| تعداد نشریات | 44 |
| تعداد شمارهها | 1,922 |
| تعداد مقالات | 15,635 |
| تعداد مشاهده مقاله | 47,683,153 |
| تعداد دریافت فایل اصل مقاله | 19,193,425 |
Gendered Agency in the Global Success Primary English Textbooks: A Source-Verified Behavioural Analysis | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Applied Research on English Language | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| مقالات آماده انتشار، اصلاح شده برای چاپ، انتشار آنلاین از تاریخ 20 آبان 1405 اصل مقاله (539.89 K) | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| نوع مقاله: Research Article | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| شناسه دیجیتال (DOI): 10.22108/are.2026.149908.2811 | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| نویسنده | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Thuan Duc Pham* | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Faculty of Foreign Languages and Information Technology, Hoa Lu University, Ninh Binh, Vietnam | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| چکیده | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Primary English textbooks do not merely distribute visibility; they also model who may initiate, question, direct, coordinate, decide, and solve problems. This study examined gendered first-order agency in all eight Global Success primary English textbooks used in Vietnam. A source-verified content analysis coded seven non-mutually exclusive behavioural categories in 1,013 eligible character occurrences across Grades 1–5, supplemented by qualitative contextual interpretation of selected source cases. Frequencies and gender contrasts were examined using descriptive statistics, Fisher’s exact tests with Benjamini–Hochberg adjustment, chi-square analysis, grade-adjusted logistic regression, and page- and section-clustered generalized estimating equations. Of these, 108 (10.7%) met the agency threshold. Female and male occurrences had similar overall rates (11.3% and 12.3%), and gender remained non-significant after adjustment for grade. Nevertheless, male characters more often initiated interactions and asked advancing questions, whereas female-skewed leadership was concentrated almost entirely in teacher roles. No source-confirmed occurrence met the thresholds for consequential decision-making or problem-solving. Agency varied significantly but weakly across grades: Grade 1 was markedly lower than Grades 2–5, but the later sequence showed no linear progression. The findings demonstrate that aggregate parity can coexist with behavioural, role-bound, and grade-specific asymmetries, supporting source-gated, fine-grained analysis in textbook evaluation and materials development. | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| کلیدواژهها | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Gendered Agency؛ EFL Textbooks؛ Textbook Discourse؛ Primary Education؛ Vietnam | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| اصل مقاله | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Introduction Textbooks organize linguistic input while also normalizing ways of acting and relating. Through recurring classroom, family, and occupational scenes, they model who may speak first, question, direct, coordinate, decide, or solve problems (Apple, 2018; Brugeilles & Cromer, 2009). Repetition can make these allocations appear routine rather than ideologically consequential. This function is particularly important in primary education. In Vietnam, the 2018 General Education English Curriculum embeds communicative competence within a sequenced national programme (Hoang, 2022), making the Global Success series a valuable corpus for examining represented influence across Grades 1–5. Textbook-gender research has documented visibility, firstness, occupations, family roles, activities, traits, and verbal space, showing that improved numerical representation may coexist with stereotyped positioning (Islam & Asadullah, 2018; Lee, 2014; Mills & Mustapha, 2015; Vu & Pham, 2023). Yet presence does not establish equivalent influence: a frequent speaker may remain responsive to actions initiated or organized by others. Dialogue research makes this distinction visible. Comparable speech can conceal differences in who initiates, seeks or provides information, and controls interactional development (Jones et al., 1997; Poulou, 1997; Yang, 2011). Huang (2025) likewise found that broad female-male balance could coexist with differences in particular discourse features. An opening question, a reactive answer, and a directive may occupy similar textual space while conferring different interactional possibilities. Gender equality must therefore be examined through the functions that move interaction forward, not inferred from presence or talk alone. Agency offers an appropriate lens. It is a socioculturally mediated capacity to act, enabled and constrained by social relations and roles (Ahearn, 2001; McNay, 2000, 2004). Operationalizing this construct requires caution. Broad labels such as active or empowered may exceed the page evidence, while incomplete dialogues, unassigned voices, and learner prompts can obscure actor attribution. The present study therefore uses a source gate: actor, gender, event, and enactment must be recoverable before an action can be coded as initiating, advancing, directing, coordinating, deciding, solving, or otherwise influencing a represented situation. Participant role and grade may structure these opportunities. Teacher authority, for example, should not be generalized as female or male empowerment, and adult authority must be distinguished from children’s agency. Early formulaic tasks and later extended dialogues also provide unequal textual space for influence. Gender patterns must therefore be interpreted alongside participant role and grade-specific task architecture. Three gaps follow. Few studies distinguish multiple observable forms of gendered agency across a complete primary English sequence; participant role is rarely tested as an explanation for apparent gender differences; and grade-related opportunities remain underexamined. Vietnamese studies have addressed broad representation and ideology, but not source-verified behavioural agency across the complete Global Success Grades 1–5 series. This study addresses these gaps through a corpus-assisted analysis of first-order agency, defined as a represented character’s observable influence on an interaction, task, person, or event. It separates fine-grained behaviours, excludes attribution-ambiguous and learner-elicited material, and examines gender through participant role and grade. The study seeks to answer the following three research questions: To what extent do source-verified female and male occurrences meet the first-order agency threshold, and which behavioural forms do they enact? How are first-order agency forms distributed across participant roles, particularly teacher and child/student roles? How does overall agency vary across Grades 1–5, and is gender associated with agency after grade is taken into account?
