The use of interactional metadiscourse markers by Saudi EFL male and female college students: the case of a gender-sensitive topic

The study investigates how Saudi EFL undergraduate males and females deploy interactional metadiscourse markers in argumentative essays on a gender-sensitive topic: comparative driving ability of men versus women. Writing, especially argumentative writing, requires rhetorical resources to manage stance and reader engagement. Metadiscourse (Hyland, 2005) enables writers to convey evaluation, attitude, and epistemic positioning. Prior work indicates EFL/ESL learners struggle with argumentative writing and use of metadiscourse can vary by gender, discipline, topic, proficiency, and L1. In Saudi Arabia, women’s right to drive (ban lifted in 2018) and evolving gender roles render “who drives better” sensitive. The study addresses: (1) How do Saudi EFL male and female writers use interactional metadiscourse in argumentative writing about driving? (2) How do metadiscourse patterns vary by gender and stance on this gender-sensitive topic (men for men vs. men for women; women for men vs. women for women)?

Research on interactional metadiscourse has focused on academic genres (articles, abstracts, theses) with fewer studies on student argumentative writing. Topic choice and L1 background can shape marker use (Yoon, 2021). Gender-based findings are mixed: some report males use more boosters, others that females hedge or boost more; engagement and self-mention patterns also vary by discipline and genre (Hyland & Tse, 2008; Pasaribu, 2017; Zadeh et al., 2015; Azlia, 2022; Farahanynia & Nourzadeh, 2023). Within Arab contexts, studies span columns, acknowledgements, consultations, theses, with differing gender effects (Alsubhi, 2016; Alotaibi, 2018; Ahmed & Maros, 2017; Merghmi & Hoadjli, 2024). There is a gap regarding Arab EFL undergraduates’ argumentative writing on gender-sensitive topics, motivating the present study.

Design: Explanatory sequential mixed-methods. Quantitative phase counted interactional metadiscourse markers; qualitative phase examined contexts to code and interpret usage per Hyland’s (2005) model. Participants: 144 Saudi undergraduates majoring in English translation (59 males, 85 females) with ~6 years of EFL instruction. Task and data collection: An in-exam argumentative essay (≥400 words; introduction, two body paragraphs, conclusion) on “Whether Saudi men drive better than Saudi women or vice versa.” Students chose a stance. Data collected in Autumn 2021 in Academic Writing (TRAJ 221). Prior to collection, students were taught argumentative structure but not explicitly metadiscourse. Pretest and equivalence: Before the main task, students wrote on a different argumentative topic. Scores graded by experienced instructors; Independent Samples t-test and one-way ANOVA confirmed no significant differences in writing ability between genders/groups. Corpus: 144 essays; total 46,453 tokens. Four subcorpora: females for women (66 essays; 24,104 tokens; 2,023 types), females for men (19; 7,461; 1,122), males for men (39; 8,576; 1,106), males for women (20; 6,312; 1,428). Frequencies were normalized per 1,000 words to account for size differences. Analytical framework: Hyland’s (2005) interactional metadiscourse categories: hedges, boosters, attitude markers, self-mentions, engagement markers. Only interactional resources were analyzed. Tools: AntConc (v4.0.5) for frequency lists and KWIC to validate metadiscourse function in context (e.g., disambiguating items like “pretty,” “way”). Reliability: Two coders jointly coded 25% of male subcorpus; 20% of female subcorpus for inter-rater reliability. Cohen’s kappa = 0.647 (substantial agreement). Disagreements on borderline items resolved by following Hyland (2005) and contextual checks. Statistics: Descriptive statistics (min, max, mean, SD, skewness, kurtosis). Normality assessed (skewness <2; kurtosis <7). Inferential tests: Independent Samples t-test (gender comparisons) and one-way ANOVA with LSD post hoc (four stance-by-gender groups). Alpha p<0.05. Effect sizes via eta squared (η²: small 0.01, medium 0.06, large 0.14).

