Learning and performance of able-bodied individuals using scanning systems with and without word prediction

Area of Science:

• Human-Computer Interaction
• Assistive Technology
• Cognitive Psychology

Background:

• Text entry methods are crucial for computer usability, especially for individuals with motor impairments.
• Word prediction features aim to reduce keystrokes and improve typing speed.
• Understanding the cognitive and perceptual load of such features is essential for effective design.

Purpose of the Study:

• To evaluate the impact of a word prediction feature on text transcription learning and performance.
• To analyze the trade-offs between keystroke savings and the cognitive costs of using word prediction.
• To investigate the relationship between feature use, selection rate, and text generation speed.

Main Methods:

• Two groups of able-bodied subjects transcribed text using row-column scanning systems.
• One system included a word prediction feature; the other did not.
• Subjects completed 10 trials, with group order of system use varied.

Main Results:

• The word prediction system showed no substantial increase in learning difficulty.
• No statistically significant improvement in text generation rate was observed with the word prediction system.
• The cost associated with using the word prediction feature offset the benefits of keystroke savings.

Conclusions:

• Under the study's conditions, the word prediction system's benefits were negated by its usage costs.
• Further research is needed to optimize word prediction for enhanced text generation performance.
• The balance between keystroke savings and selection cost is critical for effective assistive technology design.