翁儷禎2006-07-252018-06-282006-07-252018-06-282004http://ntur.lib.ntu.edu.tw//handle/246246/9728本研究擬以模擬研究的方法探討信度類別與量尺特性對評定量表相關係數校正的影響。模擬研究中擬操弄量表點數、量表內題目間的共變數、量表題數、受試者判斷的正確度（即個別題目之信度），跨量表題目真分數間的共變數，與點量表分配六個因素，比較以內部一致性信度α，再測信度，真分數與觀察分數相關的平方等三種信度係數校正之相關係數之差異。結果發現以內部一致性信度α 校正之相關係數在題目同質性低時易高於理論值，其他兩類信度校正之相關係數的偏誤則較小。application/pdf86717 bytesapplication/pdfzh-TW國立臺灣大學心理學系暨研究所李克式量表內部一致性信度再測信度效標關聯效度相關係數校正The Monte Carlo experiment examined the effects of the type of reliability and the properties of Likert-type rating scales on correlations corrected for attenuation. The independent variables manipulated included number of response categories, inter-item covariance, number of items, respondents’ judgment accuracy (or reliability of individual item), correlations between true score items across measures, and distribution of scores of the rating scales. The quality of the Pearson product-moment correlations corrected for attenuation by the internal consistency reliability α, the test-retest reliability, and the squared correlation between true scores and observed scores were compared. The correlation corrected by internal consistency reliability α was found to be higher than the population value when items were less homogeneous. The correlations corrected by the other two types of reliability, in general, yielded less biased results.Likert-type scalesinternal consistency reliabilitytest-retest reliabilitycriterion-related validitycorrection for attenuationreporthttp://ntur.lib.ntu.edu.tw/bitstream/246246/9728/1/922413H002018.pdf