Development of a turkish monosyllabic word recognition test for adults

The aim of this study was to develop a Turkish speech recognition test based on the idea that speech tests should be developed in accordance with the specifications of each individual language. For this purpose, primarily, the most used words of a selected Turkish corpus were chosen to create the lists used in this study. Words are also checked through `Turkish Language Association Dictionary? and Written Turkish Word Frequency Dictionary? in terms of their meanings. Semantic words those are present in both dictionaries were selected. In order to test the familiarity of the words used, the word lists were given to 380 participants from different age groups and educational backgrounds and they were asked to score the familiarity of these words. According to the participants? judgments, unfamiliar words were eliminated. Thirty-six normally hearing young adults aged between 18 to 30 were divided into two groups and were asked to listen the remaining word lists twice in six different intensity levels (in total 12 different intensity levels) from a recorded male voice. Words which are used rarely and most frequently were identified and eliminated in order to provide the homogeneity. Three different word lists, each composed of 50 phonemically balanced words, were developed to be used in speech recognition tests. Subsequently, these lists were divided into two groups considering the phonemic balance criteria and six lists, each composed of 25 words were developed. In order to test the balance among lists, logistic regression analysis and chi square test were performed. For lists composed of 25 and 50 words, c2 (5) =0.322 and p=0.666, c2 (2) = 0.313 and p=0.885 indicated no statistically significant difference. Besides, the internal reliability of each list was analyzed using KR-20 and was found to be above 98% for all the lists. The mean slope for logistic regression analysis was found to be 5.4 %/dB. Thus, subsequently developed 25 and 50 words lists composed for Turkish Language were ascertained to be appropriate for use in this study. For further research, to test their capability of distinguishing different pathologies, the lists should to be applied in clinical settings to patients with hearing loss. Thus, the differential diagnostic value of the developed speech recognition test will be revealed.Key words: monosyllabic word, speech recognition test, speech audiometry, word lists