家长和教师对儿童语言能力评估的对比研究

目的
验证我国家长和教师对儿童语言能力评估的准确性，比较二者差异，并分析造成差异的影响因素，以深化对两种语言评估来源特点的认知。

方法
采用横断面研究，于2024年9～12月选取南京市3所幼儿园和1所小学的280名3～7岁儿童为研究对象，分别由其主要照护家长和班级教师采用汉语普通话语言知识评估工具（ITALK-M）对词汇量、发音可辨识度、句子长度、语法合法性进行评分，儿童完成汉语普通话句子重复任务，作为实际语言能力评估指标。采用广义线性混合模型进行统计分析。

结果
儿童的实际语言能力得分与家长评分（rs=0.353, P<0.001）和教师评分（rs=0.347, P<0.001）均呈现显著正相关。只有教师评分有显著的年龄效应（大龄组高于小龄组, F=10.987, P=0.001）和显著的性别差异（女童高于男童,F=6.132, P=0.013），且该性别差异只体现在小龄组儿童中（F=4.294, P=0.038），而家长评分未体现上述差异。母亲教育水平与评分者存在显著的交互作用(F（1,2026）=6.403, P<0.001)，教师倾向于给母亲教育水平更高的儿童打出更高的分数，而家长评分中的高分更集中于母亲教育水平为中等（专科和本科）的儿童。家长评分显著高于教师评分（F=15.916,P<0.001），主要体现在发音可辨识度（t=5.343）和句子长度（t=4.188）两个语言维度，差异有统计学意义（P<0.001）。

结论
家长与教师的主观评估在儿童早期语言.筛查中都具有准确性。相较于家长评估,教师评估具有优势,在捕捉年龄发展趋势和性别差异方面更为敏感,且性别差异随着儿童年龄增长而逐渐减弱。家长可能对儿童语言能力存在高估倾向,尤其是中等教育水平的家长。

Objective
To examine the accuracy of parent and teacher ratings of children's language skills, to compare their differences, and to analyze the influencing factors, so as to deepen the understanding of the characteristics of these two sources of language assessment.

Methods
A cross-sectional study was conducted between September and December 2024 with 280 children aged 3 - 7 years recruited from three kindergartens and one primary school in Nanjing. A cross-sectional study was conducted between September and December 2024. A total of 280 children aged 3-7 years were recruited from three kindergartens and one primary school in Nanjing. Primary caregivers and classroom teachers rated the children's vocabulary, speech intelligibility, sentence length, and grammaticality using the Inventory to Assess Language Knowledge-Mandarin (ITALK-M). Children completed the Mandarin Sentence Repetition Task (MSRT) as an objective measure of language skills. Generalized linear mixed models (GLMM) were used for data analysis.

Results
Children's objective language scores showed significant positive correlations with both parent ratings (rs =0.353, P<0.001) and teacher ratings (rs =0.347, P<0.001). Only teacher ratings demonstrated significant main effects of age (older children were scored higher than younger children, F=10.987, P=0.001) and gender (girls were scored higher than boys. boys. F=6.132, P=0.013), and this gender effect was only found in the younger group (F=4.294, P= 0.038), whereas parent ratings did not show these patterns. There was a significant interaction between maternal education and rater (F=6.403, P<0.001). Teachers tended to give higher ratings to children whose mothers had higher education levels, whereas higher ratings of parents were concentrated among children whose mothers had mid-level education (associate or bachelor's degrees). Parent ratings were significantly higher than teacher ratings (F=15.916, P<0.001), particularly in the dimensions of speech intelligibility (t=5.343, P<0.001) and sentence length (t=4.188, P<0.001).

Conclusions
Subjective ratings from both parents and teachers demonstrated a certain degree of accuracy. Teacher ratings are more sensitive in capturing age and gender differences, whereas parents tend to overestimate children's language skills, particularly in the dimensions of speech intelligibility and sentence length. Additionally, maternal education levels exert differential influence patterns on the two assessment sources.