作者: David Andrich
DOI: 10.1111/J.2044-8317.1978.TB00575.X
关键词: Econometrics 、 Statistics 、 Feature (machine learning) 、 Explicit model 、 Likert scale 、 Latent trait 、 As is 、 Style (sociolinguistics) 、 Mathematics 、 Binomial (polynomial)
摘要: A binomial logistic latent trait model for attitude items constructed and scored in the Likert tradition is proposed. Unconditional maximum-likelihood estimates subject item parameters are examined applied to both simulated real data. Advantages data analyses which accrue from having an explicit questionnaires considered with one particular feature being a ready disclosure as whether or not neutral category functions expected.