Milodie Butsch is a graduate student in the Department of Speech & Hearing Sciences. At this year's first annual campus-wide student research symposium, Milodie, along with over a hundred others, met in the ballroom of the Smith Memorial Student Union to present a project, exchange ideas and celebrate student research.
Butsch presented the results of her research project: "Using Lexical Diversity to Quantify Dynamics of Neurotypical Discourse."
Language impairment is a broad term that refers to a condition in which a person is unable to express and/or understand meaning through language. Both children and adults can suffer the effects of language impairment. Symptoms can include having a hard time understanding language, difficulty following directions or organizing thoughts. Aphasia, dysphasia, specific language impairment and auditory processing disorder are common forms of language impairment.
According to Butsch, diagnosis of language impairment often begins with a language sample and analysis that provides information on lexical diversity (LD), one metric used in the determination of diagnosis. Researchers, speech therapists, and pathologists, however, have no conclusive data pointing to how long language samples should be to make a proper assessment of an individual's LD measuring language ability. Is a short sample better than a long, or vice versa? Is there an ideal sample length?
the conclusions drawn from Butsch's research shed some light on questions like these. In order to reach those conclusions, Butsch analyzed 120 language samples from adults with typical language abilities between the ages of 20 and 80. The analysis provided a score pertaining to the LD of each of the language samples. Butsch then took those same 120 language samples, shortened them to 75 words and preformed the same analysis on the truncated samples.
"The point of the exercise," Butsch said, "was to see if a shortened language sample would provide the same results in terms of LD as the longer sample. We found that the shorter language samples didn't give us the same validityof conclusions. They were weaker and less reliable. We couldn't draw strong conclusions regarding an individual's LD from them."
There may not yet be a definitive length at which a language sample can provide concrete evidence of the presence or absence of language impairment and there may never be. Butsch's research, however, provides evidence that shorter language samples are a less effective tool than longer samples when determining an individual's LD, suggesting that shorter samples are also less effective tools for diagnosing language impairments.
"This kind of research contributes to making other research more psychometrically sound," Butsch said. "It could be used to help develop psychometric tests that really get at figuring out how language impaired an individual may be."
With the first annual campus-wide student research symposium behind her, Butsch plans to focus on completing her graduate degree. She is interested in working with individuals with language impairment resulting from brain injuries and other conditions such as stroke. Beginning this fall Milodie will take on an internship at the Legacy Good Samaritan Rehabilitation Institute of Oregon where she will work with patients suffering from stroke, brain injury and other conditions that cause language impairment.
Authored by Shaun McGillis
Posted June 19, 2013