Literature Review Gender Representation in EFL Textbooks Gender analyses of language-learning materials commonly examine frequency, firstness, naming, speech, occupations, domestic roles, activities, traits, and visual prominence (Brugeilles & Cromer, 2009; Mills & Mustapha, 2015; Sunderland, 2000). These indicators reveal both who is included and the social worlds made available to each gender. Such work shows that numerical inclusion and representational equality may diverge, and contemporary materials can balance female and male presence while differentiating their social or linguistic positioning (Islam & Asadullah, 2018; Lee, 2014; Vu & Pham, 2023). These measures remain essential but do not establish comparable influence. A frequent character may mainly respond, whereas a less visible character may initiate or coordinate. Agency analysis therefore complements, rather than replaces, established measures by shifting attention from presence to enacted effect. Textbook meanings must also be distinguished from classroom uptake. Koster and Litosseliti (2021) found that teachers who supported equality often left biased textbook representations unexamined. Textbook content and teacher mediation are therefore related but analytically distinct layers of gender construction.
Agency as Observable Influence Agency is relational and situated rather than a context-free possession. Ahearn (2001) conceptualizes agency in terms of socially and culturally mediated capacities for action, which are enabled and constrained by social relations. Emirbayer and Mische (1998) emphasize its temporal and relational character, while feminist accounts locate action within embodied experience and institutional positions (McNay, 2000, 2004). Childhood research similarly cautions against equating agency with unrestricted individual autonomy (Christensen, 2025). The present study does not claim to operationalize this broader construct in full. Instead, first-order agency is treated as a deliberately bounded and empirically recoverable subset: an enactment through which a depicted character observably affects an interaction, task, person, or event. It captures represented influence but not latent capacity, unexpressed intention, resistance, unrealized possibilities for action, or constraints that do not become visible in the represented event. First-order agency may also be enabling when a character explicitly creates conditions for another to act through facilitation, organization, coordination, or provisioning. Such action qualifies only when the actor, target, and facilitating effect are recoverable from the page; inferred helpfulness or presence alone is insufficient. This behavioural construct differs from trait-based agency, such as assertiveness or dominance (Hsu et al., 2021), and from mere speech, movement, preference, participation, or compliance. Represented agency must also be separated from the situated agency of textbook users. Kostas (2023) conceptualizes gender meanings as emerging through relations among texts, children, teachers, and classroom interaction. This study analyses textbook characters’ depicted actions, not how teachers or learners may reproduce, negotiate, or resist them.
Gendered Forms of Interactional Agency Dialogue studies provide the clearest behavioural precedent. Jones et al. (1997) examined who participates and initiates; Poulou (1997) related turn position and speech to language functions; and later work considered voice, questions, requests, and information exchange (Yang, 2011). Huang (2025) showed that overall interactional balance can coexist with feature-specific differences. This scholarship moves analysis from counting turns to asking what each turn accomplishes and which participant it positions to respond. Verbal quantity therefore cannot substitute for discourse function. Results across contexts vary. Yang (2011) found more female utterances but broadly similar mixed-gender initiation and information seeking. Benlaghrissi (2022) found more female initiators, yet some subsequently received information from knowledgeable males. Initiation creates an opportunity to move interaction but does not guarantee continuing informational control. Questions likewise differ functionally: they may open or redirect an exchange, request action, or merely reproduce a drill. The analysis therefore distinguishes interactional initiation from questions that substantively advance a represented exchange and excludes decontextualized learner prompts.
Directives, Leadership, and Institutional Authority Directives and leadership make influence more explicit by orienting others toward action or organizing shared activity. A single request can shape an addressee’s next action, while leadership may sequence a lesson, game, itinerary, or shared task. Yet both depend on social position. Teachers, pupils, parents, peers, and workers have unequal rights to instruct, question, and manage others (Huang & Liu, 2024). Teacher representation is especially consequential. Female teachers may model legitimate institutional authority, yet their leadership cannot automatically be generalized to girls or women outside school
Decision-Making, Problem-Solving, and Consequential Agency Children’s literature research often treats overcoming constraints, making consequential choices, solving problems, and shaping outcomes as strong forms of agency. Vaughn et al. (2022) examined who receives such opportunities and for what purposes, while Elorza (2023) showed that agentive identities arise from configurations of processes and participant roles. These stronger forms matter because they represent characters not merely participating but altering what subsequently becomes possible. Both approaches move beyond presence to effects within represented events. A stringent threshold is necessary. Preferences qualify as decisions only when alternatives and consequences are represented; task completion qualifies as problem-solving only when a character recognizes and causally addresses a challenge. Even structural zeros remain informative because they expose forms of agency absent from the curriculum.
Participant Roles and Curricular Opportunity Structures Textbook roles and genres create unequal opportunities for agency: teachers direct lessons, doctors instruct patients, and pupils in drills often respond. A high agency rate may therefore reflect access to an authoritative role rather than a generalized gender advantage. Ignoring these structures risks attributing role composition to gender. Analyses of textbook interaction and represented action consequently support attention to adult–child, professional, and peer relations (Elorza, 2023; Huang & Liu, 2024). Opportunity may also vary across grades. Cross-grade studies show the value of analysing complete primary sequences (Agha et al., 2018; Lee, 2014; Li et al., 2017), but behavioural agency remains underexamined. Grade is treated here as a curricular opportunity structure, not evidence of developmental empowerment.
