A Training on Perspective Taking for Preschoolers with Typical Development and with High-Functioning Autism Spectrum Disorder

Perspective taking refers to the ability to understand others’ thoughts, feelings and perspectives. From previous studies, it emerges that children with a diagnosis of autism spectrum disorder show many difficulties in understanding the other’s point of view, especially in terms of thoughts and emotions. The aims of the present study were: (1) to verify the effectiveness of group training in promoting perspective taking in a small group of preschoolers including two children with diagnosis of high-functioning autism spectrum disorder; and (2) to verify whether the learning in the perspective taking skills could be transferred to everyday situations in terms of prosocial behaviors. Twelve children (mean age: 54.5 months), 10 with typical development and 2 with high-functioning autism spectrum disorder, participated in the study. The effectiveness of training was evaluated in a pre/post-test design both via experimental tests and direct observations. The results highlighted the effectiveness of the training in enhancing perspective taking ability and prosocial behaviors of preschoolers, including the two children with a diagnosis of high-functioning autism spectrum disorder. The use of non-clinical and educational methodologies can contribute to the development of children's perspective taking skills by means of daily educational activities in small group including children with high-functioning autism spectrum disorder. These results have great applicative relevance for early childhood education and care centers.


Introduction
Perspective taking refers to the ability to understand others' thoughts, beliefs, feelings and perspectives and implies being able to distinguish what individuals know about themselves in a given situation and what others know in that same situation. This ability allows us to infer what the other thinks, what they feel emotionally, how they perceive events and the motivations underlying their behavior, allowing us to give meaning to the surrounding world (Cigala & Mori, 2022;Moll & Meltzoff, 2011). Perspective taking appears to be peculiar to the emergence of positive social competencies such as empathy, cooperation and acts of altruism (Epley & Caruso, 2012).
In particular, perspective taking was conceptualized as a multidimensional construct characterized by three components: visual, affective and cognitive (Fireman & Kose, 2010). Visual perspective taking refers to the ability to make inferences about how an object presents itself to an individual who occupies a position within the space other than his/her own (Moll & Meltzoff, 2011;Pearson et al., 2013). Affective perspective taking is instead the ability to understand other people's emotional states, especially when they differ from their own (Fireman & Kose, 2010). Finally, cognitive perspective taking is the ability to infer the thoughts, motivations and intentions of other people (Baron-Cohen, 2001;Ornaghi et al., 2015). The common characteristic of all three dimensions is the ability to set aside an egocentric position adopting the other's different point of view.
Several studies ad theoretical dissertations have shown that typically in individuals with autism spectrum disorder (ASD) different social abilities related to the development of intersubjectivity are compromised with important negative influences on adaptation to contexts, especially to those that have a character of novelty (Martin et al., 2008;Spera et al., 2008;Yirmiya et al., 1998). It is possible to trace some studies that have investigated the perspective taking ability in individuals with ASD. From these studies, it emerges that children with ADS typically show many difficulties in understanding the point of view of others, especially in terms of thoughts and emotions (Baron-Cohen et al., 2013;Begeer et al., 2007;Jones et al., 2010;LeBlanc et al., 2003;Leslie & Frith, 1990;Reed & Peterson, 1990;Schwenck et al., 2012).
Despite the scientific contributions on the development of perspective taking in individuals with ASD, there is less evidence regarding the effectiveness of interventions to promote in children with ASD the ability to take the perspectives of others (Gould et al., 2011;McCollow & Hoffman, 2019). Southall and Campbell (Southall & Campbell, 2015) conducted a review that aimed to analyze the empirical research (conducted since 2000) of effectiveness of interventions targeting social perspective taking skills for individuals with ASD. A very complex framework of studies emerges from this review. In particular, some studies used a single-case research method, multiple baseline across either tasks or participants to evaluate behavior change Chin & Bernard-Opitz, 2000;Hua Feng et al., 2008;LeBlanc 1 3 et al., 2003), other research used group design (Bauminger, 2007;Begeer et al., 2011;Bernard-Opitz et al., 2001;Fisher & Happé, 2005;Solomon et al., 2004;Turner-Brown et al., 2008). Among these studies focused on group design, only a few used a pre-/post-test design (Gevers et al., 2006;MacKay et al., 2007). Most of these studies focused on the treatment with children with ASD involved school-age children / adolescents (from 6 to 16). In general, the results indicate that perspective taking skills can be taught using systematic instruction, group interventions, and technology, but generalization of skills into naturalistic settings was inconsistent (Fisher & Happé, 2005;Swettenham et al., 1996).
From these studies, some aspects seem to be emerging that still require particular attention from scholars. On the one hand, an open question remains, namely the aspect of generalization and the ability of children to transfer learning from training to everyday situations. On the other hand, the literature also highlights how generally both group and individual training have been proposed in clinical contexts and mainly conducted by therapists. In fact, even studies that use a comparison between groups usually involve a comparison between a normative group and a clinical group. Rather, there is a lack of training experiences that envisage in the same group the simultaneous presence of children with ASD and children with typical development, as the groups in school contexts.
Finally, another interesting aspect that can be found in the literature is the presence of training specifically focused on one or at most two components of perspective taking. In particular, there is a lack of training carried out with the aim of keeping the visual component inside together with the affective and cognitive one, identifying them as different from each other and not overlapping, albeit in interaction (Mori & Cigala, 2016).
Another aspect that emerges from the studies concerns the question of "what" is strengthened and promoted with specific training: the right answer to that task (the solution) or the ability to take the perspective of the other? Hence, in addition to "what," another question can be asked about "how" a specific training fosters perspective taking, namely: Which learning mechanisms, which learning processes are activated in a specific training as compared to another one?

