We analyzed the changes and characteristics of the trends of examinees during the transition period from the National Center Test to the Common Test for University Admissions. As for the movements related to the Common Test, the number of new graduate private university applicants has decreased remarkably, especially in the Tokyo metropolitan area. As of 2021, non-new graduates were in the center test generation, so they had to give up taking the Common Test. In addition, it is thought that the worsening of the 18-year-old population in rural areas led to a sharp decrease in the total number of non-new graduates. Due to COVID-19, the number of people who did not take the test on the day doubled. It became clear that a certain number of applicants do not essentially need to take the Common Test.

Item response theory (IRT) models play an important role in developing item banks when administrating large-scale standardized tests. To develop equity-focused standardized tests, it is essential that all items which are presented to focal examinee groups are unique. This study considers a case where item parameters are equated on the base scale of a field test. An anchor test design with monitor examinee groups is assumed to estimate equated item parameters in an item bank. However, the accuracy of equated item parameters has not been fully investigated among equating methods. Three equating methods were compared in this study. The results suggested that the fixed common item parameter (FCIP) method as well as the concurrent calibration method with separate calibration performed well compared to the separate calibration method only when the difference of ability distribution between monitor and focal group was large, or the number of monitor examinees was small. The results also suggested that performance was better when the number of anchor items was large.

With the massification of higher education in Japan and China, making rational selection of universities is an important issue for many young people in both countries. Japan has diversified its entrance examination system after years of reform. In China, on the other hand, the diversification of the entrance examination system is not as advanced as in Japan. It is interesting to see whether there are differences in the university decision making behaviors of high school students in Japan and China, where the degree of 'diversification' is different. This study aims to compare the actual university decision making behaviors of high school students in China and Japan as of 2019. A questionnaire survey was conducted focusing on the information gathering behaviors of high school students. The results reveal significant differences in the information gathering behavior between Japanese and Chinese high school students when selecting universities. The Japanese high school students tended to gather information earlier and more frequently than the Chinese high school students and appeared to confidently make their decision on which university to attend with clear reasons. This outcome might be due to differences in the university admission policies and resultant career guidance systems of the two countries.

In this paper, we constructed panel data by combining questionnaire survey data conducted on university students and academic data provided by the university. We analyzed the determinants of academic performance in university students by the fixed-effects model. We observed that the number of credits earned was positively influenced by the relationship with friends, while the influence of learning attitude was limited. On the other hand, the average score was negatively influenced by the relationship with friends, but positively influenced by the learning attitude. These results suggest that while the average score is supported by the learning attitude, the number of credits earned is not necessarily so. This implies that using the average score rather than the number of credits earned is more appropriate when accurately assessing the total amount of knowledge and skills of university students.

In career guidance, using the results of occupational aptitude tests is essential for searching occupational information. The Ministry of Health, Labor, and Welfare (MHLW) launched the Japanese version of Occupational-Network website (hereafter J-O-net) for jobseekers in 2020, where they published occupational database of around 500 occupations. The authors were requested to develop a module of ability test on the J-O-net for searching occupational information matching test takers’ aptitude. 3 subtests (Spatial, Verbal, and Numerical) of the General Aptitude Test Battery (GATB) were selected for developing the online module. The authors examined the test-retest reliability and confirmed some evidences of validity by examining the correlation of the subtests between the web version and the original one. A new matching mechanism for connecting the test scores and related occupational groups was investigated by attempting cluster analysis of numerical information database of the J-O-net, and resulted in the establishment of 8 occupational groups. Some challenges for the future development were examined.

This research reviewed articles that developed parameter estimation methods in item response theory models, with special focus on recent articles published mainly in Psychometrika. As a result, various new methods to manage the computational burden problem with respect to the item factor analysis and multidimensional item response models, which have high dimensional factors, were introduced. Monte Carlo integral methods, approximation methods for marginal likelihood, new optimization methods, and techniques used in the machine learning field were employed for the estimation methods. Theoretically, a new type of asymptotical setting, that assumes infinite number of sample sizes and items, was considered. Several methods were classified apart from the maximum likelihood method or Bayesian method. Theoretical development of interval estimation methods for a latent trait parameter were also proposed and they provided highly accurate confidence intervals.