Subverting the selective tradition? A self‑exploration of text selection in pre‑service teacher education

Children’s literature can be both a reflective mirror to readers’ lives and a window to new worlds, making teachers’ selection of texts for students an important professional activity. Researchers have consistently found that teachers’ choices of literature contain limited representations of ethnicities, cultures, and disabilities, and reinforce gender stereotypes. Teacher educators, when working with pre-service teachers in university settings, utilize children’s literature for literary, critical, and cultural pedagogical purposes. However, teacher educators rarely interrogate their text selections to explore patterns of representation, identity, and power. This paper describes and discusses a self-study into five children’s picture books selected for modeling aspects of early literacy teaching in a pre-service teacher education unit. Critical content analysis was used to explore representation, identity, and power in the texts. The analysis showed some connections with trends found in research into early years and school teachers’ selective traditions in the use of an older text and two texts without human characters. Other findings differed; agentive female characters, together with some variation of social, cultural, and ethnic groups and lives, were depicted in the three texts with human characters, likely because of the author’s own bias towards expanding representation in texts. This article reports on an example of one teacher educator’s selective literary tradition and shows how the texts used in education settings represent windows to specific worlds rather than standing in for “diversity.” It makes suggestions for other educators interested in interrogating their text selections and invites dialog with other educators about representation, identity, and power in the texts they teach with.


The importance of children's literature in education
Children's literature is a ubiquitous and valuable part of early years and primary education.Literature "illuminate(s) the human experience" (Saxby, 1991, p. 4), and provides encounters with new language, possibilities, ways of being, escapism, and pleasure (Nodelman, 1996).Literary texts provide a means to understand the self as well as accessing new ideas and worlds (Fellowes & Oakley, 2020); they can "allow for rich linguistic, visual, and conceptual input" (Green et al., 2022), promote critical thinking, and develop argumentative competence (Chen et al., 2021).As Bishop (1990) first identified, texts can be windows into worlds, cultures, and ways of being, they can provide sliding glass doors that enable the reader to enter and experience those worlds, and texts can be mirrors that reflect, affirm, and celebrate the reader.
Educators' choices and uses of children's literature are both important and contentious, as the texts school and other teachers choose are mirrors to specific ways of being and windows that look out on some worlds but not others (Bishop, 1990;Glazier & Seo, 2005).Martin and Spencer (2020) explain that children's literature is not neutral but is "embedded within ideological frameworks-those deleterious to marginalized identities and communities, and those that are affirmative and inclusive " (p. 388).This suggests that the selection of texts for classroom use is an important task (Darragh & Boyd, 2019), as teachers may choose texts that reinforce negative stereotypes or those that celebrate diversity.Yet substantive research has raised concerns with the limited nature of published children's literature and of teachers' text selections.For example, García (2017) discusses how the field of children's literature is dominated by monolingual Anglo-English texts and research, with writers and researchers from non-white groups marginalized; and Adam (2021) and Adam and Harper (2021) describe how early years teachers typically chose texts that reinforced dominant stereotypes and contained few examples of diverse ways of being.
Like other teacher-educators (Christ & Sharma, 2018;Daly & Blakeney-Williams, 2015;Martin & Spencer, 2020;Paley & Jipson, 2000;Simpson, 2016;Singer & Smith, 2003), I aim to subvert such trends by exposing pre-service teachers to diverse children's literature in the hope that they will in turn work with texts that celebrate a range of lives and ways of being.This paper will disseminate a self-study research project in which I explored the extent to which I espouse my values about text diversity in my own teaching context.

It addresses the research question:
To what extent are diverse lives and stories represented in a set of texts used in a preservice teacher education unit?
This paper begins with a literature review of research on the selective literary tradition in education and publishing settings, together with research on uses of children's literature in teacher education programs.

The selective tradition: representation in children's literature
The selective tradition in children's literature refers to the characters and lives typically depicted in books in schools, libraries, book awards, and publishers' lists.It has been explored over many years with stubborn trends, discussed in the subsequent paragraphs, found across historic and recent research in Australia and internationally.

