OBJECTIVE: To develop 22 Chinese Mandarin monosyllable lists with good psychometrical equivalence. This study was to evaluate the test-retest reliability of these lists when it was used in speech recognition test in normal hearing dialectal speakers. METHODS: Seven cities including Dalian, Shanghai, Hangzhou, Wuhan, Guangzhou, Fuzhou and Xiamen were selected as testing centers which contain 6 typical Chinese dialectal regions including north of China, East of China, north of Fujian, south of Fujian, Guangdong and mid-south of China. At each center, 22 local normal hearing people were selected to join this study. Every participant was tested by each recognition test of all 22 lists twice in two sessions and same test order respectively. The second run of testing was carried out within 10 days-1 month since first run of testing. RESULTS: There was a significant correlations between scores obtained at the two sessions (r = 0.682, P < 0.01). Paired student-t test had shown that a gross score of all dialectal participants was significantly higher than that of initial test to retest (P < 0.01). The mean increment of score was (2.7 +/- 10.1)%. A significant difference of test-retest score in 7 sites was 19.8% and it was equal to 5 test items. A one way ANOVA analysis had indicated that there were statistically significant difference between the score improvement of 7 test sites (P < 0.01). Another analysis had shown that there was no significant correlation between test-retest score improvement and intra-session intervals (P = 0.947). CONCLUSIONS: Mandarin monosyllabic recognition test seems to be more stable, and the present study has indicated a systematic differences in Chinese Mandarin monosyllable recognition scores between test and retest. Monosyllable recognition test is not susceptible to memory effect. Pearson's correction analysis is not suitable to evaluation for test-retest reliability.
OBJECTIVE: To develop 22 Chinese Mandarin monosyllable lists with good psychometrical equivalence. This study was to evaluate the test-retest reliability of these lists when it was used in speech recognition test in normal hearing dialectal speakers. METHODS: Seven cities including Dalian, Shanghai, Hangzhou, Wuhan, Guangzhou, Fuzhou and Xiamen were selected as testing centers which contain 6 typical Chinese dialectal regions including north of China, East of China, north of Fujian, south of Fujian, Guangdong and mid-south of China. At each center, 22 local normal hearing people were selected to join this study. Every participant was tested by each recognition test of all 22 lists twice in two sessions and same test order respectively. The second run of testing was carried out within 10 days-1 month since first run of testing. RESULTS: There was a significant correlations between scores obtained at the two sessions (r = 0.682, P < 0.01). Paired student-t test had shown that a gross score of all dialectal participants was significantly higher than that of initial test to retest (P < 0.01). The mean increment of score was (2.7 +/- 10.1)%. A significant difference of test-retest score in 7 sites was 19.8% and it was equal to 5 test items. A one way ANOVA analysis had indicated that there were statistically significant difference between the score improvement of 7 test sites (P < 0.01). Another analysis had shown that there was no significant correlation between test-retest score improvement and intra-session intervals (P = 0.947). CONCLUSIONS: Mandarin monosyllabic recognition test seems to be more stable, and the present study has indicated a systematic differences in Chinese Mandarin monosyllable recognition scores between test and retest. Monosyllable recognition test is not susceptible to memory effect. Pearson's correction analysis is not suitable to evaluation for test-retest reliability.