Title of Study:  An Experimental Study of the Effects of Cognitive Tutor Algebra 1 on Student Knowledge and Attitude

Author(s): Pat Morgan and Steven Ritter

Summary:

This well designed experimental study compared both student content knowledge and student attitudes toward math after a full year of instruction with either Cognitive Tutor or traditional instruction in Algebra 1.

Sponsoring entity(s): Carnegie Learning, Inc.

Date conducted: 2002

Location of the study: This study can be accessed through the Carnegie Learning website.

Setting(s) addressed:

• Classroom

Targeted population(s):

• Junior high school (9th grade)

Primary sources of evidence used in the study or report:

• ETS Algebra I End-of-Course Test
• Survey of student attitudes toward math

Primary Technology Application(s) Addressed:

• Instructional software

Major education topic(s) addressed:

• Math

Major findings/conclusions:

• In all participating schools, students using Cognitive Tutor had higher end-of-course scores than students in traditional classes taught by the same teachers.  This held despite school-to-school variation.
  
  
• Course grades were also higher for students using Cognitive Tutor compared to traditional classes taught by the same teachers.
  
  
• There was significant test score variation among teachers, and variation in the relative effectiveness of traditional instruction vs. instruction with Cognitive Tutor with different teachers.
  
  
• There was a slightly significant interaction between gender and improvement with Cognitive Tutor, with boys having more improvement over traditional curriculum than girls, although the gender effect was not seen for course grades.
  
  
• No interaction was seen between ethnic background and improvement with Cognitive Tutor.
  
  
• Students using Cognitive Tutor were more confident of their ability in math than those in traditional classes.
  
  
• Students’ attitudes toward math are more strongly influenced by their teacher than the curriculum used.

Practical implications of the findings:

• Interestingly, the largest effect sizes comparing Cognitive Tutor to traditional instruction were for the teachers with the lowest scoring students in their traditional classes.  This suggests that the instructional software might have its greatest impact for students with teachers who struggle to engage them or to meet their learning needs in a traditional classroom.

Reviewer Comments:

This rigorously designed study used a randomized experiment design with same-teacher controls and measures of both students’ content knowledge and attitudes.  Data analysis accounted for student mobility, gender and ethnicity.