| Research Abstract |
Rating judgment and choice, which are ubiquitously used in psychology, should not be conceived of as mere picking-up of numbers from psychological continuum. They are the end-products of vacillating and constructive decision processes on a time dimension. In order to construct process models of rating judgment and choice, models that highlight response times and rating trajectories are proposed.
A) Review of previous models. From the viewpoint of psychological process model, studies on rating judgment and choice are summarized. A review paper will be completed shortly
B) Proposal of an original model Poisson-difference model(PDM) is proposed. Quantitative and qualitative properties of PDM are contrasted with other existent models and a parameter estimation method for PDM is developed. It was a great theoretical advance that mathematical relationship between PDM and Levi process was brought to light.
Another progress was the development of the process model for Cloze Test with Predetermined Alternatives(CTPA), a test format used often in Japanese schools.
C) Parameter estimation Theory and practice must go together. A parameter estimation program for PDM on MS-EXCEL is developed. As for CTPA, an EM-algorithm based program for true value estimation is implemented on MS-EXCEL as well.
D) Experimental study Based upon the results of A), B), and C), concordance of data and theory was checked. Dynamic rating scale(DRS) was proposed by Shiina(2004) to monitor the rating behavior more directly. In DRS, subjects are asked to make category ratings for stimulus presented on a PC screen using a PC mouse with the trajectories of cursor recorded. The basic fact found was that the trajectories often vacillated between rating categories. The vacillation looked smaller when the target sentences were simple math problems than when the sentences express abstract and ambiguous ideas, indicating that the vacillation reflects subjects' internal deliberation and hesitation.
|