Whose stories are portrayed in children's literature?
Researchers have explored who is represented in books for children, particularly in major roles like the protagonist, and how they are framed within the text.Research in the past has identified underrepresentation or negative stereotypes of ethnic minorities, women, and people with physical disabilities (Jipson & Paley, 1991;Luke et al., 1986;Smith et al., 1987), and social class Jones (2006).More recently, Ferguson (2019) identified that the most popular children's books published in 2018 were dominated by white, male characters, with few characters from minority ethnic groups, one character with a disability, and no same-sex parented families.Adam (2021) and Adam et al. (2021) identified similarly monocultural representations in the children's literature selected by early childhood educators and pre-service teachers in Western Australia, noting that the few representations of minority groups were "othering."The UK's Center for Literacy in Primary Education "Reflecting Realities" survey (2020) found a slight increase in both the proportion of minority-ethnic characters in children's books published between 2017 and 2019, and greater diversity in how these characters are represented.However, the five percent of minority-ethnic characters identified in the 2019 books sits in sobering contrast to the 33.5% of school children belonging to a minority-ethnic group in the UK in 2019.
Children's book award winners are often utilized by school and other teachers because of their status and perceived quality (Caple & Tian, 2021;MacKay et al., 2017).Examinations of award-winning books in English have uncovered privileging of a narrow sub-set of lives and ways of being.For example, Caple and Tian (2021) explored the shortlisted books in the early childhood category for the Children's Book Council of Australia [CBCA] book of the year awards over a 20-year period, finding an even gender split in boy and girl main characters, but that 89% of the main characters were white.The authors note that this large sample of texts "fail[s] to portray the varieties of ability, sexual identities, and ethnic backgrounds that constitute Australian society today" (p.187).

Stereotypes and authenticity
Researchers have conducted plot analyses to interrogate the roles afforded to specific groups.For example, Luke et al.'s (1986) analysis of Australian pre-service teachers' text selections showed that nearly three-quarters of the texts had a problem that was resolved by a male character, and that female characters' roles within the texts were limited and stereotypical.Such findings are repeated in contemporary research; for example, Ferguson (2019) identified that limited roles were afforded to female, ethnic minority, or disabled characters in the most popular picture books published in the UK in 2018.In US and Australian contexts, Adam and Harper (2021) explored characters' gender identities across 44 early childhood educators' children's book selections; concluding that 85% of the texts portrayed gender traditional stereotypical roles.Text era can influence representations, with older texts being more likely to contain stereotypical or offensive depictions of minority or marginalized groups (Adam, 2022;Ferguson, 2019;Madsen et al., 2021).
Researchers have also identified that author identity can impact on the extent to which texts tell authentic and nuanced stories or conform to reductionist stereotypes (Bishop, 2003).For example, Trousdale (1990) identified significant differences between Black and white authors' portrayals of the Black experience, with Black subservience and acceptance of discrimination a core theme in texts by white authors.In the Australian context, non-Aboriginal authors have depicted Aboriginal lives and historic events (e.g., Gwynne, 1998;Marsden & Tan, 1998) though their authority to do so has, at times, been contested (Gray, 2004;Leane, 2016).Gardner (2020) observed that children's books with Black characters are much more likely to be written by non-black authors and highlighted the ways in which children's literature continues to value whiteness and white ideological constructs.Ferguson (2019) similarly associated a lack of diversity in authors and in character representations when discussing representation in 2018's best-selling children's books.
The recurring trends reported in this section appear to arise both from underrepresentation of diverse groups in published and award-winning texts, and from teachers' selection of texts.Indeed, these two elements are likely linked.These findings suggest that many students may be exposed to a diet of predominantly monocultural texts that reinforce stereotypes unless educators explore their biases and consciously select texts with diverse representations and ways of being.

Texts as sites of learning in pre-service teacher education
Teacher education is a site where pre-service teachers can explore a range of children's literature and consider the texts they will use in their own practice.Many teacher educators emphasize the value of teaching with children's literature (Freeman et al., 2011;Martin & Spencer, 2020;Meacham & Meacham, 2014;Simpson, 2016) while some explore its potential for eliciting growth in pre-service teachers.For example, Singer and Smith (2003) and Martin and Spencer (2020) explored pre-service teachers' responses as they engaged with Black and multicultural literature.Both studies found that participants' identities influenced their interpretation and connection with texts.Martin and Spencer further identified that a key tension for pre-service teachers was teaching with texts that might be considered controversial.Christ and Sharma (2018) modeled culturally relevant text selection and pedagogies with pre-service teachers, who in turn selected texts and designed learning experiences for students.Their participants were able to match texts in accordance with students' ethnicity but found it challenging to consider other dimensions of culture and identity; and were reluctant to discuss race and social justice with their students, limiting the opportunities for critical consciousness.While children's literature titles were identified in these three studies, the research focus relates to transforming pre-service teachers' practices and beliefs.The texts chosen to achieve these aims are not interrogated.
Other research has explored uses of children's literature for particular curriculum purposes.Bradbery (2013) describes how children's literature helped to develop sustainability knowledge with pre-service teachers; interrogating the environmental values portrayed in teaching texts but not other aspects of representation.Freeman et al. (2011) and Daly and Blakeney-Williams (2015) discuss how children's literature has broad benefits for pre-service teachers, including its potential to enhance cross-cultural competence and support inclusivity.They provide illustrative examples of texts but not an analysis of these texts' content.Simpson (2016) describes several advantages in using literature in pre-service teacher education including: developing an aesthetic appreciation for literature; developing curriculum content understandings; and providing examples of quality writing.However, Simpson does not unpack educators' book choices to explore how they support learning in these areas, nor consider what kinds of representations they offer.Meacham and Meacham (2014) engaged in self-study in their uses of children's literature in teacher education.Their study differs from those reported above in that the researchers are explicitly reflective about their text choices and use their connection between text interpretation and lived experiences as a basis for teaching.They share personal reflections about texts chosen but do not engage in text analysis.

