普通话儿童词汇相邻性测试词表总体信度评估

发布时间：2018-04-29 11:05

本文选题：言语测听 + 总体信度　；参考：《首都医科大学学报》2011年06期

【摘要】：目的对等价性的普通话儿童词汇相邻性测试词表(mandarin lexical neighborhood test,M-LNT)进行可行性、可靠性以及效能的评估,通过词表的得分-强度(performance-intensity,P-I)函数曲线、言语识别阈以及临界差值等指标,以期获得M-LNT词表的总体信度,为临床听力障碍患者的使用建立基准曲线。方法选取听力正常者52名,对双音节难、易词表和单音节难、易词表的4类12张词表进行测试。实验先期获得材料的P-I函数曲线的6个测试强度,依次采用随机区组设计方法,受试者随机分到奇数位或偶数位强度组,测试顺序为先易后难,3张表顺序轮换。获得每一强度下的识别率得分,用Matlab 7.1软件进行P-I曲线拟合、计算曲线的斜率、言语识别阈和临界差值。结果 4类词表的每类词表间的结果差异无统计学意义(P0.05),表间等价。对拟合P-I曲线分析得出单音节易词表、单音节难词表、双音节易词表和双音节难词表平均阈值分别为25 dB SPL、31 dBSPL、20 dB SPL、24 dB SPL,50%斜率分别为2.8%/dB SPL、3.0%/dB SPL、5.8%/dB SPL、4.1%/dB SPL,95%置信区间下的总体临界差值分别为±22.2%、±26.1%、±32.7%、±29.9%。结论本实验以普通话儿童词汇相邻性测试词表为测试材料,4类词表的总体临界差值分别为±22.2%、±26.1%、±32.7%、±29.9%,总体信度可靠,并且获得P-I曲线,建立了M-LNT词表言语识别率的基准曲线。
[Abstract]:Objective to evaluate the feasibility, reliability and efficacy of the equivalent mandarin lexical neighborhood test (M-LNTs) of Mandarin children's lexical contiguity test, and to evaluate the score and strength of the lexical table by using performance-intensityP-I function curve, speech recognition threshold and critical difference, etc. The purpose of this study was to obtain the overall reliability of the M-LNT lexicon and to establish a baseline curve for the use of clinical hearing disorders. Methods A total of 52 subjects with normal hearing were selected to test 12 word tables of 4 categories which were difficult to use dicyllabic words monosyllabic and monosyllabic words. Six test strengths of P-I function curves of materials were obtained in advance. The random block design method was used in turn. The subjects were randomly divided into odd or even digit strength groups. The test order was in the order of three tables: first easy and then difficult. The recognition rate score of each intensity was obtained. P-I curve fitting was carried out with Matlab 7.1 software. The slope, speech recognition threshold and critical difference of the curve were calculated. Results there was no significant difference in the results of each category of lexicon among the four categories of lexicon (P 0.05), but the difference between the tables was equivalent. The analysis of fitting P-I curve shows that monosyllabic words are easy to be used, and monosyllabic words are difficult to be used. The average threshold values of the double syllable word table and the double syllable word table are 25 dB / 31 dB / 20 dB / 24 dB / L / 50% respectively, and the overall critical difference is 卤22.2dB, 卤26.1, 卤32.7, 卤29.99.995%, respectively, with a slope of 5.8dB / L / d SPLL = 5.8dB / d / d SPLL = 4.1dB and 95% / 95% confidence interval, respectively. Conclusion in this experiment, the overall critical difference of the four categories of lexical lists using the contiguity test of Putonghua vocabulary as the test material is 卤22.2, 卤26.1, 卤32.7, 卤29.9, respectively, and the P-I curve is obtained, and the benchmark curve of the speech recognition rate of the M-LNT vocabulary is established.
【作者单位】：首都医科大学附属北京同仁医院耳鼻咽喉头颈外科;北京市耳鼻咽喉科研究所教育部耳鼻咽喉头颈外科重点实验室;
【基金】：国家自然科学基金(30872859) 北京市自然科学基金(7093117) 国家科技支撑计划(2008BAI50B01) 教育部人文社会科学研究项目基金(09YJA880096)~~
【分类号】：H193.1

【参考文献】