SRCLD Presentation Details
  Title  
       
    Applying Item Response Theory to the Development of a Screening Adaptation of the Goldman-Fristoe Test of Articulation  
Author(s)
Tim Brackenbury - Bowling Green State University
Michael Zickar - Bowling Green State University
Benjamin Munson - University of Minnesota
Holly Storkel - University of Kansas

SRCLD Info
SRCLD Year: 2016
Presentation Type: Poster Presentation
Poster Number:
Presentation Time: (na)
Categories
Abstract
Item Response Theory (IRT) is a form of statistical modeling that has been successfully used to develop effective and efficient protocols in multiple areas within communication sciences and disorders (e.g., Baylor et al., 2013). With the goal of developing a data-based screening measure, the present study applied an IRT analysis to 155 children’s productions of the sounds-in-words subtest of the Goldman-Fristoe Test of Articulation-2 (GFTA-2; Goldman & Fristoe, 2000). A total of 108 individual phonemes from the GFTA-2’s 53 target words fit the two-parameter logistic model. Most of the model’s information was in the negative range of the trait continuum, meaning that these items provided the most precise measurement at the low ability, suggesting that they are a good fit for a screening tool. A regression analysis on data from 90 of the participants revealed that their performances on the 3 phonemes with the highest discrimination scores in the model were significantly correlated with their standard scores from the GFTA-2. Sensitivity and specificity calculations for these three phonemes were commensurate with those of other child speech and language screening measures.