Analytical Framework and Research Gap The literature supports a behavioural, relational, and source-bound framework. Seven overlapping categories capture interactional initiation, advancing questions, directives, leadership and coordination, consequential decisions, problem identification and resolution, and other goal-directed influence. The framework distinguishes interactional movement from institutional organization and from actions that visibly change outcomes. Overlap is retained when one action performs several functions; unsupported higher-order dimensions are avoided. Few studies combine an observable agency threshold, participant-role analysis, grade comparison, and occurrence-level source verification. Existing work often illuminates one dimension—visibility, dialogue structure, social role, or represented action—without connecting these dimensions to a reproducible behavioural denominator. Applying the four elements to the complete Vietnamese Grades 1–5 series makes it possible to identify which behaviours female and male characters enact, from which roles, and under which curricular conditions.
Methodology Research Design and Theoretical Orientation The study used a source-verified, quantitatively led content analysis supplemented by qualitative contextual interpretation (Creswell & Plano Clark, 2018; Krippendorff, 2019). The audited corpus was compared by gender, role, and grade, while selected cases were re-examined to explain how textual context produced the observed patterns. Guided by sociocultural and relational accounts of agency (Ahearn, 2001; Emirbayer & Mische, 1998; McNay, 2000, 2004), the study operationalized first-order agency as a source-recoverable action that observably influenced an interaction, task, person, or event. This bounded construct captures represented enactment, not latent capacity, intention, resistance, or learners’ classroom agency; textbook representations provide curricular affordances whose uptake remains an empirical question.
Corpus and Scope The corpus comprised all eight Global Success primary textbooks: one volume for Grades 1 and 2 and two for Grades 3–5. Their 92 units and 20 reviews produced 112 coded sections. The complete series was retained to preserve grade, book, section, and page provenance. The starting master corpus contained 1,228 potentially gendered character occurrences.
Unit of Analysis and Eligibility An occurrence was one recoverable gendered character in an immediate source-grounded event. Verbal and visual evidence established the actor, action, and observable effect; gaze, framing, angle, salience, composition, and image–text relations were not systematically analysed. Repeated characters were coded separately when they appeared in different events, so occurrences were not unique persons or illustrations. This event-bound unit prevented a recurrent character from receiving a fixed agency trait across contexts. Gender was assigned from explicit names, pronouns, labels, or sufficiently unambiguous visual-contextual evidence. Female-male analyses used 914 eligible occurrences (451 female; 463 male);
Source-Gate Audit and Analytical-Corpus Construction Each record passed a source-binding gate. The book and page were located, the recorded event and actor were checked against verbal and visual evidence, and enactment was classified as represented, modelled, learner-elicited, or ambiguous. Extracted text assisted location but never replaced inspection of the published page. Five volume-specific page offsets were corrected, with all changes retained in a revision trail. Records were excluded for mismatched sources, unreproducible speaker or gender attribution, learner-completion prompts, or unassigned chants, songs, joint utterances, and generic turns. These rules prevented pedagogical invitations to learners from being treated as actions already enacted by textbook characters. Of 1,228 potential occurrences, 178 were excluded during initial source binding because of source mismatches, unrecoverable actor or gender attribution, learner-completion prompts, or unassigned turns, leaving 1,050. A secondary page-level check of all 153 provisionally agentive records confirmed 69 unchanged, corrected and retained 39, recoded 8 as non-agentive, and excluded 37. Thus, 215 records were excluded overall, producing 1,013 source-valid occurrences: 451 female, 463 male, and 99 mixed, unspecified, or other occurrences; 108 met at least one agency criterion.
Coding Framework and Decision Rules Agency was coded through seven binary, non-mutually-exclusive categories: interactional initiation (INT_INIT), advancing questions (QUEST_ADV), directives (DIRECT), leadership and coordination (LEAD_COORD), consequential decision-making (DECIDE), problem identification and resolution (PROBLEM), and other goal-directed agency (OTHER_AG). Interactional initiation began or materially advanced an interaction, topic, or activity. Advancing questions opened, redirected, investigated, clarified, or developed an exchange; routine repetition and display prompts were excluded. Directives sought to shape another identifiable character’s action. Leadership and coordination required sustained organization of a group or shared activity rather than a single instruction. Consequential decisions required an explicit choice among alternatives with a discernible outcome; preferences alone were insufficient. Problem identification and resolution required recognition of a challenge and a causal contribution to its solution; routine exercise completion did not qualify. Enabling actions were included when the source represented an identifiable facilitative effect. Sustained organization or coordination was coded as leadership and coordination, whereas discrete acts of service, creation, provisioning, or relational support were coded as other goal-directed agency. Background assistance, inferred helpfulness, or presence without a recoverable effect did not qualify. Each category was coded 1 when present and 0 when absent; unresolved source cases were excluded. Categories could overlap, as when an opening question met both initiation and advancing-question criteria. Overall agency equalled 1 when any first-order criterion was met and 0 otherwise; it was a threshold indicator, not a cumulative score. Participant roles were grouped as teacher, child/student, or other adult/worker. Supporting fields recorded context, mode, configuration, target, consequence, evidence, and rationale, but participant role provided the formal comparison for RQ2. For descriptive analysis, other adult/worker roles were subdivided into family/caregiver, occupational/professional, and other or insufficiently specified adult roles; small gender-specific cells precluded inferential testing.