Research gap
This literature review has shown that there is a body of research highlighting a selective literary tradition in early years, school, and pre-service teachers' text selections, books published for children, and award-winning children's books.Teacher-educators' research highlights the potential of teaching with children's literature, together with the importance of exposing pre-service teachers to diverse texts.However, no research was found that subjected teacher-educators' text choices to the types of content analysis and critique found in studies on teachers' text choices (Adam, 2021;Adam & Harper, 2021;Jipson & Paley, 1991;Luke et al., 1986;Smith et al., 1987).Analysis of one's own text choices is important as it provides a means for teacher educators to explore representations and consider the extent to which their chosen literature matches with their aims and values.Literature is not neutral (Martin & Spencer, 2020), and without critical analysis of the texts chosen for teaching, a limited range of authors' representations and ideologies may dominate (Iyer & Ramachandran, 2020).Moreover, in a research climate where school and early years teachers' text choices are often critiqued, it seems fair and equitable for teacher educators to share, examine, and critique their own text choices.

Methods
This research draws on Self-study of Teacher Education Practices [S-STEP] (Hamilton & Pinnegar, 1998;Loughran, 2004), a process by which teacher educators can explore, learn about, and transform their practice.S-STEP is not a method with pre-defined steps, tools, or analytical techniques; though many S-STEP studies begin with a living contradiction or possible gap between the teacher educator's values about teaching and their actual practice (Loughran, 2004).Rather, the teaching and research context drives the selection of research methods (Loughran, 2018).

Research context
My teaching and research context was an undergraduate, pre-service teacher education unit of study at a mid-sized Australian university.The unit focused on early reading development and provided foundational information about oral language, phonological awareness, phonics, vocabulary, fluency, and comprehension.It was part of early childhood and primary teacher education courses.Learning activities were delivered across an 11-week period in the first half of each year.These comprised a weekly 1-hour lecture which provided an overview of core concepts and research findings and a weekly 2-hour seminar in which teacher educators and pre-service teachers engaged in activities that developed practical understandings.They explored each of the topics, how they typically develop, and how they can be taught and assessed.
While the content and scope of education units is designed at the university level and approved via external accreditation, the specific resources used for modeling and demonstrating teaching ideas are typically chosen by academic teaching staff (Paley & Jipson, 2000;Simpson, 2016).Children's literature-a common resource in schools-is also used in pre-service teacher education (Christ & Sharma, 2018;Martin & Spencer, 2020;Paley & Jipson, 2000;Simpson, 2016;Singer & Smith, 2003).In both school and higher education environments, texts are selected for: the ways in which they exemplify a specific aspect of the English curriculum, for example, genre (Opatz & Nelson, 2022), language devices (Sowa & Lacina, 2011), or visual elements (Small & Callow, 2021); their subject content knowledge (Allan, 2014;Bradbery, 2013), or their capacity to develop critical thinking (Christ & Sharma, 2018;Martin & Spencer, 2020;Turner et al., 2023).The unit drew on and utilized a range of practical and academic resources, including textbook readings, media clips, developmental continua, practical examples, teaching resources, assessment tools, together with children's literature.
Children's literature was used to model aspects of language and literacy learning, often using the gradual release of responsibility model (Pearson & Gallagher, 1983), where the teacher educator would read the text to pre-service teachers, pausing to identify and discuss one or more specific aspects such as a phonological element, type of sentence, or vocabulary term, as they would when teaching students.During reading, the teacher educator and preservice teacher would discuss that aspect (for example, rhyming words or action verbs) and the potential of the book for developing students' knowledge and skill in that area.After reading, pre-service teachers would discuss the text's broader potential for teaching and learning, and respond to open ended discussion questions, such as "how might students with different life experiences understand this text?"The texts chosen needed to exemplify one of the core skills taught in the unit, such as phonological awareness or vocabulary development, and be broadly engaging, interesting, and worthy of discussion.I selected texts drawing on my prior knowledge and experiences as a school teacher, through browsing at public and university libraries, and reading children's literature blogs, book reviews, and book awards media.

