Comparing single case design overlap-based effect size metrics from studies examining speech generating device interventions

Mo Chen, Jolene K. Hyppa-Martin, Joe E. Reichle, Frank J. Symons

Research output: Contribution to journalArticle

15 Scopus citations

Abstract

Meaningfully synthesizing single case experimental data from intervention studies comprised of individuals with low incidence conditions and generating effect size estimates remains challenging. Seven effect size metrics were compared for single case design (SCD) data focused on teaching speech generating device use to individuals with intellectual and developmental disabilities (IDD) with moderate to profound levels of impairment. The effect size metrics included percent of data points exceeding the median (PEM), percent of nonoverlapping data (PND), improvement rate difference (IRD), percent of all nonoverlapping data (PAND), Phi, nonoverlap of all pairs (NAP), and Taunovlap. Results showed that among the seven effect size metrics, PAND, Phi, IRD, and PND were more effective in quantifying intervention effects for the data sample (N = 285 phase or condition contrasts). Results are discussed with respect to issues concerning extracting and calculating effect sizes, visual analysis, and SCD intervention research in IDD.

Original languageEnglish (US)
Pages (from-to)169-193
Number of pages25
JournalAmerican journal on intellectual and developmental disabilities
Volume121
Issue number3
DOIs
StatePublished - May 2016

Keywords

  • Effect size metric
  • Evidence-based practices
  • Intellectual and developmental disabilities
  • Single case design
  • Speech generating devices

Fingerprint Dive into the research topics of 'Comparing single case design overlap-based effect size metrics from studies examining speech generating device interventions'. Together they form a unique fingerprint.

  • Cite this