The Current Study
The aim of the present study was to verify the possibility of promoting visual, affective and cognitive perspective taking in preschool children by means of a small group training.
In particular, starting from the questions emerging from the analysis of the literature (Cigala & Mori, 2022;Mori & Cigala, 2016;Southall & Campbell, 2015), this contribution aims to implement a training, structured in such a way as to take into account the criticalities highlighted in previous studies. First, since perspective taking is a complex and multicomponent competence, it was decided to propose training that included different activities to enhance all the dimensions involved in this skill: visual, affective and cognitive perspective taking.
Secondly, starting from the evidence that most of the previous studies were carried out in clinical settings and that many of these studies seemed to highlight a lack of generalizability of the learning, it was decided to implement training that could take place in a daily non-clinical context, such as the educational one, where children with high-functioning autism spectrum disorder (HFASD) share experiences together with other typically developing children. The reason behind these choices lay in the assumption that in order to allow children to consolidate their learning and transfer them to other activities and tasks different from those specifically used in training, it is very important to implement training that is ecological. Ecological training refers to training courses that offer methods, activities, interactive formats and learning spaces that are familiar to children and that they have the opportunity to experience in the daily context. In this perspective, it was considered interesting to propose a group training that envisaged the co-presence of typically developing children and some children with HFASD who share the scholastic educational context day after day.
Finally, another critical aspect concerns the interactive format proposed in the training and the role of the trainers. In this regard, based on previous empirical evidence (Cigala et al., 2015;Mori & Cigala, 2019), we focused on training consisting of different types of problem-solving and tasks, which involved an active interaction between children. The role of the adult trainer in this training is not simply that of "teaching the right answer", but that of proposing activities and tasks, activating and coordinating the discussion, and allowing children to directly experience different points of view of all group training participants (visual, affective and cognitive) regarding the different situations proposed.
Therefore, the first aim of the study was to verify the effectiveness of a training group in promoting affective, visual and cognitive perspective taking. This training group was conducted in an educational context and engaged a group of preschool children that includes typically developing children and some children with HFASD.
The second aim was to verify whether, following the implementation of the training, there was also an increase in prosocial behavior between peers exhibited by training group participants in the kindergarten. Starting from previous empirical studies that highlight a relationship between prosocial behavior and perspective taking (Berti & Cigala, 2020;Cigala et al., 2015;Emen & Aslan, 2019;Farrant et al., 2012;Longobardi et al., 2019;Mori & Cigala, 2019) we aimed to verify whether the learning in perspective taking abilities reached by children after training could be transferred to everyday situations in terms of the frequency of prosocial behaviors among peers.
In particular, on the basis of previous studies conducted on typically developing children (Mori & Cigala, 2019), we expected that children would improve their performance in their affective, cognitive and visual perspective taking abilities as a result of training. It was also hypothesized that the frequency of peerto-peer prosocial behaviors would also increase following training group both for the group as a whole and for the children with HFASD.