Living contradiction
It was in relation to the children's literature I had chosen for use in the unit that I identified a possible living contradiction in my teaching practice.As a school teacher and a teacher educator, I aim to teach with literature depicting a range of lives and ways of being.I explicitly share this value with the future teachers I work with, highlighting dominant trends in the stories told and absent in literature for children.I asked myself, to what extent was I modeling the use of such texts in this unit of study?The identification of this living contradiction became the first step in my research process, which is shown below and elaborated on in the subsequent paragraphs.

Research process
1. Identify living contradiction.2. Choose appropriate methods of enquiry.I began with an audit of the children's literature used across the unit.I defined children's literature as texts with literary and aesthetic value, with the capacity to be read both for pleasure and for pedagogical purposes.This meant that readers designed to support students with their reading development were excluded from the audit due to their narrow pedagogical purpose, as were texts that formed part of a formal reading assessment.As the early weeks in the unit were focused on understanding early reading skills and their typical development, other resources were used in those weeks.Seven pieces of literature were shared across the second half of the unit, when teaching and assessment activities were explored.
Of these texts, five were explored in their entirety and were used for both teacher modeling and student exploration and discussion.This excluded: The birthday cake mystery (Tjong-Khing, 2012), as only a single page spread from this text was used to explore storytelling and language samples; and 123 turtles and geckos (Yunupinu & Stubbs, 2021) which was shared by tutors as an additional example of a text for interacting and developing language with young children, but not explored by pre-service teachers in groups.The resulting group of texts comprised five children's picture books, each written in English (see Table 1).Each of these texts was used in a single seminar, and some were also introduced in that week's lecture.
The next stage involved reflective journaling about the five selected texts.I asked myself, what was my motivation in including each, and how did each text's (a) literacy teaching potential, and (b) depiction of particular ways of being influence my decision?For example, my motivation for selecting Press here was its potential for teaching about verbs and commands, together with its innovative design which uses haptics to encourage students to interact with the text.These journal entries are summarized in Table 1 in the mentor text purpose and personal motivation columns.
I engaged in a review of literature on school and other teachers' text selections for students and examined how others had evaluated texts.This review supported my development of a set of criteria for analyzing my texts.These criteria encompassed text and author information, and text content.Each criterion is briefly described in the following section, together with citations showing the academic literature it draws from.

Evaluative criteria
The texts were read, viewed, and annotated across two and a half months in mid-2022.This period directly followed the use of these texts for teaching in early 2022.

Text and author information
Author and artist public biographies from professional webpages or publishers were searched for information about identity-including ethnicity, nationality, and other selfidentified traits.The purpose of this was to examine the diversity of authors and illustrators whose work was explored in the unit (Ferguson, 2019;Luke et al., 1986), and to consider whether authors and artists were writing about and illustrating cultural groups and activities to which they belonged (Gardner, 2020;Madsen et al., 2021;Trousdale, 1990).The year of initial publication was recorded as a measure of how recent each text was (Adam, 2022;Ferguson, 2019).Publishers' websites and children's book lists were searched for mention of whether each text had been shortlisted for or won any children's book awards.

Text content
Key details about the characters and their role/s in each text were gathered from the text and images of each book and recorded.These included: ethnicity, culture, gender, and family makeup (Adam, 2021;Adam & Harper, 2021;Jipson & Paley, 1991;Luke et al., 1986).
While the text and author information consisted of factual statements drawn from online sources, the text content analysis used multimodal exploration of the linguistic, spatial, and visual information (Callow, 2013;Cazden et al., 1996;Janks, 2010;Kalantzis & Cope, 2012) in each text.For example, some information about characters was directly stated in the text (who the characters were in Who sank the boat), some told through the illustrations (the ethnic make-up of the family in Things in the sea are touching me) and some aspects were inferred from clues in the text and images (that the representations of the 26 th of January offered by his daughter's school angered Dad in Day break).A summary of each plot and setting was recorded, including information about character agency (Adam & Harper, 2021;Gardner, 2020;Luke et al., 1986).

Text analysis
The analysis explored these texts as sites of representation.Information for each text in relation to each criterion was recorded (see Table 1).Some of this information could be analyzed using descriptive statistics, for example, the number of texts with a female protagonist or the number of texts where the author was from the ethnic group represented in the story.Critical content analysis (Beach et al., 2009;Bradford, 2017;Short, 2017) was used to analyze the texts' content.Theoretically, this work draws on social justice and sociocultural perspectives on literacy and literature, namely, that representation matters and that there is potential for alienation when students do not see people, lives, and ways of being like their own in their texts and resources (Adam, 2021;Gardner, 2020;Jones, 2012;McNair & Edwards, 2021).I annotated the pages of each text with research notes.Examples of these are shown in Fig. 1 which explores a page spread from Day break.This shows aspects of visual analysis (for example, vectors, gaze, and intertextuality), linguistic analysis (for example, phrasing), Dated stereotype of teacher standing in for a white settlercolonial view of Australia.