Calibration, Repeat-Pass Consistency, and Source Verification The codebook was calibrated on 50 occurrences, comprising ten cases from each grade and including relevant boundary cases. A separately seeded set of 100 records, with 20 from each grade, was reserved for repeat-pass consistency assessment. Source auditing removed 39 records, and one unresolved attribution case was excluded, leaving 60 paired cases. The same coding procedure was applied in two separately ordered passes, with first-pass decisions unavailable during the second. Records were matched only after both passes were completed. Agreement between passes was summarized using percentage agreement and Cohen’s kappa, interpreted cautiously for sparse categories (Cohen, 1960; McHugh, 2012; O’Connor & Joffe, 2020). Overall-agency agreement was 95.0% (κ = .886); agreement was 98.3% for interactional initiation (κ = .955) and 96.7% for other goal-directed agency (κ = .783). Advancing questions, directives, and leadership and coordination showed complete agreement, although the latter two contained only one positive case each. Consequential decision-making and problem-solving contained no positive cases, so kappa was not estimable. These statistics indicate repeat-pass consistency rather than independent human intercoder reliability. The principal researcher subsequently checked all boundary cases and all 153 provisionally agentive occurrences against the original textbook pages before the analytical corpus was finalized. Because a suitably trained independent human coder was not available, the study is explicitly presented as a single-researcher, source-verified analysis. Source verification and repeat coding strengthen consistency and auditability but do not substitute for independent intercoder agreement.
Quantitative Analysis Frequencies and rates were calculated for overall agency and each behavioural category. Sparse female–male contrasts were examined using two-sided Fisher’s exact tests, with odds ratios and 95% confidence intervals reported only when estimation was meaningful. No odds ratio or inferential comparison was calculated for consequential decision-making or problem identification and resolution because both categories contained no positive occurrences. For the five category-level contrasts containing positive cases, raw p-values were adjusted using the Benjamini–Hochberg false discovery rate procedure (Benjamini & Hochberg, 1995). This procedure was selected because the category-level analyses were exploratory and aimed to limit the expected proportion of false discoveries while retaining sensitivity across a small set of related and sparse behavioural outcomes; familywise-error procedures such as Bonferroni correction would address a stricter confirmatory objective. Raw and adjusted p-values were considered alongside effect sizes, confidence intervals, and cell frequencies. Grade variation was examined using chi-square and Cramér’s V. The grade-adjusted logistic model included gender and grade as main effects, with grade treated categorically. Sensitivity to within-text dependence was assessed with marginal logistic GEE models using an exchangeable working correlation and robust sandwich standard errors (Liang & Zeger, 1986). Both included gender and categorical grade; observations were clustered by page in the primary model and by section in a broader check. Page- and section-clustered GEE sensitivity models were also fitted for interactional initiation and advancing questions, the two category outcomes with sufficient positive occurrences. Cluster-aware models were not used for directives, leadership and coordination, or other goal-directed agency because their positive counts were too small for stable category-level estimation. Consequential decision-making and problem-solving contained no positive cases and therefore could not be modelled. Gender-by-grade interaction terms were not fitted because only 108 agency-positive occurrences were available and several within-grade cells were sparse. Teacher-excluded and child-only sensitivity analyses assessed whether institutional authority explained apparent gender patterns. Because occurrences were clustered within pages and recurring characters, all inferential results were interpreted as exploratory descriptions of this finite textbook corpus rather than population-level estimates.
Qualitative Contextual Analysis The qualitative component was deductive and explanatory, using the seven predefined behavioural categories rather than emergent themes. Cases were selected after quantitative coding to represent recurrent patterns, gender contrasts, rare forms, and boundary decisions; selection supported interpretation rather than prevalence estimation. Selected pages were re-examined for actor, gender, enactment, action, and effect, considering wording, turn position, addressee, sequence, role, target, and consequence. Excel supported case retrieval and retained source locators, excerpts, and rationales; interpretation relied on direct page inspection, without automated theme generation. Figure 1 summarizes the procedure.
Figure 1. Decision Sequence for Source Verification and Behavioural Agency Coding
Results Among the 1,013 source-valid occurrences, 108 (10.7%) met at least one first-order agency criterion. The gender analysis included 914 female or male occurrences; none of the remaining 99 occurrences coded as mixed, unspecified, or other met the agency threshold. The results are organized around gender and behavioural forms (RQ1), participant roles (RQ2), and grade variation and grade-adjusted, cluster-aware gender comparisons (RQ3). Source-verified examples are used to interpret key contrasts and clarify coding boundaries.
Gender and First-Order Behavioural Forms Female and male occurrences had similar overall agency rates: 51/451 (11.3%) and 57/463 (12.3%), respectively. The absolute difference was one percentage point. Fisher’s exact test detected no association, p = .682, OR = 0.91, 95% CI [0.61, 1.36], φ = −.016 (Table 1). The interval included modest effects in either direction. Overall agency was a threshold indicator: an occurrence with several concurrent behaviours counted once in Table 1. Behaviour-specific assignments are reported separately below and should not be added to infer additional agentive occurrences.
Table 1. Source-Valid Corpus and Overall Agency by Gender
Note. Formal female-male comparisons exclude mixed, unspecified, and other occurrences. The unit of analysis is the character occurrence.
