Is it REAL? Research evaluation for audiology literature.
Palmer-CV; Mormer-E; Ortmann-A; Byrne-D; Ye-Y; Keogh-L
Hear J 2008 Oct; 61(10):17-28
For the clinician, statistical significance can be considered the first step in interpreting differences between treatments. If there is no statistical difference between treatments (and you have verified that enough participants were included to achieve the power needed to find a difference if one existed and that extreme variability in the data caused by poor outcome measures or a poorly selected subject group didn't impact the ability to see a difference), then you can proceed confidently knowing that these two particular treatments aren't different. This certainly would influence clinical decisions and recommendations. In terms of published and presented data, we generally encounter treatments that have been found to be significantly different. After all, most treatment studies are designed to find differences that promote a particular new technology or signal processing scheme. Interestingly, clinicians would like to know when treatment results aren't different from each other, but non-significant findings are often difficult to publish and are rarely in the interest of manufacturers to present. There are times, however, when a finding of no significant difference would be of interest and would be the goal of the investigation. We were involved in a study like this when we wanted to compare a new auto-adaptive directional hearing aid to a "gold standard." In this study, we were comparing new treatments (auto-adaptive directional hearing aids) to how normal- hearing individuals function (gold standard) in the same listening environment. So, not being different from normal-hearing individuals would be considered an excellent outcome. As technology advances, I think we'll see more studies using normal function as the comparison condition, so we can expect to see more results where no difference is very positive and will help the clinician make recommendations of specific technology. It would be very motivating to be able to tell patients that a group of people similar to them using a particular signal processing scheme functioned as well as normal-hearing individuals.
Statistical-analysis; Measurement-equipment; Hearing; Hearing-disorders; Hearing-loss; Hearing-protection; Hearing-tests; Equipment-design; Protective-equipment; Biological-function; Sensory-disorders
Dr. Palmer at Audiology, 4th floor, Eye and Ear Institute, 203 Lothrop Street, Pittsburgh, PA 15213
Journal Article; Academic/Scholarly
The Hearing Journal