Measurement errors in voice-key naming latency for Hiragana

Area of Science:

• Linguistics
• Psycholinguistics
• Speech Science

Background:

• Voice-key technology is frequently used in speech production research, including naming latency studies.
• Previous research has not fully addressed the potential for measurement errors in voice-key latency data, particularly for simple linguistic units like single hiragana symbols.
• Understanding these limitations is crucial for accurate interpretation of naming behavior.

Purpose of the Study:

• To investigate the limitations and possibilities of voice-key naming latency research specifically for single hiragana symbols.
• To evaluate the accuracy of standard voice-key measurement procedures by comparing data with established speech-analyzer waveform data.
• To identify sources of measurement error and propose methods for their mitigation.

Main Methods:

• Analysis of three datasets of voice-key naming latency for single hiragana symbols.
• Comparison of voice-key data with reference voice-waveform data from Sakuma, Fushimi, and Tatsumi (1997).
• Examination of potential measurement errors introduced by phoneme characteristics and procedural aspects.

Main Results:

• Standard voice-key measurement procedures can introduce substantial errors in hiragana naming latency data.
• These errors can obscure the actual effects of variables influencing hiragana naming behavior.
• Specific analytical adjustments, such as applying Sakuma et al.'s deltas and excluding error-prone initial phonemes, can reduce measurement errors.

Conclusions:

• Voice-key latency research on single hiragana symbols requires careful consideration of measurement error.
• Standard procedures may not be sufficiently accurate for precise analysis of naming behavior at this level.
• While errors can be mitigated for simple symbols, they may be less significant when analyzing complex linguistic materials like words.