The Association Between Reading and Emotional Development: A Systematic Review

Reading practice is associated with numerous psychological benefits. However, its influence over individual emotional dimensions has generally been underestimated by research. Only recently has it been recognized across different developmental stages but evidence is still scarce. The aim of this systematic review is to shed light over the association between reading and the several (and sometimes hardly distinguishable) socio-emotional constructs that we have identified in literature: interpersonal skills and prosocial behavior; emotional and behavioral symptoms; emotional regulation and expression; empathy and theory of mind; emotional knowledge and comprehension; and emotional responses. A total of 50 studies were analyzed, including all age groups, various settings, research drawings, and different emotional constructs in order to create a comprehensive view of the association between reading and emotions. Results show that overall reading practice has a positive impact on socio-emotional development, whatever its declination, regardless of age, gender or setting of implementation.


Our Theoretical Framework: A Summary
We can therefore summarize that reading has substantial power over the many mechanisms that we have enclosed in the term "socioemotional competence". Reading simulates real life (Gerrig, 1993): readers experience the characters' emotional states in the stories and train themselves to grasp and understand their own and others' emotions in a better way (Mar & Oatley, 2008). By coming into contact with the stories and the characters' emotions, a resonance process is activated in which we bring our previous experiences closer to the events of the story, giving meaning to the reading in a clearer and more precise way, living the underlying emotions more intensively and decisively characterizing the reading experience in general (Mumper & Gerrig, 2019).
This immersion in reading enables the reader's empathy (Johnson, 2012) and the inference of other people's mental states. These are fundamental skills to adapt to the real world (Mar et. al, 2011;Mumper & Gerrig, 2017). In line with this point, the SPaCEN explanation (Mar, 2018) also highlights how stories foster the development of our mental capacity and social cognition. These different contributions have as a key assumption that reading has a strong influence over emotions and are able to highlight various aspects of this association. Reading allows us to experience, know and recognize emotions, to understand and grasp our own and others' emotional states, as well as to re-propose these socio-interpersonal skills in real life, improving our emotional and behavioural skills and favouring a continuous adaptation to the environment.

The Purpose of This Study
Our study is rooted in the debate of literature about the role of reading in fostering emotional development. On the basis of the above theoretical considerations, the purpose of this study is to investigate the association between reading and emotional development across the recent psychological and educational literature, trying to add evidence to this still little explored research field. For this purpose, following an overview of the whole life path, with no any age restrictions, we collected empirical studies in which the association between reading and emotional development was explored, in order to investigate whether reading leads to improvements in emotional dimensions. As Kozak and Recchia (2019) discussed, there is a lack of studies concerning school-age: for this reason, our review tries to fill this gap by adding evidence of this life period.

Retrieval of Studies and Inclusion Criteria
In order to obtain studies examining the effects of reading on emotions, the PICOS method was taken as a reference (Moher, Liberati, Tetzlaff, & Altman, 2009). In this approach, each letter identifies an element of the research question: participants (P), intervention (I), comparisons (C), outcomes (O) and study design (S). This method led us to define the research question in a structured way by establishing seven inclusion criteria.. The first criterion was to select articles published from 2011. This time range was chosen to focus on the most recent literature. A second criterion was to select peer-reviewed articles. All empirical studies were included: evidence based, studies based on qualitative analysis, observations or mixed methods; the purpose was to investigate the impact of all types of reading on emotions and this could be observed through numerous types of research. Moreover, since there are many different definitions and means of assessment related to emotional concepts (Gerdes et al., 2010), we decided to take into account different strategies to evaluate emotional development. Non-empirical studies were excluded. The third criterion for inclusion was language: contributions published in English, Italian, and Spanish were included, contributions published in other languages were excluded. Fourthly, in terms of populations considered, the analysis covered all ages, with no age restrictions. Studies conducted on a population with a learning disability or learning disorder were excluded from the analysis in order to focus the analysis on the impact of reading activity on emotional factors, while trying to limit the influence of other variables. Fifth, as far as the setting is concerned, all studies with home environments, labs, schools, universities or other services with an explicit educational purpose as their setting were included. The analysis focused on reading in a broad sense, and therefore no limitations were placed on the modalities and characteristics in which this activity could be carried out. As a sixth criterion, studies involving the use of individual or shared, silent or loud reading were included: reading training, reading sessions and educational laboratory based on reading. Studies that did not include reading interventions but assessed reading habits or language skills were also included. Therefore, socio-emotional development programs with an impact on reading performance were excluded. The use of reading through technological media was excluded. Studies including learning programs of a second language were also excluded. The same procedure was adopted for the emotional dimension; studies that investigated emotional development, emotional intelligence, recognition of emotions were analyzed. We identified as the seventh criterion the inclusion of skills concerning these areas and excluding results different from the objective.

