Reliability or Difference? A Comment on the Reliability of the S/Z Ratio A comment on the article “Reliability of the S/Z Ratio in Normal Children’s Voices” (Fendler & Shearer, LSHSS,1988) is warranted. By title, purpose, and method the study was designed to examine the test-retest reliability of the S/Z ratio. By analysis, however, this was not done. Seventy-eight, first- and second-grade ... Letter to the Editor
Letter to the Editor  |   January 01, 1990
Reliability or Difference? A Comment on the Reliability of the S/Z Ratio
 
Author Affiliations & Notes
  • Charles L. Madison
    Washington State University Pullman, WA 99164-2420
  • J. Anthony Seikel
    Washington State University Pullman, WA 99164-2420
Article Information
Research Issues, Methods & Evidence-Based Practice / Letters to the Editor
Letter to the Editor   |   January 01, 1990
Reliability or Difference? A Comment on the Reliability of the S/Z Ratio
Language, Speech, and Hearing Services in Schools, January 1990, Vol. 21, 60. doi:10.1044/0161-1461.2101.60
History: Received March 6, 1989 , Accepted July 3, 1989
 
Language, Speech, and Hearing Services in Schools, January 1990, Vol. 21, 60. doi:10.1044/0161-1461.2101.60
History: Received March 6, 1989; Accepted July 3, 1989
A comment on the article “Reliability of the S/Z Ratio in Normal Children’s Voices” (Fendler & Shearer, LSHSS,1988) is warranted. By title, purpose, and method the study was designed to examine the test-retest reliability of the S/Z ratio. By analysis, however, this was not done. Seventy-eight, first- and second-grade children were tested on 2 occasions, by different clinicians, to establish the temporal-reliability of the S/Z ratio. Test-retest reliability, or stability over time, is typically analyzed using a simple correlation statistic, assuming data are numerical, related, and normally distributed. Fendler and Shearer appeared to base their conclusion that “the S/Z ratio was found to be an acceptably reliable measure” (p. 4) on the evidence that there was no significant difference between test and retest scores. The conclusion of acceptable reliability is not justified when the data are treated in this way, and neither is the conclusion of inter-judge reliability as implied by Fendler and Shearer. The example presented in Figure 1 should help to illustrate the point. These test and retest distributions will have equivalent means and standard deviations, will not be significantly different, and will be significantly negatively correlated.
First Page Preview
First page PDF preview
First page PDF preview ×
View Large
Order a Subscription
Pay Per View
Entire Language, Speech, and Hearing Services in Schools content & archive
24-hour access
This Article
24-hour access