A Participatory Exploration of Writing Skills and Predictors of Writing Quality Among Autistic and Non-Autistic College Students

Background: Although essential for success in college, the writing skills of autistic college students have received limited empirical attention (Gerstle & Walsh, 2011; Jackson et al., 2018). Autistic people often exhibit writing difficulties (Brown et al., 2014; Zajic et al., 2018). Difficulties with Theory of Mind (ToM) and/or global coherence may contribute to narrative challenges in autism (Brown & Klein, 2011; Joliffe & Baron-Cohen, 2001). However, some autistic adults report communicating more effectively through writing than speaking (Davidson & Smith, 2009) and autistic people with developed language skills may not struggle with ToM (Happé, 1995). Therefore, autistic college students may not experience specific challenges writing. As writing demands change throughout the lifespan (Bazerman et al., 2018), more information about the writing abilities of autistic college students is critical.

Objectives: To compare the writing skills of autistic and non-autistic college students and to examine predictors of writing quality.

Methods: Autistic (n = 26; 92.6% male; 64% White; Mean age = 21.3) and non-autistic college students (n = 24; 43.5% male; 45% White; Mean age = 20.5) completed a measure of ToM (RMIE; Baron-Cohen, 2001), a global integration task (Joliffe & Baron-Cohen, 2011), a personal essay (wherein they were asked to share something they found meaningful by highlighting what it meant to them and whether or not it would mean the same thing to others), and a short story. After realizing that our initial hypotheses/coding lacked autistic perspectives, we asked two autistic college students to develop their own hypotheses and coding schemes. One neurotypical and one autistic coding pair developed coding schemes and obtained reliability. The neurotypical coding scheme included global coherence, ToM, and non-literal language. Autistic coders coded for creativity, readability, specificity, repetitiveness, emotional dullness, and point of view. We used paper.rater.com to assess length, errors, and quality (AutoGrader).

Results: Autistic students (20%) less frequently included multiple perspectives in their essays than non-autistic students (67%; p = .001). Their essays trended toward lacking global coherence (autistic 20%; non-autistic 0%, p = .051) and being more creative (15% autistic; 0% non-autistic; insufficient instances for chi-square analysis). Autistic students’ essays exhibited fewer simple sentences and heightened writing quality. Autistic students used fewer simple sentences and more complex vocabulary in their stories (ps < .02). No qualitative group differences in fiction were observed. A linear regression revealed that autism classification (r = .55) and RMIE (r = .53, ps < .004), but not age or gender, were positively associated with writing quality. Global integration was not included in the regression due to collinearity with RMIE, r(38) = .04, p = .01 but was not correlated with writing quality.

Conclusions: Autistic students exhibited subtle difficulties using ToM in their personal essays (wherein attention to multiple perspectives was prompted) but not in open-ended fictional writing. Though demonstrating variability, findings suggest that autistic college students leverage narrative skills to help them succeed in college and understand others. Future participatory explorations of the writing of autistic college students should examine institutional factors that contribute to writing development longitudinally.