Behavioural distributions differed despite overall parity (Table 2). Interactional initiation occurred in 23 female (5.1%) and 48 male occurrences (10.4%), OR = 0.46, exploratory adjusted . Advancing questions showed a similar contrast: 19 female (4.2%) and 42 male occurrences (9.1%), OR = 0.44, adjusted . Fifty-eight records received both codes because an opening or redirecting question simultaneously moved the exchange and performed the grammatical form under analysis. Leadership and coordination showed the reverse pattern: 13 female (2.9%) and one male occurrence (0.2%), OR = 13.71, exploratory adjusted .
Table 2. First-Order Agency Forms by Gender
Note. Codes are non-mutually exclusive; 58 occurrences received both INT_INIT and QUEST_ADV. Reported -values reflect two-sided Fisher’s exact tests. After applying Benjamini–Hochberg false discovery rate (FDR) correction for multiple testing across categories, differences for INT_INIT (adjusted ), QUEST_ADV (adjusted ), and LEAD_COORD (adjusted ) remain statistically significant. Dashes indicate non-estimable structural-zero contrasts.
As the role analysis shows, this contrast was not broadly distributed across female characters. OTHER_AG was descriptively female-skewed (14 vs. 6), p = .072, while directives were rare (three female; one male). No occurrence met the consequential decision-making or problem-solving thresholds; consequently, odds ratios and gender comparisons for these structural-zero categories were not estimable. The 108 agentive occurrences generated 170 assignments because codes could overlap. Most duplication arose from initiation combined with advancing questions; therefore, category counts describe behavioural repertoires and are not additive. Cluster-aware analyses retained both principal contrasts. For interactional initiation, adjusted female-to-male ORs were 0.47, 95% CI [0.27, 0.81], p = .007 (page-clustered), and 0.45, 95% CI [0.23, 0.85], p = .014 (section-clustered). For advancing questions, corresponding ORs were 0.46, 95% CI [0.23, 0.90], p = .024, and 0.42, 95% CI [0.19, 0.94], p = .034. Both male-skewed patterns persisted after adjustment for grade and clustering. Agency Across Participant Roles Agency varied by role. Teachers had the highest rate (17/35, 48.6%), followed by other adults/workers (19/134, 14.2%) and children/students (72/745, 9.7%). Teachers contributed fewer agentive records in absolute terms than children but had a rate approximately five times higher. The male-teacher rate in Table 3 is based on only four records and is descriptive.
Table 3. Overall Agency by Participant Role and Gender
Note. The male-teacher denominator is four. Role groups are analytical aggregations in Agency Corpus v1.1. Role breakdowns exclude 99 mixed, unspecified, or group occurrences where gender was non-binary or unassigned (valid n = 914 across assigned role groups).
Role stratification explained the female-skewed leadership result: 12 of 13 female leadership occurrences involved teachers, as did the single male case. Mary’s sequenced Sydney itinerary was the only non-teacher female instance. Leadership was therefore concentrated in institutional teacher activity rather than broadly distributed among female characters. Although the leadership and coordination contrast was substantial, it was based on only 14 positive occurrences and should therefore be interpreted as a large but imprecisely estimated, role-bound pattern. Among children/students, agency occurred in 26/347 female (7.5%) and 46/398 male records (11.6%), exploratory p = .063. Excluding teachers also produced no gender association: female 37/420 (8.8%), male 54/459 (11.8%), OR = 0.72, p = .183. Other adult/worker rates were similar (15.1% female; 13.1% male). Among 134 female or male other adult/worker occurrences, family/caregiver roles contained 5 agentive cases (4/49 female; 1/33 male), occupational/professional roles 8 Thus, female-skewed leadership did not create a general female advantage outside teacher roles. Children supplied 72 of the 108 agentive occurrences, other adults 19, and teachers 17. These shares reflect unequal denominators: children formed most of the corpus, whereas only 35 teacher records were eligible. Both counts and within-role rates are therefore needed to interpret role effects.
Grade Progression and Grade-Adjusted Gender Agency varied significantly but weakly by grade, χ²(4) = 13.32, p = .010, Cramér’s V = .115. Grade 1 was markedly lower (2/113, 1.8%) than Grades 2–5 (8.8%–13.2%), but the latter rates fluctuated rather than rising linearly: Grade 3 fell below Grade 2, while Grades 4 and 5 were similar (Table 4; Figure 2).
Table 4. Agency Rates across Grades 1–5
Note. Female and male denominators do not sum to grade totals because other gender categories remain in overall denominators. Tests are exploratory.
No within-grade female-male contrast was significant. Grade 2 showed the largest numerical female advantage, whereas Grade 4 showed the largest male advantage; rates were closely matched in Grades 3 and 5. A main-effects logistic regression model including gender and grade, with grade entered as a categorical covariate, retained the overall null gender association, adjusted OR = 0.86, 95% CI [0.57, 1.30], p = .479. Gender-by-grade interaction terms were not fitted because only 108 agency-positive occurrences were available and several within-grade cells were sparse, making stable interaction estimates inappropriate. Overall parity was therefore not an artefact of different grade distributions. The page-clustered GEE (360 clusters) yielded an adjusted female-to-male OR of 0.88, 95% CI [0.58, 1.33], p = .535; the section-clustered model (108 clusters) yielded OR = 0.86, 95% CI [0.54, 1.37], p = .524. Both retained the null gender association after grade adjustment.
Figure 2. Distribution of Overall Agency Rates across Grades 1–5 Note. Bars represent discrete grade-level agency rates and should not be interpreted as evidence of a continuous developmental trajectory.
