不同铆测验设计下多维IRT等值方法的比较

实际应用中测验往往具有多维结构, 如果仍采用单维IRT方法进行等值, 会得到不准确的结果。因此对于多维结构的测验, 需要使用多维IRT等值方法来实现参数的转换。基于共同题设计, 文章通过模拟研究的方法, 考察了不同铆测验设计下几种多维IRT等值方法的表现, 同时考虑了测验长度、两个维度题目数量的比例、铆测验长度、铆测验的选择策略、两个维度之间的相关和等值群体的能力水平差异六个因素的影响。所比较的多维IRT等值方法有：均值/均值(MM)方法, 均值/标准差(MS)方法, Stoking-Lord (SL)方法, Haebara (HB)方法, 最小平方(LS)方法。结果显示：(1) SL, HB和LS方法得到的等值误差均方根最小, 且在各条件下表现较为稳定。(2) MM和MS方法在非等组条件下呈现出很大的误差均方根。(3)铆测验设计对SL, HB和LS方法的等值结果没有显著影响。(4)在两个维度之间的相关较高, 测验长度和铆测验长度较长, 等值群体的能力水平没有差异的条件下, SL, HB和LS方法得到的等值误差均方根最小。     

Right-hemispheric language evidence from cortical stimulation

While the left-hemispheric dominance for language in most right-handers is unquestionable, clinical observations suggest that the right, nondominant hemisphere has a considerable capacity for language. Under certain neuroanatomical conditions, the right-hemispheric capacity becomes functional and, thus, measurable. This study focuses on evoked language responses during electrical stimulation of the right hemisphere in three left-dominant patients undergoing temporal lobe resections for medically intractable seizures.     

Teacher attitudes toward stuttering

Attitudes of 100 elementary school teachers toward stuttering were studied using the Teacher Attitudes Toward Stuttering (TATS) Inventory. Teacher attitudes, as indicated by TATS Inventory scores, were compared with teacher knowledge of stuttering, age of teacher, number of years teaching experience, educational level, and whether the teacher had personal experience with a stutterer. Results indicated that significant positive correlations existed between teacher attitudes toward and knowledge of stuttering. Significant negative correlations were demonstrated between teacher attitudes/knowledge of stuttering and the presence of a stuttering child in the classroom. The applicability of the TATS Inventory as a clinical instrument in its current form and the research needed to enhance its flexibility are discussed.