Date of Graduation


Degree Type



College of Education and Human Services


Communication Sciences and Disorders

Committee Chair

Kimberly Meigh, Ph.D.

Committee Member

Kimberly Meigh, Ph.D.


Purpose: This study used a two-stage model of motor learning (Latash, 2012) to evaluate bad and good variance during a speech learning task. Stage 1 learning involves minimizing bad variance, or overall error, whereas Stage 2 learning involves changes in good variance while ceiling accuracy is maintained. Using two different variables to evaluate speech complexity (manner of articulation and syllable length), two hypotheses were evaluated: 1) A two-stage model exists during speech motor learning and 2) complexity of stimuli will alter the achievement and maintenance of each stage during learning.

Methods: Twelve native English speakers with normal speech and hearing skills repeated stimuli that varied by manner of articulation (stops versus fricatives) and syllable length (4 versus 6-syllables) while having their tongue movements tracked using the WAVE Speech Research system. Complexity of the stimuli were ranked: simple = stops, 4-syllable, complex = fricative, 6-syllable.

Results: Visual analysis was conducted by evaluating each participant’s productions for accuracy and duration per block. Results suggest there are two-stages of learning during a speech motor learning task. Three patterns of learning were observed in achievement of Stage 1 learning, and Stage 1 was achieved more quickly for simple stimuli. Additionally, differences in duration based on complexity of the stimuli were also observed during Stage 2 learning.

Conclusion: A two-stage model was observed contrary to traditional speech motor learning theories. This study suggests variables other than accuracy may be related to learning and generalization. The complexity of stimuli also influences learning outcomes.