Source-Confirmed Behavioural Contrasts Source-confirmed cases clarify the threshold. In Grade 2, a girl’s “Is there a fox?” opened and advanced an exchange, whereas the boy’s “Yes, there is” was reactive (Book 1, p. 28). Speaking alone was therefore insufficient; the turn had to create interactional movement. Directive agency appeared when language targeted another character’s action, as in a male doctor’s “Open your mouth!” (Grade 3, Book 1, p. 30). Female and male teachers met the leadership criterion when they organized lessons or games, not merely when they stood before a class (Grade 3, Book 1, p. 50; Grade 4, Book 2, p. 43). OTHER_AG captured purposeful effects outside the specified forms. Mai helped her brother with mathematics, and Minh and Mai planted a tree (Grade 5, Book 1, pp. 50, 74). The former was not coded PROBLEM because the page did not depict recognition, solution development, or causal resolution of a challenge. Mary’s email sequencing a Sydney itinerary was the only non-teacher female leadership case (Grade 5, Book 2, p. 57). It confirmed that student coordination was possible but rare. Across the cases, no character made a consequential choice among alternatives or completed an explicit problem-solving process. Taken together, the findings answer the three research questions directly. RQ1 showed similar overall female and male agency rates but differentiated behavioural forms: interactional initiation and advancing questions were male-skewed, whereas leadership and coordination were female-skewed. RQ2 showed that the leadership contrast was concentrated in teacher roles rather than broadly distributed among female characters. RQ3 identified significant but weak and non-linear grade variation, while grade adjustment and page- and section-clustered models consistently retained the null overall gender association.
Discussion Overall Agency Parity and Differential Positioning Female and male occurrences crossed the overall agency threshold at similar rates, and grade adjustment did not change this result. The series therefore does not construct a simple hierarchy in which one gender is generally more agentive. Neither gender was systematically confined to visibility, routine participation, or response. This broad balance qualifies findings from materials where numerical inclusion coexists with male advantage or stereotyped positioning (Islam & Asadullah, 2018; Lee, 2014; Vu & Pham, 2023). Aggregate parity nevertheless masks how influence is distributed. Dialogue research shows that similar speech or participation may conceal who initiates, seeks information, directs, or controls development (Huang, 2025; Jones et al., 1997; Poulou, 1997). Here, near-equal totals were assembled through different behavioural repertoires, which a single active/passive or empowered/disempowered measure would obscure.
Gender Differentiation in Behavioural Form Male occurrences were about twice as likely to initiate and to ask questions that advanced an exchange. This is scripted positioning, not an intrinsic male disposition: textbook dialogues allocate opening turns, target questions, and routine answers through recurring character assignments (Ahearn, 2001; McNay, 2000). The pattern concerns access to interactional movement, not greater knowledge or leadership in every exchange. Such ordinary choices can cumulatively position boys and men as interactional movers. The result is more specific than greater male talk. Earlier studies found either female initiation followed by male informational control (Benlaghrissi, 2022) or broadly similar initiation despite unequal utterance totals (Yang, 2011). In the present corpus, reactive answers were non-agentive and questions qualified only when they created observable movement. Because initiation and advancing questions overlapped strongly, the exploratory contrasts represent one recurring question-led pattern, not two independent male advantages. Leadership moved in the opposite gender direction but should not be combined with initiation into an overall empowerment score. Opening interaction and organizing collective activity enact different relations. Keeping them separate shows that female influence was not absent; its most distinctive form was institutionally located.
Female Leadership as Institutional Rather Than Generalized Empowerment Participant role explains the female-skewed leadership pattern. Twelve of 13 female cases belonged to teachers, whose lessons and games modelled women as legitimate institutional authorities, consistent with Yang (2011). The pattern is substantively positive because women are shown directing collective activity rather than merely assisting it. Nevertheless, it reflects the authority attached to a participant role rather than a general female advantage. Once teachers were removed, females showed no overall advantage; among pupils, the descriptive male difference remained statistically inconclusive. Women could lead where the curriculum assigned them teacher authority, while girls rarely coordinated peers. Mary’s exceptional itinerary demonstrates that female student leadership is feasible and could be distributed more routinely across peer, family, and project contexts.
Restricted Consequential and Problem-Oriented Agency No occurrence depicted consequential choice or causal problem resolution across the five grades, although characters expressed preferences, answered questions, assisted others, and completed activities. Short, structured activities may restrict such opportunities, but task type and linguistic complexity were not analysed; the cause of these structural zeros therefore remains undetermined. These zeros do not imply that primary learners cannot act complexly; they describe the represented textbook world. Reclassifying preferences, correct answers, or a reference to a mathematics problem would weaken the observable threshold and romanticize routine participation (Christensen, 2025; Schwebel, 2016). Retaining the categories makes their absence visible and identifies a space for curriculum development.
Grade Progression as a Changing Opportunity Structure Grade 1’s very low rate plausibly reflects formulaic naming, recognition, repetition, and brief responses, which offer few opportunities to redirect or coordinate events. Later books include longer dialogues, games, emails, projects, visits, and social scenarios in which actor, target, and effect are more readily recoverable. Grade therefore operates as a curricular opportunity structure rather than merely a chronological variable. The later sequence did not rise steadily: Grade 3 fell below Grade 2, and Grades 4 and 5 were similar. As cross-grade research cautions (Agha et al., 2018; Lee, 2014; Li et al., 2017), such fluctuations may be associated with differences in genre, topic, dialogue design, or participant role rather than progressive empowerment. Because task type and genre were not prespecified coding variables, their contribution could not be tested in the present study. Grade adjustment and the page- and section-clustered models nevertheless retained the overall null gender association.
