1977 Volume 2 Issue 4 Pages 137-147
項目反応の三重クロス得点と目標マトリックスとをクロスして得られる反応構造を,得点のパターン別出現率によって典型的なベン図パターンに変換する.つぎに,ベン図について項目間の形成関係を階層化し,反応構造を2項目間の上下位関係に変換する.このデータを用いて,項目関連マトリックスを作成し,さらに形成関係項目連関図を作成する.形成関係項目連関図を項目重要度のレベルによって分割し,目標項目(群)を核とする教授項目群(教授ブロック)に分類する.項目群に小数のルールを適用して,項目の系列化をおこなう.この系列化の方法をいくつかの事例に適用した結果,概念学習や法則学習など,階層性があって相互に連合関係のある教授項目を系列化する方法として有効であることが確かめられた.
A structure of responses obtained from triple cross item responses combined with specicfiation matrices of objectives is transformed into typical patterns of Venn diargam by frequency of score patterns. A hierarchical relationship of the items is determined for the set of Venn diagrams, by which the item response structure is converted into the set of hierarchical orders between two items. On the basis of these data item relationship matrices are constructed, and then item structural contingency diagrams are drawn to indicate the relationship among items. The item structural contingency diagrams are divid ed by level of importance, and they are classified into groups of instructional items or instructional blocks with the nucleus of objective items. The sequencing of items is carried out by applying some rules to the group of items. Application of this method to several cases proved effective as a method of se quencing of instructional items which are mutually associated and have hierarchical orders, such as those in the learning of concepts and rules.