Search strategies
This analysis was carried out in May 2020 and was conducted using two electronic databases: ERIC and PsycInfo. The ERIC research input was: (reading OR reading aloud) AND (emotion OR empathy OR emotional intelligence) AND (emotional development OR understanding of emotions) NOT (diseases OR brain loss OR disability OR autism spectrum OR syndrome OR disorder) NOT (second language). The PsycInfo research input was: (reading OR oral reading OR reading aloud) AND (emotion OR emotional development) AND (empathy OR emotional intelligence OR recognize emotion) NOT (diseases OR brain loss OR disability OR autism spectrum OR syndrome OR disorder) NOT (second language). This method of analysis has been used in previous research (Barquero, Davis, & Cutting, 2014;Salceda, Alonso, & Castilla-Earls, 2014). As shown in Figure 1, the ERIC search produced 558 results and PsycInfo search produced 220 results: 6 duplicate items were removed and the result was 772. The first analysis that was carried out was based on reading the title and the abstract of each article to identify those relevant to our criteria: 673 contributions were excluded because they did not meet the inclusion criteria, the remaining 99 articles were included. Subsequently a full reading of the articles was carried out, 33 articles that did not meet the inclusion criteria were excluded and 16 theoretical articles were used for the drafting of the introduction. The bibliographic research led to a final total of 50 studies that met the criteria for inclusion in the review of the descriptive literature.

Data Collection
The final articles (50) that have been analyzed are presented and discussed in the results. The information describing the area of intervention, the research design, the samples, the evaluation tools, the intervention and the outcomes is presented in Appendix A, B and C. Ten of the fifty articles are qualitative studies, three are case studies and two are neuroimaging studies. As regards the age group considered, Appendix A collects 13 articles including a sample of children (of which two articles had as participants children from kindergarten to primary school), Appendix B collects 24 articles with a sample of school age participants (from primary to upper high school), Appendix C collects 13 articles with a sample of adults. In line with the assumptions of the review, all the studies analyzed focused on reading and emotions variables as objects of investigation. A descriptive analysis of the results will be presented in relation to the reference age and the research design used in the studies; this description will be divided according to the three reference evolutionary periods: childhood, school-age, adulthood. Further information about research designs is reported in Tables 1, 2 and 3. Subsequently, the overall results related to reading components are presented. After that, the focus will be on the results related to emotional constructs.

Descriptive Results
The samples of the total 50 studies cover an overall age range from 18 months to 73 years old. These data are based on the youngest and oldest ages considered, excluding studies that did not indicate the age range.

Childhood
Regarding the Childhood sample (see Appendix A), out of a total of 13 articles, 2 studies were identified on participants in the nursery age group (Betawi, 2015;Drummond, Paul, Waugh, Hammond, & Brownell, 2014), 9 studies on a sample of participants in the Preschool age group (e.g. Aram, Deitcher, Shoshan, & Ziv, 2017;Baker & Rimm-Kaufman, 2014) and 2 studies following participants' growth from the preschool to primary school (Riquelme & Montero, 2016;Riquelme, Munita, Jara, & Montero, 2013). Of the total of 13 articles, 10 studies have a quantitative research design (of which 4 with control groups), 2 are longitudinal studies without a specific intervention and 1 study has a qualitative research design.

Adulthood
Finally, concerning the Adulthood sample (see Appendix C), out of a total of 13 articles, 9 studies involved participants from the university (e.g. Glenn, 2012;Hoggan & Cranton, 2015), and 4 studies enrolled samples of adult participants not belonging to an educational level (e.g. Altmann, Bohrn, Lubrich, Menninghaus, & Jacobs, 2012). Of the total 13 articles, 4 studies have a quantitative research design, 5 studies have a qualitative research design, 2 studies use both quantitative and qualitative research designs and 2 are functional neuroimaging studies.

Reading Components
The analysis of the studies allowed us to report some results on the reading components. Out of a total of 50 studies, 10 studies did not implement any specific reading intervention, they evaluated the association between reading and socio-emotional competence in other ways: Baker (2013) for example, explored the association between numerous family literacy activities conducted by fathers and mothers with their children (such as storytelling, reading books, number of books provided) and the subsequent development of socio-cognitive skills in children; in contrast, other authors measured reading skills (e.g. Morgan et al., 2012;Undheim et al., 2011), reading habits (e.g. Dylman et al., 2020;Tabullo, Jimé nez, & Garcí a, 2018) and language skills (e.g. Beck et al., 2012) and their association with emotions. The other 40 studies implemented different reading interventions. Some studies conducted intensive reading training with long-term interventions; these studies implemented reading practices in various forms and modalities, such as individual reading (e.g. LaForge, Perron, Roy-Charland, Roy, & Carignan, 2018), or reading aloud through mediated, shared or interactive modalities (e.g. Abdolrezapour & Tavakoli, 2012;Betawi, 2015;Lysaker & Tonge, 2013). Some research developed comparisons between various types of reading, in order to observe the most effective one on emotional variables (e.g. Riquelme & Montero, 2016). Other research used sessions of individual or shared reading, sometimes based on a single moment (e.g. Angantyr et al., 2016;Chisholm et al., 2017), other times based on several moments limited in time (e.g. Schapira & Aram, 2019). Finally, some studies have implemented educational workshops based on reading and discussion (e.g. Chisholm et al., 2017;Gordon, 2017).