Participants
Twelve children participated in the present study aged between 3 and 5 years (mean age of 54.5 months), of whom 6 males and 6 females, all attending the same kindergarten. This group included 2 children with diagnosis of High autism spectrum disorder (HFASD), one male and one female, named her Fabio and Laura. All participants were recruited from the school attended.
This study is part of a larger project that involves the collaboration between the School Institution and the Clinical Center for the diagnosis, treatment and study of communication and socialization disorders in a city in northern Italy. At the time of the study, Fabio and Laura were attending this center.
Fabio (62 months) was an only child and Laura (43 months) had a younger brother, also diagnosed with autism spectrum disorder much more severe than his sister. They were diagnosed with autism disorder by a multi-professional team (including neuropsychiatrists, psychologists and speech therapists) of the Clinical Center for the diagnosis, treatment and study of communication and socialization disorders, according to DSM-5 criteria (APA, 2013) using the Autism Diagnostic Observation Schedule (ADOS-2) (Lord et al., 2012) and Autism Diagnostic Interview-Revised (ADI-R) (Rutter et al., 2005). Both children presented the absence of comorbidities (e.g., attention deficit / hyperactivity disorder APA 2013) and non-verbal IQ in the normal range (full-scale IQ on the Italian version of Wechsler Preschool). Direct observations conducted in school prior to the study revealed that both children not exhibited aggressive or severe disruptive behaviors. Fabio engaged quite frequently in play sessions and some noncompliant behaviors (screaming and refusing to do work); Laura had limited play skills and engaged in repetitive behaviors (hand flapping, repeating questions, and perseverations).
At the time of the study, the children attended the Clinical Center once a week and participated in psychomotor activities and strengthening the main autonomies; all activities were conducted individually and no intervention was provided to enhance socio-emotional skills.
Participants were divided into 2 groups balanced by gender, age and section: 6 children made up the control group and 6 children in the experimental group. The two children with HFASD were included in the experimental group. None of the children had previously taken part in similar research.
Consent for conducting the study was obtained from the school principal, and, prior to the data collection, the parents' informed written consent was obtained, for each child who participated, in compliance with the ethical norms set down by the American Psychological Association. Sally & Annie (Wimmer & Perner, 1983) The caterpillar sleeping on a pillow (Flavell et al., 1981) Emotional decentralization task (Harwood e Farrar, 2006) Nesquik box with pencils The Enchanted Valley (Hughes & Donaldson, 1979) Rainbow and his friends (Pfister, 2009) Pen-flower (re-adapted Flavell et al., 1986) The crocodile (Flavell, 1986) POST-TEST TEC (Pons & Harris, 2000) The mice "Stefano e Alberto" (Liverta Sempio et al., 2005) The turtle sleeping on a pillow (Flavell et al., 1981) Emotional decentralization task (Harwood e Farrar, 2006) Bottle water with candy (Perner et al., 1987) The Lego farm (Hughes & Donaldson, 1979) Pingu and his friends (Von Flue, 1990) Syringe-Highlighter (re-adapted from Flavell et al., 1986) The dog (Flavell et al., 1981)

Procedure and Materials
The research design was a quasi-experimental procedure that included a pre-test, training and post-test phase. In the pre-test and post-test phase, each lasting onemonth, affective, cognitive, visual perspective taking and prosocial behaviors were assessed (see Table 1).

Perspective Taking
Three tasks were proposed to each child, administered individually, in order to detect each component of perspective taking in pre-test and post-test phase.

Prosocial Behaviors
Non-participating ecological observations of prosocial behaviors were conducted: Each child was observed for 45 min on 3 different days, at 3 different moments: during unstructured play before the meal, during the meal and during unstructured play after the meal. Prosocial behaviors were detected by a trained research assistant who was blind to the specific aims and hypotheses of the study. The frequency of emission of each prosocial behaviors was coded by means of an observation grid that includes more than fifty prosocial behaviors divided into three categories: helping, consoling/supporting and sharing. This coding grid had already been used in previous research (Cigala et al., 2015;Cigala & Mori, 2022) and the observer was trained to use it in preschool contexts. Cronbach's alpha score for the reliability of the internal consistency of the coding grid is 0.69 for each behavior separately.