Teacher and Cook are above students but girl raises herself to their level and meets their gaze.
English Curriculum link: "build literal and inferred meaning about key events, ideas, and information in texts" ACELY1660 (Australian Curriculum English, n.d.).Explore with reference to background knowledge and its role in comprehension.

Broad theme of asserting Aboriginal survival and resisting the settler-colonial narrative. Short phrase broken up for emphasis and stretched across page, alluding to time.
Personal objects, suggesting a whole rounded person in contrast with the teacher stereotype.
Warrior stance, embodied gesture rather than just words.Intertextuality (e.g.Adam Goodes).Back turned against colonialism.and broad thematic messages (asserting Aboriginal survival).The research notes were analyzed in tandem with the text and illustrations using a social justice lens.The aspects I sought to interpret in each text were: representation and identity-who and what was represented in each text, and how were they represented; and power-to what extent were characters agentive?The analysis was a recursive process with constant comparison between my notes and the texts themselves.The trends and patterns resulting from the analysis were then compared to patterns identified in earlier research on selective traditions in literature for children.Loughran (2004) explains that "a defining feature of self-study is that it is available for such public critique and dissemination, rather than solely residing in the mind of an individual" (p.26).This self-study becomes a shared task through publication with the aim to connect and collaborate with other researchers interested in exploring their own text selections.

Findings
The following section begins by summarizing each text in relation to the evaluative criteria.These summaries expand on the information provided in Table 1, including an overview of the text content analysis.
Who sank the boat (Allen, 1982) is a narrative story of anthropomorphic animals-a cow, a sheep, a donkey, a sheep, a pig, and a mouse-in a rural, pastoral location.While gender is suggested with some gendered pronouns and in some images (a cow with an udder and a sheep knitting) it does not play an obvious role in the narrative.The story tells how the animals get into a boat one by one until the boat sinks.The text uses a mix of third-person narration and second person questions to the reader, finishing with "You DO know who sank the boat."Agency lies with the reader who is invited to know more about how the boat sank than the unsuspecting animals.
Press here (Tullet, 2011) uses instructions and haptics to direct the reader to engage with dots on each page in a variety of ways.Commands and praise are prominent throughout the text and each page shows changes or a new pattern.The commands are rich with verbs, adverbs, and conditional language.While not a traditional narrative with a beginning, middle, and end; the excitement rises and falls, with the text ending by cycling back to the start.
Yobbos do yoga (Gwynne & Joyner, 2013) is a light hearted narrative of a girl, her anxious, yoga loving father, and their new yobbo neighbors-all illustrated as Caucasian.Dad is skeptical about the yobbos and grumbles about their loud music but his daughter is curious, visiting the yobbos to retrieve a ball and asking them to turn their music down.Dad's assumptions are challenged when his car will not start and is fixed by the yobbos.The story ends with Dad and the neighbors enjoying each other's hobbies-yoga and air guitar.Agency resides with the girl whose lack of prejudice results in positive outcomes for herself and her Dad.
Things in the sea are touching me (Keegan & Stapleton, 2019) narrates a day at the beach for Ma, Mum, and their daughter.Mum and child are illustrated as Māori or Pasifika and Ma as Pakeha.The text uses rhythm, rhyme, and alliteration to tell of the ocean flora and fauna that "touch" the girl and give her a start, culminating in Mum playing a trick on Ma.Scientific terms and explanations are an unobtrusive part of the story-for example, what salps are and how mangrove forests grow.Agency is not strongly shown by any character though it is Mum and Ma who explain what each thing "touching" the girl is and Mum who tricks her partner.
Day break (McQuire & Chun, 2021) explores Invasion/Survival Day (26 th January) from the perspective of an Aboriginal girl who is navigating the different messages she receives about Australia and colonization from friends, school, her Dad, and her Nan.The main characters are Aboriginal and live in an extended family grouping comprised of Nan, Dad, and the girl, in an urban setting.Dad and Nan disrupt the taken-for-granted information about Australia's history provided by dominant Anglo-Celtic Australian histories and conveyed through the child's teacher, offering nuanced perspectives and ways of rebutting these narratives.Agency rests with the collective assertion of Aboriginal survival rather than with an individual character.

Discussion
This discussion will consider the extent to which these texts depicted representation, identity, and power for particular groups, including those traditionally marginalized by selective literary traditions.It connects to the research literature and identifies the ways in which these texts did and did not conform to typical selective traditions in children's literature.

