Abstract
After describing the main features of TeX-Ray, the author reports the result of an experimental study performed to validate the data set which was generated by TeX-Ray from editorials of four major newspapers. The validation was performed by regressing the TeX-Ray-generated data set on poll support ratings for Koizumi cabinet, during the period from April, 2001 through September, 2004. The regression analysis demonstrated an existence of unusually strong systematic relationship between editorial treatments of the cabinet and the cabinet support ratings, strongly indicating that the TeX-Ray-generated data set was a valid one.