Theoretical and Methodological Contribution Theoretically, the study treats represented agency as a repertoire of observable, relational behaviours rather than a personality trait inferred from prominence. It distinguishes behavioural influence from trait-based agency-communion (Hsu et al., 2021) while operationalizing socioculturally mediated agency at the level of recoverable textbook evidence (Ahearn, 2001). This approach avoids labelling whole characters as active or passive across contexts. Gender, role, and curricular opportunity jointly organize who can influence whom and how. Methodologically, the source-gated procedure demonstrates why actor, gender, event, and enactment must be treated as construct-validity conditions rather than clerical details. Source verification prevents learner prompts, unassigned voices, reactive participation, and unsupported actor attributions from entering behavioural agency counts. Overlapping codes remain useful for identifying actions that perform more than one function, although sparse categories and occurrence clustering require cautious interpretation. The contribution therefore lies in a transparent and reproducible procedure for identifying represented behavioural patterns rather than in population-level claims from scripted textbook records.
Pedagogical and Materials-Development Implications Materials writers should supplement character counts with a behavioural-role audit recording who initiates exchanges, advances questions, directs others, coordinates activities, makes consequential choices, and resolves problems across gender, role, and grade. This would expose distributions concealed by aggregate visibility. Task designers can rotate initiator, questioner, planner, and coordinator roles across girls and boys; extend female leadership beyond teachers; represent boys in supportive roles; and embed age-appropriate choices and problem scenarios. These are design audits, not quotas, and should be adapted to curricular aims. Series-level review should track whether agency repertoires expand across grades through a compact gender-by-role-by-behaviour matrix. Teachers can also rotate dialogue roles and extend routine responses into questioning, planning, choice, and problem-resolution activities. Although the findings concern one Vietnamese series, the source gate and behavioural-role audit are transferable to other graded EFL programmes. Cross-context research can test whether numerical balance similarly conceals differences in interactional and consequential agency without generalizing the present distributions.
Conclusion This study found no overall female-male hierarchy in source-verified first-order agency: rates were similar, and gender remained unrelated to agency after grade adjustment. Aggregate parity nevertheless concealed differentiated enactment. Male characters more often initiated and asked advancing questions, whereas female-skewed leadership was located almost entirely in teacher roles. The latter represents women as legitimate authorities but does not establish a general advantage for girls. Grade 1 offered markedly fewer opportunities than Grades 2–5, but the later sequence showed no steady progression. No verified case represented consequential decision-making or problem-solving. The study’s contribution is a source-gated behavioural account that reveals patterns hidden by visibility, speech, or a single agency label. It shows that gendered influence must be interpreted through behavioural form, participant role, and curricular opportunity. Source verification also demonstrates that speaker attribution and represented enactment materially affect the substantive conclusion. Materials can build on the series’ broad balance by distributing initiation, questioning, peer leadership, consequential choice, and problem resolution more widely across genders and grades. Several limitations qualify the findings. First-order agency excludes latent capacity, intention, resistance, and unenacted constraints. Although page- and section-clustered GEE models retained the overall null association and two principal behavioural contrasts, recurring-character dependence could not be modelled because stable identifiers were unavailable. Sparse categories limited inference, and neither task/genre variation nor multimodal positioning was systematically analysed. The corpus represents one Vietnamese series and does not demonstrate classroom uptake. Repeat-pass and source checks strengthened consistency but did not constitute independent human intercoder reliability; independent replication is therefore needed. Future studies should use trained coders, stable character identifiers, predefined task/genre categories, multimodal analysis, classroom reception, and cross-series comparison.
Declaration of Conflicting Interests The author declared no potential conflicts of interest.