Equivalence: No significant pretask difference in writing ability between males and females (Males 16.85±6.232; Females 18.29±3.386; t=-1.613, p=0.111). Across four groups, ANOVA indicated a small difference in writing scores (F=1.685, p=0.047), with males-for-females scoring lower than females-for-females, not expected to affect metadiscourse use. Overall by gender (per-essay means; Table 6): - Boosters: Females 22.34±7.893 vs Males 12.59±6.349; p<0.001; η²≈0.304 (large). - Attitude markers: Females 24.01±8.430 vs Males 15.24±5.937; p<0.001; η²≈0.251 (large). - Hedges: Females 10.33±4.917 vs Males 8.51±4.337; p=0.023; η²≈0.036 (medium). - Engagement markers: no significant difference (Females 3.00±2.948; Males 3.20±2.981; p=0.686). - Self-mentions: Females 5.39±3.704 vs Males 3.59±3.696; p=0.005; η²≈0.055 (small). Normalized frequencies per 1,000 words indicated higher female rates for attitude (64.66 vs 54.33), boosters (60.16 vs 44.26), and self-mentions (14.50 vs 11.95), with males slightly higher in hedges by normalization (29.95 vs 27.81) and engagement (11.08 vs 8.07). By gender × stance (ANOVA; Table 7): - Boosters: Significant differences (p<0.001). Means: Males-for-men 11.26; Males-for-women 15.20; Females-for-men 20.68; Females-for-women 22.82. Female groups > male groups; males-for-men lowest. - Attitude markers: Significant (p<0.001). Means: 13.77; 18.10; 29.79 (highest, females-for-men); 22.35. All pairwise differences significant. - Hedges: Significant (p<0.001). Means: 7.56; 10.35; 13.47 (highest, females-for-men); 9.42. Females-for-men significantly > others; males-for-men lowest. - Engagement markers: Not significant (p=0.682). - Self-mentions: Significant (p=0.024). Females-for-women 5.65 > males-for-men 3.51 (significant); other pairwise not reported significant. Lexical tendencies: Frequent attitude words included good, better, even; males used easily, important, long, hard; females used bad, important, careful. Boosters commonly think, believe, many, a lot, all; ‘all’ used for generalizations by both genders. Common hedges: can, some, most. Self-mentions: I and my most frequent; females also used we to project solidarity. Engagement: you most common, with females also using should; males used rhetorical questions more. Headline results: Females used attitudinal lexis, hedges, self-mentions, and boosters more than males. Females arguing for men’s driving used hedges significantly more than other groups. Females supporting female drivers used self-mentions more than males arguing for men’s driving. Engagement markers showed no significant gender or stance effects.

Findings address the research questions by showing that in argumentative essays on a gender-sensitive topic, interactional metadiscourse use varies by gender and stance. Attitude markers dominated, followed by boosters and hedges, consistent with argumentative genre demands; engagement resources were least used. Females’ higher use of self-mentions, hedges, boosters, and attitude markers suggests a more personalized, assertive, yet cautious style, potentially reflecting recent sociocultural shifts (e.g., post-2018 empowerment) and a desire to assert capability and solidarity (use of we). Boosting often accompanied positive evaluation (e.g., careful, important) and self-mentions, signaling ownership and confidence. Stance modulated patterns: females arguing for men hedged the most, aligning with culturally cautious positioning when supporting the opposite gender; females supporting women self-foregrounded more than males supporting men. Males arguing for men used the fewest boosters/attitudes, possibly relying on presumed communal consensus in a traditionally male-associated domain. Overall, results reinforce that topic sensitivity and stance, not gender alone, condition metadiscourse choices and identity projection in EFL argumentative writing.

The study contributes evidence that on a gender-sensitive topic, Saudi EFL female undergraduates employ more attitude markers, boosters, hedges, and self-mentions than males, while engagement markers show no gender difference. Stance matters: females advocating men’s driving hedge most; females supporting women use more self-mentions than males supporting men. These patterns underscore that metadiscourse use is shaped by topic and stance, not deterministically by gender. Pedagogically, explicit instruction on the pragmatic functions of metadiscourse (especially boosters and attitude markers) can enhance argumentative writing quality and help students calibrate stance and engagement effectively.

- Uneven and relatively small subcorpora, especially males supporting females; fewer male enrollees overall; some essays under 400 words. - Single institutional context (King Saud University). - Single genre (argumentative essay), single gender-sensitive topic (driving), single age group; exam conditions may constrain performance. Future work should use larger, multi-institutional corpora; compare genres (including spoken debates); examine additional gender-sensitive vs neutral topics; consider pre-/post-policy change comparisons; and include cross-linguistic analyses to probe L1 influence and cultural norms.