[慶應義塾大学との産学共同研究]認知診断モデルによる学力分析で学習量と習得率に相関性あり:[Joint Research with Keio University] Correlation between study time and learning percentage using a Cognitive Diagnostic Model to evaluate academic skill

2016. 9. 5

株式会社リクルートマーケティングパートナーズ(本社:東京都中央区、代表取締役社長:山口 文洋)が運営する、リクルート次世代教育研究院(院長:小宮山利恵子)は、「認知診断モデルによるスタディサプリ学習者の学力分析」を慶應義塾大学 星野崇宏教授との共同研究の進捗をご報告致します。

Progress report of the joint research program “Academic skill analysis of Study Supuri users using a Cognitive Diagnostics Model” between Recruit Institute of Next Generation Eduction, operated by Recruit Marketing Partners, and Associate Professor Takahiro HOSHINO, Keio University Graduate School of Economics.

Validating Learning in Mathematics

行動経済学や心理学、統計学において先進的な研究を行っている慶應義塾大学 星野崇宏教授と共に、同氏の専門領域である「認知診断モデル」(※1)を活用した学習者の学力分析の研究を4月より開始。この度、高校1年生「数学」において認知診断モデルに基づくコンポーネント(※2)毎のオンライン講義動画サービス『スタディサプリ』と学習習得率について検証を行いました。

Research into analysis of learner academic skill using a Cognitive Diagnostic Model (CDM) with Professor Takahiro HOSHINO of Keio University began in April of 2016. Professor HOSHINO is a leading researcher in the fields of Behavioral Economics, Psychology, and Statistics who specializes in CDM (*1). We have conducted a validation of learning effectiveness of the online lecture video service “Study Sapuri” in 10th grade mathematics on a CDM per component level(*2).




We have found the per component learning effectiveness is positively correlated to amount of time using the Study Sapuri application (noted as “using app” in the graph), with higher academic skill scaling with time using the app. The rate of increase in academic skill is highest in two components; “Basic Calculation”, and “Behavior of Quadratic Equations”. We are finding learning effectiveness is pronounced with the fundamental skills.


Future research will look at visualizing the effectiveness of each grade level and each lecture unit, and increasing the precision of components. This data will be used to improve the service and content effectiveness and quality, better understand learner behavior, and provide feedback to learners.


※1 認知診断モデル:

*1 Cognitive Diagnostic Model (CDM):
A Quantitative Psychology Model based on Cognitive Psychology, Neuroscience, and Learning Psychology which analyses test response and behavioral data to approximate the hidden internal “learner understanding, knowledge, and cognition” state to provide guidance to more effective learning methods.


*2 Component:
Significant Learning Elements that may be approximated using the CDM. A cognitive element, which represents a more effective evidence for intervention in comparison to learner competency of a teaching unit.


■本研究内容:Research Content
・対象(Target):スタディサプリ「到達度テスト」(Study Sapuri “Achievement Test”)
・学習時間(Study Time):スタディサプリ視聴時間+単元確認テスト受講時間(Study Sapuri + teaching unit verification test time)
・対象数(Learner Pool Size):7758
・解析データ期間(Data Collection Period):2015年4月9日~10月15日(April 9th, 2015 to October 15th, 2015)

■慶應義塾大学 星野崇宏教授との共同研究概要:Abstract of Joint Research with Prof. Takahiro HOSHINO of Keio University Graduate School of Economics
Development of a efficient education methods which adapts to learner strengths and weaknesses utilizing Behavioral Economic Principles)
・開始時期(Research Initiated):2016年4月1日(April 1st, 2016)