The training
The training includes nine sessions (three for each perspective taking dimension) with a three-weekly frequency. The activities were proposed to the experimental group in a quiet room inside the kindergarten. During this period, the control group carried out daily curricular activities with their teachers. According to ethical standards, at the end of the research, we proposed the training to the control group as well. The training had already been used with children of this age with typical development and the measure treatment fidelity inter-rater reliability calculated across the sessions was 80% (Mori & Cigala, 2019). We considered that this training could prove effective also with children with HFASD since it involves the use of figurative material, simple tasks and active movements in space that allow children to experience the different perceptual points of view. The training was characterized by the following key aspects: (1) Addressed to a small group of children.
(2) Use of stimuli and methodologies (such as the reading of stories, dramatization and graphic-pictorial activities) pleasant and familiar to children and ecological in the sense of already widely used in the educational context. (3) The realization inside the kindergarten, i.e., a familiar context, which children know well, because they attend it every day, with their companions and their educators. (4) The adult's twofold role: (a) putting forward situations and stimuli to promote the children's awareness of their own point of view (i.e., visual, cognitive and affective) and the awareness of the existence of points of view different from their own; (b) highlight the coexistence of different looks and perspectives creating in the various meetings a perspective taking space, in which different perspectives can coexist and be explained. The complete description of the single session of the training is reported in Appendix (Mori & Cigala, 2021).

Scores Coding
Before the analysis of the results, it was necessary to recode the scores obtained by the children in the individual perspective taking tasks and in the observation session. In order to have a single perspective taking indicator for each dimension, first, three new scores for each dimension (affective, cognitive, and visual PT) were calculated according to following formula: 1 3 PT index = (score on task1/max score on task 1) + (score on task2/max score on task2) + (score on task3/max score on task3)* 1 3 Then, the values were multiplied by 100 in order to increase readability. In this way, we obtained three new scores ranging from 0 (poorer performance) to 100 (best performance): visual PT, affective PT and cognitive PT. Moreover, a global Index of perspective taking was computed by averaging the three scores described (TOT PT).  In a previous study, in order to evaluate the extent to which the different test would measure similar construct, a confirmatory factor analysis was performed for pre-test and post-test measures separately (see Mori & Cigala, 2019). More precisely, a model with three latent factors (i.e., affective, cognitive, and visual PT) each measured by three indicators was tested using maximum likelihood estimation with robust standard error. For pre-test, results indicated a good fit of the three-factor model, v2 (23) = 8.13, p = 0.99, CFI = 1.00, RMSEA = 0.00 [90% CI 0.00-0.00], p = 0.99, SRMR = 0.022. Moreover, all the observed variables were significantly measured by the intended latent factor (ps ≤ 0.05). Factors core determinacy was 0.81 for affective, 0.77 for cognitive, and 0.93 for visual dimension of PT. Similar results were obtained for post-test measures. Indeed, goodness of fit was acceptable, v2 (23) = 33.02, p = 0.08, CFI = 0.966, RMSEA = 0.046 [90% CI 0.000-0.079], p = 0.54, SRMR = 0.044, and all observed variables were significantly represented by the intended latent dimension (ps < 0.01). Factor score determinacy was 0.96 for affective, 0.84 for cognitive, and 0.84 for visual dimension of PT.
Finally, an indicator of prosocial behavior was also computed by summing the frequency of every single observed behavioral category (TOT PROSOCIAL).

Pre-test
Post-test

Group Results
Descriptive statistics were calculated for all the variables detected at the pre-test and post-test for both the control group and the experimental group (Table 2). In order to highlight the initial level of the participants' skills, a comparison for two independent groups (non-parametric test of Mann Whitney) was made with the experimental/control group as an independent variable and all the perspective taking and prosocial behaviors scores at the pre-test (affective PT, cognitive PT, visual PT, total PT and prosocial TOT) as dependent variables. From the analysis, no significant differences emerged. By observing the mean values, it can be noted that the experimental group is slightly less competent than the control group in different variables, but the difference between the two groups does not reach significance.
To verify the progress of the two groups from the pre-test to the post-test, a nonparametric analysis was performed for two dependent groups (Wilcoxon test) with the pre/post-test as an independent variable and the perspective taking and prosocial behaviors scare as dependent variables. The analysis shows statistically significant improvements in the experimental group between the pre-and post-test in the affec- In the control group, the comparison between the pre-test and post-test variables does not show any significant differences (see Fig. 1).

