Noboru Takahashi

This paper reports the characteristics of a grammar-discourse test which was developed as a subscale of the ATLAN (Adaptive Tests for Language Abilities). ATLAN, which can be used via the Internet, is a collection of sub-scales used to assess different language abilities, mainly in middle childhood. In Study 1, we prepared 67 items from 8 grammar categories for elementary school children and 67 items from 12 grammar categories for preschool children. Participants were 309 children from the first to third grades and 258 preschoolers from 3-6 years of age. One hundred and twenty eight items were chosen based on the results of study 1, as an item pool for a grammar-discourse subscale of the ATLAN. In Study 2, 59 preschool children responded to the two subscales of the ATLAN, grammar-discourse and vocabulary, in addition to the LC Scale (Language Communication Developmental Scale; Otomo et al., 2008). The results of multiple regression analysis, in which the child's score of LC Scale was a dependent variable and the two ATLAN scores were independent variables, showed that 48% of the variance on the LC Scale was explained by the two ATLAN subscales. A discussion focused on the possibility of expanding the ATLAN tests.

The present research used item response theory to develop the Adaptive Tests for Language Abilities (ATLAN) for elementary school children. These tests, which are available on-line and also for stand-alone computers, are comprised of 2 subtests: vocabulary and kanji reading, both of which are componential language skills needed for Japanese reading comprehension. Children from preschool to junior high school (N =3,067) participated in determining the parameters of each item in the vocabulary subtest, whereas 4,463 children from elementary to junior high school participated in the development of the kanji reading subtest. Because the results of a simulation using the Adaptive Tests for Language Abilities adequately replicated the original data, the parameter estimation was judged to be reliable. The tests also have validity, because the results from each subtest moderately correlated with existing vocabulary and kanji tests. Advantages of the Adaptive Tests for Language Abilities are that they need fewer items than existing tests for measuring these abilities, and that different patterns of items are shown for different children. Copyright © 2005-2009 National Institute of Informatics. All Rights Reserved.

The relationship between reading ability and other skills that are components of reading ability was analyzed longitudinally in the present study. The children who were participants in the study were attending public school in Osaka, and had participated in a previous 2-year longitudinal study (Takahashi, 1996a). They participated in the study while in the first, third, and fifth grades. At the end of first grade, the speed of naming hiragana words was quicker in those children who had acquired the skill of reading hiragana script before entering school than in children who had learned it in school naming speed strongly determined the children's reading comprehension. As the children became older, the differences among the children in naming speed diminished, and the relationship of speed to comprehension also became weak. When the children were in the fifth grade, the speed of naming kanji words was not related to reading comprehension. That is, the efficiency of processing at the level of encoding did not determine the children's reading ability when they were in the fifth grade. The size of their vocabulary had, on the other hand, a continuously strong influence on reading comprehension, which could be explained by reading comprehension in the third grade. Reading ability and vocabulary had, therefore, a reciprocal relationship: elementary school children increased their vocabulary through reading, and the strength of their vocabulary determined their reading comprehension.

This paper reports 3 experiments about how Japanese preschool children play one of their favorite word games, "shiritori" (cap verses). It also concerned the relationship of shiritori to phonological awareness and reading kana syllabary scripts. Experiment 1 and 2 indicated that two kinds of phonological processing ability are needed to play shiritori without adults' help : (1) isolating the last syllable of a word and (2) retrieving words that begin with a specific syllable. The results do not mean that children have good vocabulary, but rather that they need to recompile their mental lexicon that has a sound index. Experiment 3 showed that there are close interrelationships among shiritori, reading of kana script, and phonological awareness. It also indicated that children who have insufficient phonological processing abilities can participate in shiritori activity with adults' help. It was concluded that children begin to participate in word play activity peripherally, and that helps them to refine their phonological awareness which benefits them to learn to read kana script.