Black and Wiliam (1998a, 1998b) demonstrate that formative assessment is one of the most effective strategies for promoting student learning. Since the publication of their reviews, formative assessment has gained increasing international prominence in both policy and practice.
 However, despite this early innovation, the theory and practice of formative assessment are currently at a crossroads. It is widely understood that problems emerge when formative assessment is being reduced to a mini-summative assessment or to a series of teaching techniques for coaching to improve grades and levels.
 On the one hand, a serious threat to the effectiveness of formative assessment occurs when it is assimilated into larger accountability systems such as National Curriculum Assessment in England. On the other hand, a defense of formative assessment is offered by some researchers who suggest that the threat stems from misinterpretation of the evolved form of formative assessment.
 In this paper, although I am alert to the rich potential of the evolved form of formative assessment, I suggest that the threat stems not from “misrepresentation” but that it exists in the original theoretical framework of Black and Wiliam and the early ARG definition of Assessment for Learning. I will illustrate that this type of formative assessment becomes “convergent assessment” (Torrance & Pryor, 1998) and identify the widespread notion of “closing the learning gap”(Sadler, 1989) as the mechanism of “convergent assessment”. I also claim that formative assessment characterised by “convergent assessment” can lead to the practice of “criteria compliance” (Torrance, 2007). Together these claims point towards the theoretical problems of the evolved form of formative assessment and lead to a discussion of the main dilemmas for formative assessment: the kind of learning that is taking place, the effects of explicit learning objectives, the tension of accountability pressure and high-stakes summative assessments. Finally, by applying the suggestions above, this paper provides a critical analysis of recent assessment policy in Japan, emphasizing criterion-referenced approach in classroom assessment and proposes a pathway for developing formative assessment further.

抄録全体を表示

 In this study, we test the assimilation thesis by comparing the academic achievement between native students and first and second generation immigrant pupils. It is the first empirical study that systematically analyzes the native-immigrant achievement gap in Japan. Although numerous studies have examined the achievement gap, most of them are based on small-scale case studies and have failed to test the effects of multiple factors simultaneously, using large-scale nationally representative data. Since the number of immigrant (foreign) students is relatively small in Japan, we constructed a pooled dataset of PISA by combining all five waves from 2000 through 2012. The dependent variable is the test score in reading literacy, and we tested the effects of three key independent variables: immigrant generation, parental socioeconomic status and language spoken at home. A multilevel analysis was performed to examine both individual and school-level variations, followed by a multiple imputation method to deal with missing values of parental socioeconomic status. The major findings are three-fold. First, first generation immigrant students perform more poorly in reading literacy, but there is no significant difference between second generation and native students. Second, parental socioeconomic status has a positive effect on academic achievement, but the effect is not robust enough to mediate the impact of immigrant generation. Third, Japanese spoken at home is an important determinant of the native-immigrant gap in academic achievement. The same results were obtained for mathematics and scientific literacy test scores. Although these findings echo previous studies, they underscore the importance of language use at home. Our empirical results suggest that it is important to expand opportunities for Japanese language learning for both immigrant students and parents.

抄録全体を表示

 In recent years, teachers’ increased workloads have become an issue for policy, and have been multiply pointed out, deriving as they do from peripheral duties such as paperwork, in academic research as well. However, these mentions have not been based on sufficiently solid proof. Here, this paper compares teacher working hours surveys extant from the 1950s-1960s and from the late 2000s, using a general linear model. Results show that it is not necessarily the case that teachers in the late 2000s are spending more time on peripheral duties like paperwork than their 1950s-1960s counterparts, but that the time they spend on educational activities (in particular, extracurricular activities) is longer.

抄録全体を表示