A Game to Target the Spelling of German Children with Dyslexia
Maria Rauschenberger, Silke Fuechsel, Luz Rello, Clara Bayarri, Azuki Gòrriz · 2015 · ASSETS '15: Proceedings of the 17th International ACM SIGACCESS Conference on Computers & Accessibility · doi:10.1145/2700648.2811345
Summary
This demonstration paper presents the adaptation of Dyseggxia, a successful Spanish spelling game for children with dyslexia, to the German language. The original Spanish version was evaluated in an 8-week study with 48 children (ages 6-11) and showed that error-based game exercises significantly improved spelling compared to a Word Search control condition. The German adaptation required substantial linguistic work because dyslexia manifestations are language-dependent. The adaptation process involved three phases. First, researchers collected 47 texts (homework, dictations, essays) written by students ages 8-17, including 32 from children diagnosed with dyslexia and 15 from students with high spelling error rates. Second, errors were manually extracted and annotated with phonetic and visual features specific to German orthography—including characters absent from Spanish (ß, ä, ö, ü) and the distinct letter forms of the German handwriting system (Lateinische Ausgangsschrift). Third, linguistic analysis addressed German's semi-opaque orthography (where letter-sound relationships are less regular than Spanish) and complex syllabic structure. The resulting German version includes 2,500 manually created word exercises across eight exercise types and five difficulty levels. Two exercise types are German-specific: capital letter and wrong capital letter, addressing the frequent misspelling pattern of missing capitalization on German nouns. The "cut into words" exercises include German-specific morphemes like trennbare Verben (separable prefix verbs).
Key findings
The paper demonstrates that adapting dyslexia interventions across languages requires more than translation—it demands comprehensive linguistic reanalysis. German's characteristics created specific challenges: semi-opaque orthography means letter-sound mappings are less predictable than Spanish; complex morphology and longer words required adjusting difficulty level criteria; and the unique capitalization rules for nouns created an entirely new error category requiring dedicated exercises. The researchers identified existing German tools (Lernserver, Klex 11, CESAR schreiben 2.0) but found none based on error analysis of actual dyslexic writing. Lernserver showed promising results in a 2008 evaluation where 78.2% of 3,798 students improved their writing. However, the error-based approach—using the specific mistakes children with dyslexia make as the foundation for targeted exercises—represents a distinct methodology grounded in research showing that (1) dyslexic writing errors reflect specific underlying difficulties and (2) readers with dyslexia cannot consciously detect errors in words. The study cites longitudinal research showing that children diagnosed with dyslexia at age 8 achieved lower school performance and higher unemployment rates at age 25, even with effective treatment and above-average socioeconomic status. In Germany, only 25% of poor spellers achieve average spelling performance during primary school, underscoring the need for effective interventions.
Relevance
This work illustrates an important principle for accessibility technology: interventions validated in one language cannot simply be translated to another. Dyslexia manifests differently depending on orthographic transparency—languages with consistent letter-sound relationships (like Spanish or Italian) present different challenges than semi-opaque (German) or opaque (English) orthographies. Practitioners developing or deploying dyslexia tools should verify that the underlying linguistic analysis matches the target language. The error-based methodology offers a replicable framework: collect authentic writing samples from the target population, systematically annotate errors using language-specific phonetic and visual features, analyze patterns linguistically, then design exercises targeting identified error types. This approach could extend to other languages or even other writing difficulties. The integration into a game format addresses motivation challenges inherent in remedial practice. The planned longitudinal school evaluation will provide valuable data on whether error-based games are differentially effective across language types—a question with implications for global deployment of dyslexia interventions.
Tags: dyslexia · serious games · spelling · German · children · learning disabilities · educational technology · error-based learning · gamification