Author and illustrator identity
The texts explored in this study were from Australia (Day break, Yobbos do yoga), France (Press here) and New Zealand (Things in the sea are touching me), with the author and illustrator of Who sank the boat residing across both Australia and New Zealand.There were an equal number of male and female authors and illustrators.No information was found about Herve Tullet's, Pamela Allen's, or Andrew Joyner's ethnicity, while commentary on Phillip Gwynne's earlier books clarifies that he is not Aboriginal (Gray, 2004).In contrast, Amy McQuire belongs to the Dharumbul and South Sea Islander peoples, while Matt Chun (Chinese-Australian), Linda Keegan (Chinese-Pakeha), and Minky Stapleton (Afrikaans) define their respective ethnicities and national connections on their professional websites.Other aspects of cultural identity, such as class affiliation or sexuality, were not commented on in any of the author biographies.
It is unclear whether Pamela Allen is a farmer, Phillip Gwynne or Andrew Joyner are either yobbos or yoga enthusiasts, or Linda Keegan or Minky Stapleton parent in samesex, mixed-ethnicity families.It is also unclear whether these aspects of authorial identity impact on the representations in these lighthearted texts.In contrast, Day break, by Dharumbul and South Sea Islander Amy McQuire, specifically addresses the conflict between Aboriginal and settler-colonial perspectives about, and experiences of, the 26 th of January in Australia.Children's literature about minority groups frequently represents cultural or ethnic groups as singular and homogeneous (Galda et al., 2000) or seek to portray a happy multicultural vision of society (Gardner, 2020).Day break subverts this trend by portraying three Aboriginal family members who respond to dominant messages about the 26 th January in nuanced ways: the girl asks questions of all, engages in activities, and observes; Dad is more reactive, depicted in a warrior spear throwing pose when contesting the teacher's claim of white discovery of Australia; and Nan is gentler, more contemplative, but with clear messages about the importance of country and remembering.In contrast, the white teacher and friends are depicted as singular, monocultural stereotypes, standing in for a settler-colonial view of Australia and parodying homogenous depictions of minority ethnic groups.Through these rounded and varied Aboriginal characters, McQuire's Day break responds to Trousdale's (1990) and Gardner's (2020) appeal for diverse and authentic representations by authors belonging to the groups they write about and Madsen et al.'s (2021) centering of authorship in texts about First Nations peoples.

A different gender bias
Three of the five texts in this study had human characters; and each of these had a child protagonist within a family grouping.Interestingly, these children were all girls, a finding that differs from the findings of others that more protagonists in children's literature are boys (Adam & Harper, 2021;Ferguson, 2019;Luke et al., 1986).It was also unexpected as there was no conscious intention to prioritize girl characters.Each of these girls is presented in a non-traditional family unit-a girl with her Nan and Dad in Day break, a girl with her Mum and Ma in Things in the sea are touching me, and a girl with her Dad in Yobbos do yoga.While it was not a conscious choice to show a range of family groupings in these texts, children's literature representing one parent or separated parent families is common and represents a large section of society (Australian Bureau of Statistics, 2021).In contrast, children's books with same-sex parented families remain uncommon, and many of those that exist are educative texts about having same-sex parents rather than depicting diverse families involved in adventures and everyday events (Sunderland & McGlashan, 2012).Overall, the human character texts utilized in this unit offer gender and family representations that differ to the selective traditions identified and discussed in the literature review (Adam & Harper, 2021;Jipson & Paley, 1991;Luke et al., 1986;Smith et al., 1987).