Categorization of the Emotional Constructs
As mentioned above, under the label of "socio-emotional competence" are grouped many different concepts whose borders fade and are intertwined. This phenomenon is reflected by our results. Indeed, many articles use different names to refer to the same mental process (e.g. "social imagination" and "Theory of mind"); others focus on different, singular aspects of the same, wider, emotional concept (e.g. "cognitive empathy" vs "affective empathy"). In other cases, some wider emotional constructs overlap one another (e.g. the emotional intelligence, among its several aspects, comprehends also empathetic skills, overlapping the construct of "Empathy").
This "Babilonia Tower" is reflected even in the tools used to assess emotional components. Sometimes the same aspects are evaluated with different instruments. For example, beyond the qualitative studies, the most used tools for empathy are at least four: the Interpersonal Reactivity Index (IRI; Davis, 1983) (Pons & Harris, 2000) is the most used instrument (five studies investigating emotional knowledge). In order to better sort out our findings, we decided to decompose the wider emotional concepts and to focus on the singular socio-emotional aspects evaluated by each study, grouping our sample of 50 articles on the base of the same area of assessment (see Tables 1, 2 and 3). Referring to the theoretical background, we focused mainly on constructs associated with the emotional experience elicited by reading and the ability to infer, recognize and understand one's own and others' emotions and to relate to the environment in an adaptive and socially appropriate way. Thus, six labels were inferred: (a) Social-relational and interpersonal skills (prosocial behavior, cooperation, assertiveness, adaptability, social awareness); (b) Emotional and behavioral problems (such as anxiety and depression, emotional maladjustment, social problems, conduct problems, aggressive behavior, ADHD symptoms); (c) Emotional and behavioral adaptive competences (emotional regulation and expression, self-control, impulse control, emotional stability, self-esteem); (d) Ability to understand and sympathize with others' mind (empathy, theory of mind, understanding of emotions, social imagination, social understanding, sympathy, perspective-taking); (e) Emotional knowledge (emotional vocabulary; knowledge about nature of emotion; emotional recognition); (f) Emotional responses while reading (immersion experiences, flow, peak experiences, emotional processes and states while reading).

Reading and Social-Relational and Interpersonal Skills
The relationship between reading and social-relational and interpersonal skills is investigated by 13 studies. Among these, seven focus on childhood and six on the school-age period (four are set in primary schools, one in a middle school, one in a high school). All articles underline a positive association between reading and the development of interpersonal skills since the early months of life and up to the end of high school. Focusing on early childhood, two relevant studies (Baker, 2013;Baker & Rimm-Kaufman, 2014) and one longitudinal research (Ferretti & Bub, 2017) reveal a significant and positive association between home literacy activities (e.g. shared book reading) provided by parents and children's development of socio-emotional skills (such as prosocial behaviors, engagement and focused involvement while interacting with others) and interpersonal skills (forming and maintaining friendships, comforting or helping other children, etc.). Still regarding the domestic environment, four studies analyze parents-child interactions during shared book readings and found that kids are exposed to a greater and more variable emotional and mental states talk while reading picture books with their parents then when playing with them (Drummond et al., 2014). Moreover, the more the parents and the child discuss about social-emotional issues, the more the child will tend to show prosocial and socially appropriate behaviors (Aram et al., 2017;Drummond et al., 2014;Schapira & Aram, 2019), as shown also in the school-age period (Curtis, Zhou, & Tao, 2020). The same association between reading and interpersonal skills development is reported in the nursery (Betawi, 2015). As far as the primary school is concerned, reading interventions are proven to foster prosocial behaviors and attitudes (Jones, Brown, & Lawrence Aber, 2011), compliant and focused class behaviors (Brokamp et al., 2019), social awareness and metacognition (Venegas, 2019). With regards to middle school and high school, reading interventions promoted social-relational skills (Tijms, Stoop, & Pollock, 2018) and sociability (Abdolrezapour & Tavakoli, 2012). Emotionalbehavioral problems [2] Emotional and behavioral adaptive competencies [3] Understandsympathize with others' mind [4] Emotional Knowledge [5] Emotional responses [ immersion experiences, flow, peak experiences, emotional processes and states while reading.

Reading and Emotional and Behavioral Problems
The association between reading and the presence of emotional and behavioral problems is investigated by 8 articles: 4 studies are set in childhood, while the other four regard the school-age period (two in primary school, one in middle school, one in high school). As far as childhood is concerned, two studies with large samples (Baker, 2013;Baker & Rimm-Kaufman, 2014) and one longitudinal research (Ferretti & Bub, 2017) underline that home literacy interactions (e.g. shared book reading) and the presence of a large number of books at home have a negative and significant correlation with children's externalizing behaviors, conduct problems, negativity, and hyperactivity and impulsivity while interacting with others. Similarly, this result appears also in a nursery after a reading intervention (Betawi, 2015). As regards to school-age, poor readers in primary school result to feel more anger, sadness, and anxiety and perceive to be more unpopular than stronger readers (Morgan et al., 2012), while having reading difficulties at middle school predicts anxiety and depression, social problems, and aggressive and delinquent behaviors (Undheim et al., 2011). On the other hand, a two year reading program in schools reduces hostile attribution, depressive and ADHD symptoms and, after two years, aggressive behavior (Jones et al., 2011). Another five weeks reading intervention on high school hyperactive students brought a significant decrease in hyperactivity. However, this gain was not maintained after two weeks by the end of the treatment (Pang & Zhang, 2011).

