Assessment of reliable change using 95% credible intervals for the differences in proportions

a statistical analysis for case-study methodology

Rachael Unicomb*, Kim Colyvas, Elisabeth Harrison, Sally Hewat

*Corresponding author for this work

Research output: Contribution to journalArticle

10 Citations (Scopus)

Abstract

Purpose: Case-study methodology studying change is often used in the field of speech-language pathology, but it can be criticized for not being statistically robust. Yet with the heterogeneous nature of many communication disorders, case studies allow clinicians and researchers to closely observe and report on change. Such information is valuable and can further inform large-scale experimental designs. In this research note, a statistical analysis for case-study data is outlined that employs a modification to the Reliable Change Index (Jacobson & Truax, 1991). The relationship between reliable change and clinical significance is discussed. Example data are used to guide the reader through the use and application of this analysis. Method: A method of analysis is detailed that is suitable for assessing change in measures with binary categorical outcomes. The analysis is illustrated using data from one individual, measured before and after treatment for stuttering. Conclusions: The application of this approach to assess change in categorical, binary data has potential application in speech-language pathology. It enables clinicians and researchers to analyze results from case studies for their statistical and clinical significance. This new method addresses a gap in the research design literature, that is, the lack of analysis methods for noncontinuous data (such as counts, rates, proportions of events) that may be used in case-study designs.

Original languageEnglish
Pages (from-to)728-739
Number of pages12
JournalJournal of Speech, Language, and Hearing Research
Volume58
Issue number3
DOIs
Publication statusPublished - 1 Jun 2015

Fingerprint Dive into the research topics of 'Assessment of reliable change using 95% credible intervals for the differences in proportions: a statistical analysis for case-study methodology'. Together they form a unique fingerprint.

  • Cite this