and

Statistical
Study of the Denton ISD Math Tutoring Projects

Fall
2005, 2006, 2007 and 2008

Many programs, such as Title 1, require that federal funds are only used where a statistical study has been published supporting the value and expected results of the expenditure being requested. The study should be conducted by an independent organization such as a university or college. The statistical study of a math tutoring project by the University of North Texas (UNT) fulfills this criteria. The target group was at the Denton, Texas ISD. The 2008 study has been completed, and results are coming soon.

The following reflects the outcomes as published by University of North Texas.

Three hundred eighteen (318) children from grades three to eight were identified for tutoring, based on Denton ISD math baseline testing. Each student was diagnosed for competence in numbers and operation using GoToLearn’s diagnostic instrument. Of the 318 in the identified group, 152 were randomly selected for tutoring, and this group was tutored for six weeks, sometimes once and sometimes twice weekly.

Tutoring was structured individually by providing interns with the results of the diagnostic testing. Interns were introduced to the district Math Tutoring materials (TIPs), the TEKS, methods for tutoring young children, and manipulatives for teaching mathematics.

Each week, each intern communicated with both mentor teachers in person and the university math instructor online to discuss the progress of the tutoring sessions. Following the tutoring, all students, both tutored and non-tutored, were post-tested. The test results indicate that, at the post-test level, the tutored students differed significantly from the non-tutored students at the p<.05 level.

Three hundred ninety three (393) children from grades three to eight were identified for tutoring in 2006. Of the identified group, 243 were randomly selected for tutoring.

This study shows two samples (pre and post test) of the same students which determines the t-test be dependent and paired. The data below represents the output of the t-test performed by the University of North Texas.

In the first table below, the paired samples statistics reflects a positive shift of the Mean test scores from 46.24 to 57.31.

The Paired Samples Correlations table below verifies a correlation exists of students who did well on the first test also did well on the post test.

The test results in the table below, while using GoToLearn software, demonstrates significant improvement in test scores.

The GoToLearn software diagnostic served several purposes. First, it showed the interns how to consider needs of the learner as related to the TEKS, as each question is linked to a TEKS. Second, the software’s ease of use and reporting allowed the interns to compile information in an electronic form for reporting purposes.

Finally, by using the same items for both the pre- and post-tests, interns were able to assess the effect of their tutoring on student learning

This document is published by GoToLearn and taken from formal UNT-issued reports.