Reading and Emotional and Behavioral Adaptive Competencies
Ten articles examine the relation between reading and emotional and behavioral competencies. Among them, three are focused on childhood and seven on the school-age period (five are set in primary school, one in middle school, one in high school). These studies underline how reading predicts a better emotional well-being and stability from the first months of life to the end of high school. The childhood studies show that domestic literacy activities foster self-control (controlling temper, accepting peer ideas for group activities, etc.) (Baker & Rimm-Kaufman, 2014). Reading interventions appear to bring to similar results in nursery (Betawi, 2015), while mediated reading programs are associated with better emotional regulation in the transition from kindergarten to primary school (Riquelme & Montero, 2013, 2016. Still regarding primary school, reading classes appear to improve self-confidence (Casserly, 2011;Venegas, 2019) and emotional stability (Brokamp et al., 2019); literacy interventions support emotional regulation (Jones et al., 2011); bibliotherapeutic book clubs foster emotional expression and self-confidence in middle school (Tijms et al., 2018), while one reading training with books with high emotional contents leads to better emotional intelligence in high school (Abdolrezapour & Tavakoli, 2012). Emotionalbehavioral problems [2] Emotional and behavioral adaptive competencies [3] Understandsympathize with others' mind [4] Emotional Knowledge [5] Emotional responses [6]  immersion experiences, flow, peak experiences, emotional processes and states while reading.

Reading and the Ability to Understand and Sympathize With Others' Mind
With 30 studies focused on it, the ability to understand and sympathize with others' mind is the most represented emotional construct in our review: five articles are focused on childhood, 15 on the school-age period (ten are set in primary school, two in middle school, three in high school), and 10 on adulthood. One of these studies covers the highest age period of our review: 17-73 years old (Koopman, 2016). The childhood studies reveal that story-book and reading training evoke theory of mind and empathic understanding in self, others, and even animals (Grazzani & Ornaghi, 2011;Karniol, 2012). Socio-emotional utterances generated during shared-book reading interactions are associated as well with empathy growth both in childhood (Schapira & Aram, 2019) and in school-age period (Curtis et al., 2020). Mediated reading fosters empathy even in the transition from preschool to school and in primary school Venegas, 2019), with stable effects even after six months by the end of the intervention (Ornaghi, Brockmeier, & Grazzani, 2014). Similar results were found in middle school (Stutler, 2011), while the use of graphic novels elicited empathy and historical empathy (Chisholm et al., 2017). As about high school, literature circles (Boatright & Allman, 2018) and reading training with social-emotional content books (Abdolrezapour & Tavakoli, 2011) promote empathy. Laboratory studies with adolescents and university students underline that engaging (Gerrig et al., 2016) and foregrounding texts (Koopman, 2016), as well as books with negative contents (Altmann et al., 2012), can foster empathy. As about adulthood, a study with university students underlines that women read more than men and show a stronger association between reading appreciation and empathy; instead, reading non-fiction (such as university textbooks) does not associate with empathic scorings (McCreary & Marchant, 2017). Similarly, textual difficulty is negatively associated with empathy (Junker & Jacquemin, 2017). On the other side, fiction reading significantly relates to theory of mind abilities (Tabullo et al., 2018). A series of qualitative studies on university students and pre-service teachers reveals that reading classes and bibliotherapeutic sessions promote empathic reflections and connections (Chamberlain, 2019;Hoggan & Cranton, 2015), empathic understandings of issues such as suicide and bullying (Pytash, 2013), and perspective taking and reflections about racism, reducing prejudice and stereotypes (Glenn, 2012; Slay & Morton, 2020). Only two studies do not find a significant increase in empathy over the course, respectively, of a 15 weeks reading workshop (Junker & Jacquemin, 2017) and a four week class reading intervention (Gordon, 2017).

Reading and Emotional Knowledge
The association between reading and emotional knowledge is examined by 13 studies: six refer to childhood and seven to school-age (six are set in primary school, one covers the last year of middle school and the high school). These studies investigate the emotional knowledge: with this term we refer to the knowledge about the emotions' nature, the emotional vocabulary and ability to recognize the emotional facial expressions and to label them with the right term. Except for two studies (Gordon, 2017;Laforge et al., 2018), all of these articles reveal that reading has a positive association with emotional knowledge: the more one reads, the more he/she will become an 'expert' of emotions. This result is extended to all the ages considered in these studies. More specifically, reading practice has a significant impact on the recognition of emotions and this evidence is found both in preschool ( (Beck et al., 2012;Kumschick et al., 2014;Lysaker & Tonge, 2013), with a stable effect in follow-up (Ornaghi et al., 2014). Emotional vocabulary is also increasing as a result of reading practice (Dylman et al., 2020;Kumschick et al., 2014;Ornaghi et al., 2014). Parents' socio-emotional discourses evaluated in conversations during and after reading in the family have an impact on children's understanding of the causes of emotions (Schapira & Aram, 2019). The study by Beazidou and colleagues (2013) highlights the techniques used by teachers to foster emotional knowledge during reading in preschool: they use movements and voice as mediators of emotional understanding and expression, urge pupils to find the cause of an emotion, compare the feelings of different characters and lead to imagine what will happen next. Emotionalbehavioral problems [2] Emotional and behavioral adaptive competencies [3] Understandsympathize with others' mind [4] Emotional Knowledge [5] Emotional responses [ immersion experiences, flow, peak experiences, emotional processes and states while reading.

