Conceptualising computerized adaptive testing for measurement of latent variables associated with physical objects

2014 Joint IMEKO TC1-TC7-TC13 Symposium: Measurement Science Behind Safety and Security

Camargo, F.R.^1 ; Henson, B.^1

School of Mechanical Engineering, University of Leeds, Woodhouse Lane, Leeds

LS2 9JT, United Kingdom^1

关键词: Affective engineering;    Computerized adaptive testing;    Financial benefits;    Independent variables;    Latent variable;    Physical features;    Physical objects;    Stochastic framework;   

Others  :  https://iopscience.iop.org/article/10.1088/1742-6596/588/1/012012/pdf
DOI  :  10.1088/1742-6596/588/1/012012

来源: IOP

【 摘 要 】

The notion of that more or less of a physical feature affects in different degrees the users' impression with regard to an underlying attribute of a product has frequently been applied in affective engineering. However, those attributes exist only as a premise that cannot directly be measured and, therefore, inferences based on their assessment are error-prone. To establish and improve measurement of latent attributes it is presented in this paper the concept of a stochastic framework using the Rasch model for a wide range of independent variables referred to as an item bank. Based on an item bank, computerized adaptive testing (CAT) can be developed. A CAT system can converge into a sequence of items bracketing to convey information at a user's particular endorsement level. It is through item banking and CAT that the financial benefits of using the Rasch model in affective engineering can be realised.