Mirrors and windows to Australian lives
Two of the five books were by Australian authors and illustrators, with Australian content, themes, and language.Day break contains multiple linguistic and visual references to Australian cultural knowledge and historical understandings, including country as a place of belonging and connection, valorization of soldiers from particular wars, the belief in Australia as terra nullis pre-European settlement, and the inter-generational trauma of the Stolen Generations.Day break uses visual juxtaposition of symbols, color, and environments to further illuminate the different perspectives being depicted.Yobbos do yoga uses colloquial language to describe music loving car enthusiasts and depicts these yobbos as cheerful men with long hair, beards, a mullet haircut, band t shirts, and flannel shirts.Dad conforms to a certain middle-class stereotype with his pink outfit, ponytail, yoga, and neuroticism; though, interestingly, no colloquial words are used to describe him.These linguistic and visual symbols conjure up associations for many readers though Clarke (2021) suggests these are dated and no longer representative of Australian society.Broader understandings of culture as multifaceted and not limited to ethnicity are uncommon in children's literature (Jones, 2006) but are important as readers may conflate culture with "otherness" and struggle to discuss their own cultural identities (Christ & Sharma, 2018;Glazier & Seo, 2005).The central theme of "don't judge a book by its cover" is accessible in Yobbos do yoga, even if the reader is not familiar with the cultural stereotypes depicted.In contrast, Day break requires understanding of the contested nature of Australian history and contains multiple layers in the text and images allowing for deep connections.Luke et al.'s (1986) analysis of pre-service teachers' chosen texts examined characters' agency within the narrative to consider which types of character were afforded power.In my own selection of texts, agency and power were not always prominent.In Press here, the journey style of narration and lack of characters meant that there was no conflict to resolve.The animals in Who sank the boat are represented as equal and the "problem" that occurs at the end with the boat sinking is resolved by the reader's understanding rather than the characters' actions.Mum and Ma in Things in the sea are touching me are responsible for identifying the things in the sea, reassuring their daughter, and playing a gentle trick.In contrast, Day break and Yobbos do yoga have clearer points of conflict within their narratives.Day break shows how power is contested through examples of tensions between Aboriginal and settler-colonial perspectives about Australia's past in the first half of the book.The second half emphasizes Aboriginal survival through remembering and Country.Agency comes from the collective "we"-not just Nan, Dad, and the girl, but Aboriginal people and ancestors.Yobbos do yoga has the most obvious child agency as the girl is wiser, seeing the potential in her new neighbors before her prejudiced father and is proven to be right when they are receptive to turning their music down, helping fix Dad's car, and eventually becoming Dad's friends.These findings differed from others' research findings that female and minority-ethnic children's book characters were less likely to be agentive (Adam & Harper, 2021; Centre for Literacy in Primary Education, 2020; Luke et al., 1986).

Celebrated texts
Press here and Who sank the boat were winners of one or more significant awards for children's books.Interestingly, both were texts with non-human characters with either no clear setting (Press here) or a non-country-specific pastoral setting (Who sank the boat).Caple and Tian's (2021) analysis of 20 years of the CBCA shortlisted books for the early years award found that 47% of the books' characters were non-human, suggesting that such texts commonly receive accolades.Day break and Things in the sea are touching me were listed as notable or shortlisted for, but did not win, national children's book awards in Australia and New Zealand respectively.Each of these texts told specific human stories of: an extended Aboriginal family, and a mixed ethnicity, same-sex parented family.Each of these texts could be construed as contentious, for example, by those who believe families should have heterosexual parents, or by those who regard Australia's settler colonial past with pride and patriotism.The strong representation of non-human characters in award winning texts raises the question-might texts with non-human characters and non-controversial content be seen to transcend cultural and national boundaries and therefore garner more approval from award panels?This is an area for further exploration, especially given that award winning texts are often recommended for classroom use (Caple & Tian, 2021;MacKay et al., 2017).

Text diversity
This analysis showed that some of the texts explored in this research appear to have provided alternate representations to those often utilized in classrooms, including agentive girls, diverse family groups, and contemporary cultural understandings.It also showed that my text selection conformed to aspects others have critiqued like using no character or non-human character texts (Press here and Who sank the boat) and an older text (Who sank the boat).These five texts reflected certain realities and not others, providing support for Jipson and Paley's (1991) claim that we are all part of a selective tradition, making literary choices that reflect our own personal social and cultural values.These texts provide mirrors and windows to select worlds rather than showcasing a diverse representation of peoples and worlds.For example, no human boys were main characters, and there were no transgender or gender non-conforming characters, or disabled characters.A small subset of ethnicities and cultures was portrayed across the three texts with human characters, and all human characters wore western clothing with no religious symbols or attire.
While representation in these texts did not conform to the selective tradition identified by many (Adam, 2021;Adam & Harper, 2021;Ferguson, 2019;García, 2017;Gardner, 2020;Jipson & Paley, 1991;Jones, 2006;Luke et al., 1986), each text portrayed a particular way of being rather than standing in for "diversity."Each text is one representation, and, over time, many texts have the potential to show a rounded view of the world.A selective tradition of literature that excludes most groups or relegates them to minor roles is not acceptable.But I ask whether describing literary choices as "diverse" accurately captures educators' text selections.We create learning environments that reflect our values and I question whether these would typically represent all perspectives or groups.For example, I do not plan to purchase or teach with the unashamedly pro-gun picture book My parents open carry (Jeffs et al., 2010), nor the transgender-critical Johnny the Walrus (Walsh & Reece, 2022), as these contain values that are incompatible with my own and with the broad goals of education-"to become successful learners, confident and creative individuals, and active and informed citizens" (ACARA, n.d.).Loughran (2004) explains that "self-study may be an attempt to better understand how to manage the dilemma… rather than a search for the correct response to a specific question" (p.26).In this research context, engaging in self-study provided the opportunity to engage in a detailed analysis of the texts I work with, uncovering some surprising details and trends.It also helped me to identify types of stories that are not represented in this sample or in my wider teaching picture book collection.Most importantly, this self-study enabled me to conceptualize representation in children's literature as a dilemma to work more consciously with.My living contradiction-did I practice what I preached when selecting texts-had no simple response.Instead, I found ways in which I did and did not conform to selective literary traditions and realized the extent to which I privilege windows to some worlds and not others.Like Meacham and Meacham (2014) I aim to become more transparent about my rationale for selecting texts and more nuanced in the ways in which I discuss the importance of representation in children's literature.