Reading and Emotional Responses While Reading
The emotional responses elicited by reading is examined in 8 articles: 3 refer to school-age (1 is related to middle school and 2 to high school) and 5 to adulthood. Some of these articles (Stutler, 2011;Yussof, Jamian, Hamzah, & Roslan, 2013) underline that reading programs can evoke states of flow (a state of enjoyment in which sensations are heightened and thoughts and feelings fade, with a loss of time, place and self; Csikszentmihalyi, 1990), peak experiences (states of pure joy associated with moments of great truth or beauty, with a lack of awareness about time, space and self) and emotional responses.

Discussion
In do not highlight a significant association between reading and students' emotional scoring. These studies are conducted in naturalistic settings (daycare centers, a primary school and a university) with small samples and promote short term interventions (respectively, 3, 15 and 4 weeks) that foster empathic responses and emotional comprehension, but do not produce significant enhancements over time. Moreover, one of these studies does not find an effect over emotional vocabulary. However, these results may be due, as suggested by the authors, to the limited amount of time (that, unlike laboratories studies, may not have been enough in naturalistic settings) and to the use of qualitative and holistic measures (instead of the tools used in lab settings). Indeed, children's development of emotion comprehension may require longer time, as suggested by Pons and Harris (2000).
Besides these exceptions, the articles included in our review show a positive effect of reading on emotional dimensions all across several developmental stages. In fact, they cover a life span from the early months of life to adulthood, involving samples composed by infants, as well as school-age participants and adults. In contrast to the findings of Kozak and Recchia (2019), most of our articles (especially those with an educational perspective) focus on the school-age period, adding information on this still little studied evolutionary phase. Interestingly, adulthood samples are mostly composed by university students, while we found no study specifically focused on elderly or middle-aged samples. Thus, our review sheds light on a gap in the scientific literature and highlights the need of research investigating the relationship between reading and emotional dimensions even in other life stages with older participants.
Our review shows a beneficial effect related to books starting from childhood. The more a child is exposed to shared book readings, the more he/she will develop interpersonal and empathic skills and theory of mind abilities and will count on better emotional well-being and competence. Even more important, the more a child is provided with reading experiences, the less he/she will be likely to develop behavioral problems (e.g. externalizing behaviors, conduct problems, negativity towards parents). These associations are clear both in educational and family settings and are supported, in particular, by three studies (derived from the USA survey: Early Childhood Longitudinal Study, Birth and Kindergarten Cohort) that were based on very numerous samples and on quantitative and/or longitudinal research design (Baker, 2013;Baker & Rimm-Kaufman, 2014;Ferretti & Bub, 2017).
A possible way to portray these findings is suggested by a sample of 4 articles that, in both childhood and school-age participants, investigate the interactive processes elicited by shared book readings. Indeed, these studies show that reading interactions evoke more emotional utterances and mental states talk by both parents and children in comparison to other playing activities. Among these studies the longitudinal correlational research conducted by Curtis and colleagues (2020) must be highlighted, since its methodological reliability. Results show that parents' emotional talk during shared book reading was related to children's higher effortful control, that, in turn, was positively associated with children's higher sympathy and more socially appropriate behaviors. These findings are in line with the literature (Howe, Rinaldi, & Recchia, 2010), that demonstrates how this kind of interaction can foster children's better socio-emotional skills.
Similar results have been found in the school-age period and in adulthood. Scholastic settings are mostly taken into consideration and both quantitative and qualitative studies reveal that reading interventions (class readings, bibliotherapeutic interventions, reading aloud or silently, reading programs, literature circles, book clubs) and good reading habits support the development of relational abilities, emotional competence and empathic skills. Moreover, in both childhood and school-age period, there is evidence that these activities and being a strong reader enhance emotional knowledge, meaning the competence in recognizing and correctly labeling emotions and the awareness about their nature and causes. In particular two quantitative controlled studies (Grazzani & Ornaghi, 2011;Ornaghi et al., 2014) show that a 2 month reading training, based on emotional contents and followed by activities focused on emotional knowledge, promotes both empathy and emotion comprehension, with a stable effect over 6 months (Ornaghi et al., 2014).
One of a teachers' most implemented strategies is the use of multicultural literature, counter-narratives, or books with high socio-emotional contents to enhance empathy and reduce prejudice and intercultural conflicts. Even the use of historical graphic novels (Chisholm et al., 2017) has been effective.
Importantly, the environment is shown to have an influence over the association between reading experiences and emotional development, since educators and teachers can play an important role in fostering this positive relation. Three methodologically accurate studies (Riquelme & Montero, 2013, 2016Riquelme et al., 2013), conducted with a quantitative research design in naturalistic settings (in the transition from kindergarten to primary school), implemented a training based on mediated reading. Results, supported by comparisons with control groups, show that this practice fosters interactions between the reader and the group and promotes the development of theory of mind. Other studies suggest that also other kinds of interactions, such as group discussions among peers with or without an adult facilitator (with secondary school and university students), support the building of connections between the text and the readers' lives and foster reflection and awareness about one's emotions and empathic feelings. Moreover, teachers, through a supportive attitude and by providing good reading models, can improve students' school behaviors and emotional well-being and, indirectly, support their literacy skills (Brokamp et al., 2019).
This positive influence of reading over behavioral and emotional problems is underlined also by two articles (Morgan et al., 2012;Undheim et al., 2011) (one in primary and one in middle school), that show that being a poor reader and having reading difficulties correlate with the presence of anxiety, depression, social problems and aggressive behaviors. However, reading interventions have proven to be effective in reducing these manifestations. This result may depend on the finding that good readers also rely on better emotional regulation and expression that, in turn, may improve impulse control and foster psychological protective factors such as social success and greater self-esteem. Since a five week intervention addressing hyperactivity in high school students showed only a short-term efficacy (Pang & Zhang, 2011), long-lasting reading interventions should be implemented in order to obtain more stable behavioral benefits. Indeed, a reading intervention performed by Jones and colleagues (2011) showed a decrease in hostile attributions after one year, and a steeper decline of ADHD and depression symptoms over two years; however improvements in aggressive interpersonal negotiation strategies and social competencies were found only after two years. In line with the developmental cascades theory (Masten & Cicchetti, 2010), this may depend on the fact that interventions have to change, firstly, social cognitions and social-emotional dimensions in order to achieve subsequent changes in behaviors.
As suggested by this theory, these psychological and behavioral gains would lead to long-term effects, thwarting the onset of later delinquent behaviors and providing universal benefits in academic skills (Pang & Zang, 2011;Jones et al., 2011) and in other developmental domains (delayed onset of substance use and reduced academic disengagement/failure) (Eddy et al., 2003;Lochman & Wells, 2004).
This association between reading and emotional dimensions may be explained by the emotional responses elicited by reading. Reading fiction and narrative stories (Altmann et al., 2012;Hsu et al., 2014) activates neural areas related to theory of mind cognitive (dorsomedial prefrontal cortex, bilateral temporal poles and left posterior superior temporal sulcus) and emotional dimensions (left inferior frontal gyrus, mid-cingulate cortex). These articles analyze the connections between reading and emotions from a neuroscientific point of view, emphasizing the role of reading in activating neural processes linked to empathy. Other studies in lab settings underline that texts with socio-emotional contents, connotative language and with engaging features evoke strong emotional responses. In line with the literature (Oatley, 1999), first person narratives also promote a deeper immersion experience and emotional reactions.
Several studies also suggest that it is not the reading per se that promotes empathy and emotional growth: textual characteristics are important (Junker & Jacquemin, 2017). As shown by previous research (Batini et al., 2018;Kidd & Castano, 2013, 2017, fiction, unlike non-fiction texts, have resulted to foster empathy and theory of mind ( However, caution is necessary in interpreting these results, since most of our findings show bidirectional associations, not proving causal effects, meaning that more studies are needed to clarify this relationship. Despite this, the positive effects that reading interventions seem to have in reducing psychological symptomatology and enhancing empathy, as well as their activation of brain areas related to theory of mind, supports the implementation of reading programs in family and educational settings at every evolutionary stage in order to achieve better socio-emotional skills.

