Abstract
The author of the above article was informed that the entries of the approximate posterior standard deviations in Table 1 were too small and that the effects of the prior distributions should be lower than those shown in Table 1 (F. J. Abad, personal communication, October 13, 2009). The author found out that the FORTRAN implementation of Equation 6 under the one-parameter logistic model contained an error. Hence, the earlier Table 1 on page 264 and Figure 1 on page 272 should be replaced with the following new Table 1 and Figure 1, respectively. In addition, the fourth sentence on page 272 should be replaced with the following sentence: “Under Prior Specification 2 (e.g., the default setting of BILOG), assuming that the item discrimination is 1.5 and item difficulties are located within a reasonable range from -2 to 2, a sample of 2,500 is needed for the one-parameter logistic model in the Bayes modal estimation with the normal prior on b, and a sample size of nearly the same but slightly more than 2,500 is needed in the maximum likelihood estimation for the one-parameter logistic model.” The author thanks Professor Abad at the Universidad Automona de Madrid, Spain, for his careful reading of the article.
Get full access to this article
View all access options for this article.