Limitations
This small study explored five children's picture books used for modeling aspects of the English curriculum.The purpose was to interrogate my own selective literary tradition and consider the mirrors and windows provided to pre-service teachers via the texts used for modeling curriculum area content.Five texts is a small sample and this is clearly acknowledged.The narrow range of titles allowed an in-depth analysis of each text, which contrasts with the descriptive statistics reported by larger surveys of early years, school, and preservice teachers' text selections (Adam & Harper, 2021;Adam et al., 2021;Jipson & Paley, 1991;Luke et al., 1986;Smith et al., 1987).The educational purpose of these texts differed from that presented in some teacher education research in that they were not shared for aesthetic or experiential reasons (Simpson, 2016;Singer & Smith, 2003), or to explore culturally relevant pedagogy or critical thinking (Christ & Sharma, 2018;Jipson & Paley, 1991;Martin & Spencer, 2020).Rather, each text was linked to the modeling of teaching a specific aspect of the curriculum and selected to broaden pre-service teachers' minds about representation in the children's literature they might choose in their own settings.Pre-service teachers' engagement with and study of each text was limited to a single seminar rather than the extended time that a children's literature unit may offer.
Unlike many self-study projects (e.g., Abi-Hanna et al., 2014;Mahalingappa, 2018) this research did not utilize a critical friend (Pinnegar & Hamilton, 2009) or explore pre-service teachers' responses, interactions, or transactions with the texts (Martin & Russell, 2014;Pinnegar & Hamilton, 2009).This was a conscious choice.Firstly, the aim of the study was to interrogate my own selective tradition and the extent to which texts I select for learning in units of study align with my values about the role and purposes of literature in education.I believed that a self-study of my own selective tradition and personal motivations for text selection should come before research into others' responses to those texts.A secondary reason related to issues of actual and perceived power within my teaching and learning context.As is common in universities in Australia and internationally, my teaching colleagues were hired on semester contracts, reported directly to me, and were paid only for the specific teaching tasks in their contracts.It did not feel ethical to invite their participation in unpaid research activities, nor to ask them to provide feedback on my professional decisions given that I was responsible for employing them.The pre-service teachers studying this unit were required to do so as part of their degree programs.Given that I was responsible for the unit content, managing teaching staff, and overseeing marking across the unit; seeking student participation in a study evaluating my professional decisions troubled me in terms of possible perceptions of power, particularly in a first year, first semester unit with pre-service teachers who were often more familiar with the rules and power structures of compulsory education.

Recommendation
This paper concludes with an invitation to teachers and teacher educators to explore who and what is represented in your chosen teaching texts.Through interrogating our selective traditions, acceptable topics, and preferred themes, we may be better able to (a) identify gaps in representation and work on strategies for addressing these, and (b) articulate and share our rationales for excluding certain texts, authors, or topics-to ourselves, our colleagues, and our students.I welcome dialogue with others engaged in this work.

Conclusion
This paper has examined one teacher-educator's texts for teaching and has shown ways in which these conform to and diverge from common patterns of representation in teachers', book awards', and publishers' text selections.It provides evidence supporting the idea that we are all part of a selective tradition in the texts we choose for learning in educational settings.Rather than positioning oneself in opposition to this tradition, I suggest that acknowledging the specific biases and representations in the texts we choose to showcase and inviting dialogue about these may support more nuanced and fruitful discussion about the windows, mirrors, and sliding glass doors we offer to students through our uses of literature in educational settings.
(a) Audit children's literature used in unit.(b) Develop inclusion and exclusion criteria for texts for analysis.(c) Create journal entries explaining the rationale for choosing each text.(d) Engage in literature review investigating how others have analyzed teachers' selections of children's literature.(e) Develop evaluative criteria drawn from research literature.(f) Analyze each text in relation to the criteria.3. Write and disseminate research findings with an invitation to engage in dialog on our selection of children's literature for teaching.
Text copyright © Amy McQuire 2021.IllustraƟons copyright © MaƩ Chun 2021.Published in Australia by Bright Light, an imprint of Hardie Grant Children's Publishing.

Table 1
Text analysis summary table