Limitations
Although this review offers a deeper perspective with regard to the analysis of the association between reading and emotions, we acknowledge that it is possible to identify two main categories of limitations. The first concerns the inclusion criteria. Generally speaking, we maintained a global focus of research and this choice could be seen as a strength, because it led us to collect a huge set of articles on which we developed some interesting comparisons. On the other hand, we had to deal with comparing studies with different focuses. For example, we pointed out that all empirical studies were included (see the second criterion), independently from research designs. Consequently, all studies have a different setting (e.g. home environments, labs, schools, universities or other services with an explicit educational purpose) and we did not make a discrimination in this way (see the fifth criterion). Still, studies investigated the reading component in a broad sense (see the sixth criterion). In some cases, studies (e.g. Aram et al., 2016;Drummond et al., 2014;Lysaker & Tonge, 2013) focused on a shared reading modality; in others (e.g. Boatright & Allman, 2018;Chisholm et al., 2017;LaForge et al., 2018) they concerned an individual reading modality; in others, reading habits and general language skills were assessed and therefore it was not implemented any intervention (e.g. Baker, 2013;Baker & Rimm-Kaufman, 2014). Thus, it was not possible to make a unitary comparison and grasp specific mechanisms of reading that were more associated with the emotional constructs investigated. In any case, this open-minded approach led us to collect some noteworthy empirical evidence, even for future research and educational implications.
The second type of limitations relates to the data collection process. Since we reported the complexity to consider a shared, unique definition of socio-emotional competence, we used several keywords (see Search strategy) in order to account for the complexity and multidimensionality of the emotional constructs and prevent studies with different constructsbut inherent to our research focusfrom being excluded. As a matter of fact, we analyzed a huge number of articles that focused on the same emotional constructs, but also on socio-emotional competencies with different or more hues of meaning. This, although has provided a great deal of data, has often led us to compare different constructs, or similar constructs but defined in different terms in individual studies.

