Advances and Applications in Statistics
Volume 10, Issue 2, Pages 207 - 218
(December 2008)
|
|
ON A THRESHOLD MODEL FOR ESTIMATING RESPONSE BIAS AND ITS APPLICATION TO CHOICE REACTION TIME DATA
Jennifer L. Asimit (Canada), W. John Braun (Canada) and William A. Simpson (UK)
|
Abstract: Choice reaction time experiment data involves two types of input and output. Such data involves a latent component since not all input events can be matched to the output process, and it is not known which input events elicit an output event. To study response bias we introduce a two-response threshold model for data from a point process version of the experiment. The two threshold estimates allow us to make inferences regarding a bias toward one response type, regardless of any differences between the two stimulus rates. This model also enables inferences regarding inhibition/facilitation effects of the flashes on the responses. This approach is applied to experimental data with nine different flash rates. |
Keywords and phrases: intensity function, maximum likelihood estimation, ordinal logistic regression, point process, threshold model. |
|
Number of Downloads: 341 | Number of Views: 1185 |
|