Beattie R C, Raffin M J
J Speech Hear Disord. 1985 May;50(2):166-78. doi: 10.1044/jshd.5002.166.
The Auditec recordings of the CID W-22 monosyllables were used to generate test and retest intelligibility functions on normally hearing listeners and subjects with mild-to-moderate sensorineural hearing loss. The normally hearing subjects were tested with 50-word lists at SPLs ranging from 15 to 50 dB. Lists of 25 words were used with the hearing-impaired group. The functions were analyzed to assess the reliability of threshold (50% point), slope (20%-80% points), and maximum intelligibility (PB max). The 50% point was obtained at 28 dB SPL for the normally hearing listeners and at a sensation level (SL) of 12 dB respondaic thresholds for the hearing-impaired group. Very stable monosyllabic thresholds were found because 95% of the test-retest values were within 6 dB for both subject groups. Slopes of 4.9% per dB and 2.7% per dB were obtained for the normally hearing and hearing-impaired groups, respectively. Fair reliability was observed; 95% of the test-retest values encompassed a range of +/- 1.9% per dB for the normally hearing subjects and +/- 1.1% per dB for the hearing-impaired group. Although group slopes provide useful information for selecting the range and step size for generating psychometric functions, the value of routinely obtaining slope on an individual basis has not been demonstrated. Even though the same word lists were used for both test and retest measurements, reliability closely agreed with predicted results based on the binomial theorem. In contrast to the large variability for a single list of 25 words, very stable PB max scores were found when two or three scores were averaged on the plateau of the function.
使用CID W - 22单音节词的Auditec录音，对听力正常的受试者以及患有轻至中度感音神经性听力损失的受试者进行测试和复测，以生成可懂度函数。对听力正常的受试者使用包含50个单词的列表，测试声压级范围为15至50分贝。对听力受损组使用包含25个单词的列表。分析这些函数以评估阈值（50%点）、斜率（20% - 80%点）和最大可懂度（PB最大值）的可靠性。听力正常的受试者在28分贝声压级时获得50%点，听力受损组在感觉级（SL）为12分贝时获得反应阈值。发现单音节阈值非常稳定，因为两个受试者组的95%的测试 - 复测值在6分贝以内。听力正常组和听力受损组每分贝的斜率分别为4.9%和2.7%。观察到可靠性尚可；听力正常受试者的95%的测试 - 复测值每分贝的范围为±1.9%，听力受损组为±1.1%。尽管组斜率为选择生成心理测量函数的范围和步长提供了有用信息，但尚未证明常规获取个体斜率的价值。即使测试和复测测量使用相同的单词列表，可靠性与基于二项式定理的预测结果非常吻合。与25个单词的单个列表的较大变异性形成对比的是，当在函数的平稳段对两三个分数进行平均时，发现PB最大值分数非常稳定。
