Response to Johnson and Danhauer

We thank Johnson and Danhauer for pointing out our oversight of calling the ED-NST the CUNY-NST. The nonsense syllable test that we used in the Auriemmo et al (2009) study was indeed the Edgerton-Danhauer Nonsense Syllable Test (ED-NST), which we have referenced correctly throughout the manuscript. We also agree with Johnson and Danhauer on the requisite considerations in administering an open-set nonsense syllable test to children. As these considerations are relatively obvious, we assumed all researchers are aware of and account for them in their study designs. Johnson and Danhauer inquired into the articulation skills of the study participants and the administration of the ED-NST. Like many careful researchers, we were aware of the potential influence that production skills may have on the performance on an open-set test. Thus, we did what Johnson and Danhauer did to make sure that performance on the NST was indeed perceptual and not biased by the articulation skills of the children. All the children in the study were administered the Goldman-Fristoe Test of Articulation by an experienced (20 years) diagnostic speech-language pathologist. This served as a baseline measure and allowed her to determine if the children’s articulation skills were adequate for inclusion into the study. Two children exhibited no articulation errors. The errors observed for the remaining eight children were mainly /w/ and /r/ substitutions and errors with blends. While the articulation errors were noted on word and sentence levels, all the children were able to produce the consonant and vowel sounds in isolation when presented in a CV and VC context. Thus, they were able to produce all the speech sounds in isolation. We indicated that fact on p. 293 (lines 6–7) of the manuscript. We cannot imagine how any reasonable research study would not have considered that potential variable and accounted for it in its design. Johnson and Danhauer rightfully pointed out the potential difficulties in scoring an open-set test based on auditory input alone. In administering this test to our pediatric subjects, two clinicians were actively involved. An audiologist, outside the test booth, controlled the stimulus presentation. The speech-language pathologist inside the booth and facing the child scored the child’s responses. In addition, responses were audio-recorded for later validation by the speech-language pathologist. This, we felt would ensure the accuracy and validity of the scoring. This has been our standard protocol when testing pediatric subjects for some time. Indeed, even in our adult test protocol, the study clinician always faced the test subjects as the test subjects spoke into a hanging microphone inside the booth. Johnson and Danhauer questioned our testing at 30 and50 dBHL levelswithoutensuring audibility.Wewere somewhat perplexed by their comment. First, we have ensured that the default (or non-LFT) program met the standard audibility requirement by matching its output for a soft (55 dB SPL) and moderate (70 dB SPL) speech input to the DSL5.0 target (p. 294, fourth paragraph in left-hand column). In addition, the last paragraph in the ‘‘Hearing Aid Fitting’’ section in the right-hand column of p. 294 also described how we ensured that both the default (no-LFT) and LFT programs were subjectively acceptable at different input levels. Secondly, by testing at 30 and 50 dB HL levels, we were indeed obtaining a performance-intensity function, as recommended by Johnson and Danhauer, albeit at only two levels and using a normal hearing reference. However, considering that the aided thresholds of all participants were between 5 and 20 dB HL at 500 Hz (Figures 2 and 3), it was clear that all the children could indeed hear the stimuli at the 30 dB HL presentation level. We are most perplexed by Johnson and Danhauer’s assertion that we compared the NST results between the unaided and aided conditions. We never did. All the NST testing and comparisons were made between the LFT and the no-LFT (and own aid) in the aided condition only. Since the frequency-gain characteristics of the LFT and no-LFT conditions were the same, a comparison between LFT over the no-LFT condition must be a direct result of the LFT processing. Our rationale for testing at the 50 dB HL input level was to examine how LFT compared to noLFT at a typical conversational level. Our rationale for testing at the 30 dB HL level was to examine if input level would have altered the noted benefit. In other words, we want to examine if there is any additional audibility benefit provided by LFT. This is in line with our long-held belief that a hearing aid’s full benefits (including its features, such as a low compression threshold) cannot be examined simply at the conversational level. In our manuscript, not only did we demonstrate that LFT worked, we also revealed a greater magnitude of LFT improvement in the 30 dB HL condition over the 50 dB HL condition. Audibility of the softer sounds definitely played a part in the improvement—and that audibility benefit resulted from LFT processing. Johnson and Danhauer questioned our thoroughness on the presentation of our methods section and the Journal for ‘‘missing the points raised...’’ in their letter. We can appreciate Johnson and Danhauer’s motives given their unique involvement with the ED-NST. Perhaps we could have presented all the minute details of our study. We did not, because we felt such details would not be interesting for the majority of readers. There are also practical considerations such as page limitations in a printed journal, making this all but impossible. Thus, we reported what we thought was important and what we thought the readers would not already know. We assumed that any readers interested Journal of the American Academy of Audiology/Volume 20, Number 10, 2009