Conclusion and Future Directions
We are aware that the analytical work carried out can not be considered exhaustive. Despite the limitations of this review and the complexity inherent in the variables of reading and emotions that are the subject of this research, we consider the path of this analysis and the results found very meaningful to better understand the association between reading and emotions. Our broad investigation may inspire more specific foci of research: for example, the categorization we have carried out on the broad construct of socio-emotional competence, subdividing the studies into six main areas, could provide a starting point for future reviews focusing on more particular aspects; the same mechanism could also be carried out for the reading components, developing surveys aimed at evaluating interventions based on a specific type of reading, such as mediated reading. Moreover, as shown, the effects of interactive reading (Lysaker & Tonge, 2013;Riquelme & Montero, 2016) are proven to improve socio-emotional competence; so it may be interesting to argue about the implications that the continuative implementation of this practice in the schools may have over long-term periods. Finally, as discussed before, the impact of reading on adult and elderly populations emotions may be a particular matter of interest when formulating new research may be a particular matter of interest, since we have collected only one study which takes into account older participants (Koopman, 2015), even though we did not set a restriction about this criterion.  "Projective" test with carbon paper; Structured interview created ad-hoc.

Reading aloud session
All children showed more excitement after hearing the story.
Those who had empathic understanding for the protagonist's situation showed significantly higher levels of excitement than those who had not such understanding. This excitement depends on the understanding of the protagonist's psychological situation; participants with empathic understanding were also more likely to understand the negative affect associated with cartoons in which other children were the actors. They also understood better that they too, in the same situation, would have experienced a negative affect. This understanding brought to predictions of negative affect for other children and themselves in general contexts. No intervention Children exposed to more preschool routines (bedtime routine, reading routine, telling stories, singing songs, dinner as a family, and dinner at a regular time) later exhibited greater gains in socio-emotional skills in comparison to children exposed to fewer routines. All the six routines were individually and significantly associated with the outcomes.  (Caldwell et al., 1984). After reading, decontextualized references prevailed both in mothers (p<.001) and in children (p<.05). These references are associated in literature to better socio-emotional development.
References by mothers and children to the book's socio-emotional contents were associated with greater socio-emotional understanding (p<.05) and prosocial behavior (p<.05) in children. When mothers and children used more socio-emotional contents, children showed greater social understanding (p<.05) and more prosocial behavior (p<.05) than children belonging to dyads in which less socio-emotional contents were used. Emotional skills in the mediated reading group increased compared to the traditional reading group. Recognition of emotions: significant effect in the interaction between "Treatment" and "Evaluation" (p<.01) and in the interaction between "Degree" and "Treatment" (p<.05). Empathy: significant, although marginal, fourfold interaction between the factors Evaluation, Treatment, Degree and Sex (p=.07). Emotional regulation: it shows a significant effect on the interaction "Treatment" and "Evaluation"(p<.01), as well as in the interaction between the factors "Degree" and "Treatment" (p<.01).

Reading training 5 months (two sessions on
week of 25-40 minutes) The intervention had a significant impact on children in the mediated reading group, where facial recognition skills of emotions (Childhood group: p<.01; Primary group:p<.01) and empathy (Childhood group: p<.01; Primary group: p<.01) are increased in comparison to the traditional reading group. Semi-structured interview created ad-hoc.

Special
Reading Program Children were aware of their inadequacies, particularly when they read in front of their peers in class. Almost half of the children felt that their peers were aware of their difficulties, but were empathetic towards them. In some cases, children felt embarrassed, especially when others laughed that they were losing their mark in reading. They also reported that they had improved their self-help strategies and referred to their determination to solve problems. Girls showed more empathy than boys (p< .01). This main effect was qualified by a significant interaction effect between gender and condition in the group (p< .05). However, in the experimental group, it was not statistically significant. Children who own a dog showed more empathy than children who do not.
In this case, a significant gender effect was recorded for girls. Age/Empathy: negative correlation (p< .07), the youngest show more empathy). Joy/Empathy: negative correlation confirmed (p< .01). This suggests that human education programs can reduce gender differences by increasing boys' empathy.  (Bonino, Lo Coco, and Tani, 1998 (Eisenberg et al., 1995;Harter, 1982).

Reading sessions
(Two waves at a distance of 1,5 years; laboratory evaluation 2.5 -3 hours) ET was positively correlated (rs=0.12 to 0.74). The use of positively emotional words by parents and the quality of ET were positively correlated with self-control (rs=0.19 and 0.17) and children's understanding (rs=0. 16 and 0.15) reported by teachers. Parents' use of positive emotional words was positively correlated with children's social behavior reported by parents in W2 (r= 0.17). Positive correlation between parents' ET in W1 and children's self-control in W2. Even excluding social behavior in W1, the ET of parents in W1 positively predicts children's self-control in W2. Self-control in W2 was positively associated with social behavior in W2.  Brokamp, & Smits, 2012).

No intervention
Reading fluency at the beginning of the year is significantly associated with task-oriented behavior (p<.001), emotional stability (p<.05) and class behavior (p<.05) at the end of the year. Effective teachers' book introductions are a useful factor to improve students' behavior and emotional stability (p<.05).  , 2005).

No intervention
Third grade children who were weak readers (N= 331) subsequently reported feeling angrier (32%), sadder (34%) or more unpopular (20%) than peers who were not weak readers (10%); in particular, feelings of sadness, worry and anxiety were frequently reported. Girls reported feeling significantly less angry and unpopular than boys. Children who were weak readers in third grade are more than twice as likely to report feeling angry or distracted in fifth grade.  (Dunn & Dunn, 2004).  Heyerdahl, 1998).

