語言障礙學生口語說明能力之探討：語言層面的分析

Investigating performance of expository discourse in Mandarin-speaking children with language impairment: Language sampling analysis

口語說明能力是國小高年級學童的高層次口語能力，對學科學習成就有重要的影響，近年來受到不少國外研究的關注。相較之下，國內關於國小高年級以上一般學生和語言障礙學生口語說明能力的探究相當缺乏。本研究旨在探討國小高年級語言障礙學生和一般學生的口語說明能力，並探究能區分語言障礙的口語說明能力指標。研究者以改編之遊戲或運動作業與提示表收集語言樣本，進行語意、語法、錯誤及迷走語之分析。主要發現如下：（1）語意方面，語言障礙學生口語說明時使用之「總詞彙數」、「相異詞彙數」與「校正後相異詞的出現率」均顯著低於一般學生。（2）語法方面，語言障礙學生說明時的單句數、被字句數、平均主謂句長度皆顯著少於同儕。（3）錯誤分析方面，語言障礙學生在說明時單句中錯誤比率顯著較高。在個別錯誤的類型分析中，則是以「單句錯誤使用語詞的比率」和「單句語意關係不明句比率」等語意層面的錯誤明顯高於同齡學生。（4）迷走語方面，兩組學生在「迷走語詞佔總詞彙比率」、「單句迷走語數」、「單位迷走語詞彙數」和「單句放棄語句比率」等計量分析結果沒有顯著差異。（5）能區分兩組學生之語言指標分析結果顯示，以「相異詞彙數」、「平均主謂句長度」、「被字句數」之組合性指標有較佳之區分正確性。最後，根據研究發現，提出研究及實務方面的相關建議。

Purpose: Expository discourse is challenging for school-age children and has recently gained more attention because of its crucial role in academic learning. However, the number of studies on the expository language abilities of Mandarin-speaking students is limited. In Taiwan, the most commonly used method for determining whether students have difficulties with oral expression is to administer standardized tests. However, for some students with special needs who cannot be tested using standardized procedures, the validity of test results is contentious. In addition, tests that are used to assess the oral expression skills of senior elementary school students are limited and are typically used to determine the semantic and syntactic aspects at words and utterance levels. Furthermore, Taiwan does not yet have any standardized tests or procedures for assessing children’s performance in expository discourse. Therefore, in an expository context, the analysis of language samples of students with and without language impairment is necessary to fill the gap between the standardized tests and language samples in order to provide a comprehensive profile of individuals’ language development and disorders. To evaluate the expository language performance of Mandarinspeaking students, this study compared the language performance of school-age students with language impairment (LI) and chronological-age (CA) peers. In addition, we evaluated the effect of a language-sampling task on language deficits in students with LI and explored language performance indicators for differentiating Mandarin-speaking students with LI from CA peers. Methods: In total, we enrolled 40 elementary school students from grades 5 and 6 (aged 10 years; 0–12.5), comprising 20 students with LI and 20 students with normal language development. The teachers of these students referred students with LI with a formal or suspected diagnosis of language or learning disabilities in accordance with the policy of the Identification and Counseling Committee for Special Education in Keelung City, Taichung City, Taipei City, and New Taipei City. The inclusion criteria were as follows: (1) having an intelligence quotient (IQ) of ≥75 in accordance with Raven’s Standard Progressive Matrices Test; (2) having a score lower than −1.25 standard deviations of the normative mean of any two of the five following tests: the “Vocabulary” and “Similarity” subtests on the Wechsler Intelligence Scale for Children (Fourth Edition) and the total language development score, language comprehension score, and language expression score on the Revised SchoolAge Language Scale. Subsequently, 20 students with LI were included in this study (15 boys and 5 girls). Another group of 20 CA students was matched with adjustment for age and sex to the group of 20 students with LI and were recruited from the general education classes. In the two groups, we excluded students with gross neurological, sensory, motor, or intellectual impairment. Expository language samples were collected from 40 students using the revised “favorite game or sport protocol” (proposed by Miller et al., 2011; Nippold et al., 2005). The students were required to name a game or sport they liked the most and explain why they liked the sport/game. To encourage the students to produce comprehensive oral expository passages, a prompt sheet, which was adapted from Expository Scoring Scheme (Heilmann & Malone, 2014), was used. The prompt sheet included eight dimensions (object, preparations, start, course of play, rules, scoring, duration, and strategy), and a brief description for each dimension was listed as notes. The language samples were collected, segmented into the units of simple Chinese sentences, and subsequently analyzed using semantic, syntax, error, and maze measures. Statistical analyses were conducted using SPSS software, version 22.0. The students’ basic characteristics, namely age, sex, parents’ education level, and nonverbal IQ, were analyzed using a chi-square test and a two-sample t test to examine betweengroup mean equality. A series of one-way analyses of variance were performed to examine differences in language measures elicited in the expository task between the groups. Finally, a discriminant analysis was performed to determine the language performance indicators for differentiating Mandarin-speaking students with LI from their CA peers. Results/Findings: The results showed that the students with LI performed significantly poorer on some language measures, including semantic measures (i.e., total words, number of different words, and corrected type–token ratio), syntax measures (i.e., number of simple Chinese sentences, number of “bei”-sentences, and mean length of subject–predicate sentences), and error measures (i.e., percentage of total errors per simple Chinese sentence, percentage of word errors, percentage of unclear contexts). The discriminant analysis revealed that, compared with a single measure, multiple measures, including the number of different words and mean length of subject–predicate sentences and bei-sentences, had a higher accuracy of distinctive prediction for differentiating the LI from the CA students. Conclusions/Implications: This study demonstrated that students with LI exhibited deficits in the semantic and syntactic domains of expository discourse. In addition, we found that the revised favorite game or sport task can be used to effectively elicit expository language samples from school-age Mandarin-speaking students with LI and CA peers. Furthermore, this study revealed that specific language measures for Mandarin (i.e., number of simple Chinese sentences, number of bei-sentences, mean length of subject–predicate sentences, and percentage of total errors per simple Chinese sentence) can be used to differentiate the LI group from the CA group. Because expository discourse is a key domain for examining individuals’ oral expression abilities, this study addressed the gap in the current knowledge regarding spoken language performance and deficits among senior elementary school students with LI. However, the research on expository discourse in Mandarin-speaking students is quite limited. Additional studies are required to examine the overall development of oral expository discourse among students of different ages, to establish oral expository discourse developmental courses in Mandarin and explore the cutoff with empirical validity. Therefore, the study findings can be used to supplement the standardized tests for detecting students with language impairment. In addition, the findings on students’ expository language performance can be used to integrate expository discourse into communication training courses or language intervention projects to improve the oral expression skills of students with LI.