The relation between reading skills and linguistic prediction - An investigation of Dyslexia

My master's dissertation at the University of East Anglia delved into the cognitive mechanisms underlying linguistic prediction in individuals with dyslexia compared to control groups. This pioneering research was recognized for its contribution to the field and was selected for presentation at the Architectures and Mechanisms for Language Processing (AMLaP) conference in 2018, Berlin.

Research Objective
The study aimed to dissect the relationship between skilled reading and linguistic prediction capabilities, particularly focusing on the performance disparities between dyslexic and non-dyslexic individuals across varying linguistic constraints.

Methodology
The project was structured around two experimental designs:

Experiment 1: Participants completed cloze sentences, allowing us to measure natural language processing through reaction times and cloze probabilities.
Experiment 2: Participants engaged in speeded semantic-plausibility judgments while listening to sentences, with reaction times recorded from the final word's onset to response.
A 2 x 2 x 3 factorial design (Group: dyslexic vs. control x Constraint: low vs. high x Completion: high, low, and anomalous) underpinned the experiments, yielding rich datasets for analysis.

Data Analysis and Visualization
All data were meticulously compiled and analyzed using Microsoft Excel. This versatile platform enabled the efficient manipulation and examination of large datasets, facilitating the calculation of reaction times and probability measures. The findings were visually represented through graphs, also generated in Excel, which effectively conveyed the complex data in an accessible and interpretable format.

Key Findings
The analysis revealed that individuals with dyslexia exhibit challenges in linguistic prediction. Specifically:

In Experiment 1, dyslexic participants demonstrated notably weaker performance with high-constraint sentence items.
In Experiment 2, they showed a marked inability to predict low-constraint items, alongside significant baseline reaction time discrepancies.
Implications
The results of this research underscore two critical insights:

Proficient reading is intricately linked to the adeptness of linguistic prediction.
Linguistic prediction presents a substantial challenge for individuals with dyslexia, affecting their reading proficiency.

Conclusion
This dissertation contributes valuable evidence to the psycholinguistic understanding of dyslexia. It offers a nuanced view of the predictive deficits that characterize dyslexic reading, setting the stage for future interventions aimed at ameliorating these specific challenges and enhancing overall reading skills.