Cases Analysis: The Children with HFASD
Analyzing the scores of the individual children on the different tests, it is noted, confirming the group data, that the improvements of children with HFASD were evident, especially in reference to the total PT and affective PT (Fig. 2 and 3).

Laura's Profile
In particular, the scores related to Laura's perspective taking ability (Fig. 2) show that in the pre-test phase, she reached a score of 29 (range 0-58; average value of the experimental group: 39), placing her in the lowest performance compared to the group of children, both experimental and control. Following the training, Laura showed a significant improvement in perspective taking ability, reaching a score of 44 (range 0-58; average value of the experimental group: 49), placing herself closer to the average value of the group.
With respect to the different components of perspective taking, it is in the affective dimension that the most significant change occurred. In fact, Laura, starting from an affective PT score at the pre-test of 20 (range 0-45; average value of the experimental group: 29.33), reached the score of 33 at the post-test (range 0-45; average value of the experimental group: 37.33), getting closer to the average value of the group.
With regard to the visual PT (Fig. 2), Laura recorded a score of 4 in the pre-test (range 0-6; average score of the experimental group: 4.17) placing her perfectly in line with the average score of the group. After the training, it was possible to observe an improvement. In fact, the maximum score of the test was reached (average of the experimental group: 6).
With regard to the cognitive PT (Fig. 2), Laura shows in the pre-test a score of 5 (range 0-7; average value of the experimental group: 5.50), slightly below the average value of the group. Following the training there was no improvement in Laura's performance, while the experimental group improved overall (average value of the experimental group: 6.33).
In relation to prosocial behaviors (Fig. 2), Laura in the pre-test observational sessions showed an emission frequency of 20 prosocial behaviors in peer interactions, showing that she enacted fewer behaviors than the average of the experimental group (average value of the experimental group: 30.33). After the training, there was an increase in these behaviors (30) that brought Laura closer to the average emission frequency of the group (32.17).

Fabio's Profile
With respect to Fabio's profile, the scores related to the ability of perspective taking (Fig. 3) show that in the pre-test, the child reached a score of 40 (range 0-58; average value of the experimental group: 39), placing him slightly above the average value of the group. After the training, there is, however, an increase in perspective taking with a score of 47 (range 0-58; average value of the experimental group: 49).
In particular, with regard to the affective PT (Fig. 3), Fabio recorded a pretest score of 30, placing himself slightly above the average score of the group (range 0-45; average value of the experimental group: 29.33). At the post-test, the affective PT increased, reaching a score of 34 (range 0-45; average value of the experimental group: 37.33), in this case placing himself below the average of the group.
With respect to the visual PT (Fig. 3), a trend very similar to that of Laura emerged. Fabio in fact obtained a score of 4 in the pre-test (range 0-6; average score of the experimental group: 4.17) placing himself perfectly in line with the average score of the group; after the training, an improvement was shown and the maximum score of the test was reached (average of the experimental group: 6).
With regard to the cognitive PT (Fig. 3), the child's pre-test score was 6 (range 0-7; average value of the experimental group: 5.50), slightly above the average value of the group. Following the training Fabio reached a cognitive PT score of 7, placing himself above the average score of the group (average value of the experimental group: 6.33) and reaching the maximum score of the tests.
In relation to prosocial behaviors (Fig. 3), Fabio in the pre-test observational sessions showed an emission frequency of 24 prosocial behaviors in peer interactions, showing that he enacted a lower number of behaviors than the average of the experimental group (average value of the experimental group: 30.33). After the training, a considerable increase in the emission frequency of these behaviors was observed (34) which exceeded the group's average emission frequency (32.17).