No intervention
The total score for problematic behavior increased significantly from T1 to T2. The RD group showed a higher number of problems than the non-RD group in both T1 and T2 (p< 0.001). 53% of adolescents with DD in T1 also reported them in T2. Externalizing problems also increased among those who did not have DD. There was no difference between those who had received a special education and those who had not, both in T1 and T2. Reading difficulties are a significant predictor of social problems one year later. Those who received special education and experienced DD in T1 showed more social problems than those who did not. From 2 nd week: the scores increased in line with the average of the class and the average hyperactive behavior decreased significantly. By the 3 rd week, the 3 participants scored twice as high as the baseline scores. In the last week, the students showed only a quarter of the initial hyperactive behavior. After the intervention, the number of correct responses of the participants was inversely proportional to the number of unwanted behaviors. The performance in the last 3 weeks was better than the average of the class. In the generalisation phase (interruption of the intervention), performance returned to be similar The exposure of EFL students to literature response activities positively influences their EQ: the experimental group had a statistically significant score compared to the control group (p<0.05). A difference also emerged between the two groups in the relationship between reading and comprehension activities (p=0.000) and between emotional intelligence and text understanding (r=0.660, p=0.000). Educational laboratory H1. The more readers liked the protagonists of the story, the higher was the measure of "escape rooting" (i.e. the more readers hoped for his escape). Readers who found the protagonist a moral character had higher expectations for his escape (p< .01), but these expectations were not confirmed by a correlation with sympathy for the character (p <.10). H2. Individual differences in empathy were related to the judgment of sympathy and morality, but the evidence is partial; readers who reported higher levels of empathy tended to report less sympathy for the characters in the text. H3. The involvement was positive and significantly correlated to the positive outcome (p=< .05). H4. The correlation between empathy and involvement in reading is positive and significant (p< .001; p< .05). H5. The correlation between higher involvement and higher enjoyment was confirmed (p< .001).

No intervention
Girls, in comparison to boys, reported higher levels of empathy, were involved in more recreational readings and showed more appreciation for reading. The perspective taking ability was significantly associated with the appreciation for reading in the general sample (p <0.005). Girls (but not boys) who reported higher levels of reading appreciation also had higher levels of empathy (p <0.01). In males the amount of non-recreational reading is related to less empathy (p <0.05). Three results: 1) "Previous life experiences": many participants knew a person who committed suicide and expressed their willingness to understand why it happened; 2) "the efferent and aesthetic continuum": the students used reading to gain practical and theoretical information on suicide and bullying issues; 3) "transactions as future teachers": the students developed empathy for the characters in the books and began to understand the reasons and feelings that can bring to suicide. They connected this experience to their need, as future teachers, to be empathetic with students. Educational laboratory Three main results: "promoting change", "new perspectives" (awareness of new and different ways of perceiving the world) and "promoting critical reflections" (about themselves and the world). Other additional results concern: connections between history and personal life experiences, emotional stimulations by reading (and feelings of being able to express themselves freely) and identification of role models in the book characters. Stories with first-person pronouns lead to higher levels of general immersion than the stories with third-person pronouns (p< 0.05). The pronoun type has no significant influence on the ranking (p= 0.06). There is a relationship between the scores of the immersion questionnaire and the appreciation of a history (Evaluation: p < 0.001; Classification: p< 0.001). People prefrontal cortex and mPFC Bilateral anterior insula and right posterior cingulate cortex showed a stronger coupling with mPFC for individuals with a stronger tendency to feel concern for other people (p < 0.05).  (Davis, 1980(Davis, , 1983.

No intervention
Significant associations between RMET performance and: reading habits (p=.03) and personal discomfort (p= .015). For trait empathy, fantasy scores were higher in subjects who reported more frequent narrative reading times (p <.001) and started recreational reading earlier (p= .014), but tended to decrease with age (p <.001). RMET performance was significantly associated with ART scores within the male group (p= .004), but this effect was not observed in women.
The association between Mind Theory performance and narrative print was significant for men (p= .002) but not for women (p= .421).

Reading session
The foreground condition is able to evoke more emotions than other conditions. No difference in emotions after reading (p=0.038) (significant data for the first group). Empathic understanding is greater in the original text and especially after reading. In general, the degree of reflection depends on the general characteristics of the person, whether he/she presents empathic traits or not. A general type of reflection is aroused in subjects who also present a greater empathic understanding (p= .034). Cognitive and affective empathy were strongly influenced by the perception of textual difficulty: higher levels of openness, awareness of multiple perspectives, awareness of the cultural framework and affective mental attitude coincided with the perception of a more interesting and intriguing text (p < .05). Significant relationships between empathy and textual difficulty (negative; p = .05) and the ability to write (positive; p <.001). Creative writing tasks stimulate more empathic responses than tasks that require analytical writing (p <.001). Empathy did not appear to increase over time. The reading conducted by teachers, aimed at promoting empathy towards the characters, has generated racial considerations driven by the desire to weave universal connections that unite all human beings.

Chamberlain, 2018
The experience of older adults who participate in a bibliotherapy/poetr y group in an older adult inpatient mental health assessment and treatment ward Being in a reading group had a therapeutic effect; participants were able to empathise with others and share opinions and experiences. This was facilitated by a safe and tolerant environment. Using phenomenographic techniques, the act of reading aloud could be seen as a development of trust in the participants, as an activity through which they could find their voice and identity.