iluz, Adi-Japha, & Klein, 2016). However, it is reported that in Sweden, the preschool units with an organizational perspective of special support which is based on the demands on the unit, tend to have lower child-teacher ratios. Those units reduce the amount of children in the group and also increase the number of regular staff when many students in need of special support are identified (Sandberg et al., 2010). It can therefore be expected that the ratios in this study’s groups would be different, with the units with high proportion of SLL to

Structures, processes and outcomes in preschool units with high and low proportions of second language learners of Swedish : A Comparative Study