Discussion and Implications
The framework of the results highlighted the effectiveness of the training in enhancing the perspective taking ability, in fact the experimental group that participated in the training showed an increase in the perspective taking ability at the post-test compared to the pretest, while no changes were observed in the control. It is interesting to note that this increase occurred for all the dimensions of perspective taking considered, i.e., the visual, affective and cognitive.
The analysis of the performance of the two children with HFASD allows us to highlight the effectiveness of the training, in the sense that for both children the perspective taking ability was enhanced in the post-test compared to the pre-test. At the pre-test, Laura showed a lower perspective taking ability than the group average, instead following the training significant improvements were pointed out, especially in emotional perspective taking. On the contrary, the data showed that the training did not allow her to enhance her cognitive decentralization ability probably due to her young age; indeed, at the time of the study, Laura was 43 months (three and a half years old), the youngest of the group of participants together with another child. In addition to this, it should also be noted that the activities proposed in the meetings focused on cognitive perspective taking turned out to be the most complex and articulated for all children. As regards Fabio, the training seemed to have a positive impact on all dimensions of perspective taking: The child showed an increase in affective, visual and cognitive perspective taking ability from pre-test to post-test.
Starting from this data framework, hypothesis number I was verified, although, owing to the small number of participants, the results are to be considered as preliminary and future research will need to verify and assess them more broadly.
With respect to aim II, the framework of the results seems to highlight an increase in the prosocial behaviors toward peers implemented in kindergarten following the training. In particular, the increase in prosocial behaviors was evident for the experimental group, while no significant variations emerged in the control group.
This trend is confirmed by the analysis of the individual profiles of the two children with HFASD, which showed that both Fabio and Laura enhanced the emission of pro-social behaviors toward their peers in the school educational context from pre-test to post-test. The relevance of these results is also emphasized by the fact that in the present study, the prosocial behaviors of help, sharing and comfort are not evaluated through indirect tools, such as self-reports proposed to teachers, as often happens in many researches. These behaviors were detected through direct, non-participating and ecological observation carried out by trained observers who watched children in everyday moments (such as free play and lunch) for long times spread over different days (Mori & Cigala, 2019).
Furthermore, these results referring to prosocial behaviors appear to be of considerable interest, especially in light of the previous studies (Fisher & Happé, 2005;Swettenham et al., 1996), which pointed out a lack of generalizability of the learning achieved through training. Instead, the data of this study show how this training seems to have a positive impact, not only on the perspective taking ability that is specifically strengthened in the training, but also on the prosocial behaviors of help, comfort and sharing directly observed in everyday school context. In this sense, it is plausible to argue that the learning in the perspective taking abilities, enhanced from the training, could be transferred to the context of daily life, influencing interactions in the kindergarten.
In summary, although this is an exploratory study, the results highlight the effectiveness of the training in enhancing the perspective taking ability and the implementation of prosocial behaviors of help, comfort and sharing with peers. The effectiveness of this training had already been verified in previous studies with numerous groups of typically developing children (Mori & Cigala, 2019;Cigala et al., 2015); the present study brings a new contribution, as the effectiveness of this training had never been tested with groups, including both children with HFASD and typically developing children.
The use of non-clinical and educational methodologies highlights how teachers can also contribute to the development of children's perspective taking skills by means of daily educational activities that can help children to become aware of their own point of view and that of others, even in case of a specific developmental disorder like HFASD (Walker & Berthelsen, 2008).
Furthermore, this study allows for some reflection on the group, meant as a significant space that allows to children with HFASD the "experience of perspective taking." That is to say, we maintain that the group in this training was a significant context of expression, comparison, integration of the different visual, affective and cognitive points of view of the children, through the coordination and support of the adult. An adult, therefore, who does not correct and teaches the solution, rather an adult who facilitates and trains children in perspective taking experiences.
The main weakness of the study is certainly the limited number of participants; in fact, the results obtained are to be considered as preliminary and require further corroboration that may stem from other similar studies conducted with many groups of children in multiple school contexts. Concerning to future studies, a new direction could be to implement some similar training for older children adapting activities and proposals of the meetings to the specific age. Another limitation of this study is also that an alternative activity was not proposed to the control group, which would allow to further corroborate the effectiveness of the training. Furthermore, it will be interesting in future research involving more groups of children, to consider different types of training paths aimed at enhancing perspective taking by comparing their effectiveness. Finally, in following research, it will be interesting to try to further simplify some materials to enhance