Funding Details This study received no specific funding. | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| مراجع | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Agha, N., Syed, G. K., & Mirani, D. A. (2018). Exploring the representation of gender and identity: Patriarchal and citizenship perspectives from the primary level Sindhi textbooks in Pakistan. Women’s Studies International Forum, 66, 17–24. https://doi.org/10.1016/j.wsif.2017.11.009 Ahearn, L. M. (2001). Language and agency. Annual Review of Anthropology, 30(1), 109–137. https://doi.org/10.1146/annurev.anthro.30.1.109 Apple, M. W. (2018). Ideology and curriculum (4th ed.). Routledge. Benjamini, Y., & Hochberg, Y. (1995). Controlling the false discovery rate: A practical and powerful approach to multiple testing. Journal of the Royal Statistical Society: Series B (Methodological), 57(1), 289–300. https://doi.org/10.1111/j.2517-6161.1995.tb02031.x Benlaghrissi, H. (2022). Gender representation in Moroccan EFL textbooks’ conversations: A quantitative and qualitative content analysis. International Journal of Language and Literary Studies, 4(3), 43–53. https://doi.org/10.36892/ijlls.v4i3.942 Brugeilles, C., & Cromer, S. (2009). Analysing gender representations in school textbooks. Umr Ceped. Christensen, N. (2025). Relational agency, children’s literature, and childhood. In G. Fenech (Ed.), Child as citizen: Agency, activism, and citizenship in children’s and young adult literature and culture (pp. 69–89). University Press of Mississippi. https://doi.org/10.14325/mississippi/9781496858405.003.0005 Cohen, J. (1960). A coefficient of agreement for nominal scales. Educational and Psychological Measurement, 20(1), 37–46. https://doi.org/10.1177/ 001316446002000104 Creswell, J. W., & Plano Clark, V. L. (2018). Designing and conducting mixed methods research (3rd ed.). Sage. Elorza, I. (2023). Gender-inclusive picture books in the classroom: A multimodal analysis of male subjective agencies. Linguistics and Education, 78, 101242. https://doi.org/ 10.1016/j.linged.2023.101242 Emirbayer, M., & Mische, A. (1998). What is agency? American Journal of Sociology, 103(4), 962–1023. https://doi.org/10.1086/231294 Hoang, V. V. (2022). Interpreting MOET’s 2018 general education English curriculum. VNU Journal of Foreign Studies, 38(5), 1–22. https://doi.org/10.25073/2525-2445/ vnufs.4866 Hsu, N., Badura, K. L., Newman, D. A., & Speach, M. E. P. (2021). Gender, “masculinity,” and “femininity”: A meta-analytic review of gender differences in agency and communion. Psychological Bulletin, 147(10), 987–1011. https://doi.org/ 10.1037/bul0000343 Huang, P., & Liu, X. (2024). Challenging gender stereotypes: Representations of gender through social interactions in English learning textbooks. Humanities and Social Sciences Communications, 11, 819. https://doi.org/10.1057/s41599-024-03293-x Huang, S. Y. (2025). Gender and diversity in EFL textbook dialogues: Interactional structure and pedagogical implications. TESOL Quarterly, 59(3), 1176–1198. https://doi.org/ 10.1002/tesq.3310 Islam, K. M. M., & Asadullah, M. N. (2018). Gender stereotypes and education: Jones, M. A., Kitetu, C., & Sunderland, J. (1997). Discourse roles, gender and language textbook dialogues: Who learns what from John and Sally? Gender and Education, 9(4), 469–490. https://doi.org/10.1080/09540259721204 Kostas, M. (2023). Textbooks, students and teachers talk around gender: A new materialist approach to children’s agency. Teaching and Teacher Education, 125, 104052. https://doi.org/10.1016/j.tate.2023.104052 Koster, D., & Litosseliti, L. (2021). Multidimensional perspectives on gender in Dutch language education: Textbooks and teacher talk. Linguistics and Education, 64, 100953. https://doi.org/10.1016/j.linged.2021.100953 Krippendorff, K. (2019). Content analysis: An introduction to its methodology (4th ed.). Sage Publications. Lee, J. F. K. (2014). Gender representation in Hong Kong primary school ELT textbooks – a comparative study. Gender and Education, 26(4), 356–376. https://doi.org/10.1080/ 09540253.2014.916400 Li, L., Xiong, T., & He, J. (2017). The representation of gender in a popular primary school EFL textbook series in China. International Journal of Education and Practice, 5(5), 79–87. https://doi.org/10.18488/journal.61.2017.55.79.87 Liang, K.-Y., & Zeger, S. L. (1986). Longitudinal data analysis using generalized linear models. Biometrika, 73(1), 13–22. https://doi.org/10.1093/biomet/73.1.13 McHugh, M. L. (2012). Interrater reliability: The kappa statistic. Biochemia Medica, 22(3), 276–282. https://doi.org/10.11613/bm.2012.031 McNay, L. (2000). Gender and agency: Reconfiguring the subject in feminist and social theory. Polity Press. McNay, L. (2004). Agency and experience: Gender as a lived relation. The Sociological Review, 52(S2), 173–190. https://doi.org/10.1111/j.1467-954X.2005.00530.x Mills, S., & Mustapha, A. S. (2015). Gender representation in learning materials: International perspectives. Routledge. https://doi.org/10.4324/9781315764092 O’Connor, C., & Joffe, H. (2020). Intercoder reliability in qualitative research: Debates and practical guidelines. International Journal of Qualitative Methods, 19, 1–13. https://doi.org/10.1177/1609406919899220 Poulou, S. (1997). Sexism in the discourse roles of textbook dialogues. Language Learning Journal, 15(1), 68–73. https://doi.org/10.1080/09571739785200141 Schwebel, S. L. (2016). The limits of agency for children’s literature scholars. Jeunesse: Young People, Texts, Cultures, 8(1), 278–290. https://doi.org/10.1353/jeu.2016.0014 Sunderland, J. (2000). Issues of language and gender in second and foreign language education. Language Teaching, 33(4), 203–223. https://doi.org/10.1017/ s0261444800015688 Vaughn, M., Sotirovska, V., Darragh, J. J., & Elhess, M. (2022). Examining agency in children’s nonfiction picture books. Children’s Literature in Education, 53(1), 33–51. https://doi.org/10.1007/s10583-021-09435-y Vu, M. T., & Pham, T. T. T. (2023). Still in the shadow of Confucianism? Gender bias in contemporary English textbooks in Vietnam. Pedagogy, Culture & Society, 31(3), Yang, C. C. R. (2011). Gender representation in a Hong Kong primary English textbook series: The relationship between language planning and social policy. Current Issues in Language Planning, 12(1), 77–88. https://doi.org/10.1080/14664208.2011.541390 | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
آمار تعداد مشاهده مقاله: 159 تعداد دریافت فایل اصل مقاله: 28 |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||