Evaluating RTI’s Effectiveness Over the Long Term

by Kimberly Gibbons

No matter how you formally define response to intervention, most definitions contain common components: scientific, research-based instruction; the use of learning rate and level as the basis for determining effectiveness of intervention; and decisions about intensity and duration of interventions based on a student’s response to interventions across multiple tiers of service.

The St. Croix River Education District, which manages special education services for five school districts in east-central Minnesota, has been involved in implementing these components for the past 20 years, though no one knew it as Response to Intervention back then.

Our district has a long history of using data-based decision making through a problem-solving model. For the past 11 years, we’ve worked with member districts to implement an RTI model that coordinates three critical elements: (a) frequent and continuous measurement using general outcome measures, (b) research-based instruction within a tiered service delivery model, and (c) schoolwide organization principles to ensure the most effective instruction for each student.

In addition, our district uses a student’s response to intervention as part of the evaluation process for special education services in the category of specific learning disability.

Our Model
Measurement: Schools follow a protocol in which student progress is measured on three schedules: benchmark (three times per year) for all students in early childhood through 8th grade; strategic (monthly/biweekly) for students with some concern; and intensive (weekly) for students with the most intensive needs.

All districts use general outcome measures of reading (oral reading fluency), early literacy measures (letter naming and sound fluency, nonsense word fluency and phonemic segmentation and blending tasks), and mathematics (math fact fluency and math applications). These measures are sensitive to growth over time and have predictive validity with high-stakes tests. Formative data from these measures help determine the effectiveness of instruction and intervention, and these data are used in the entitlement process. If instruction is not leading to student progress, then the interventions must be changed.

Scientific, research-based instruction within a multi-tiered model: St. Croix River has worked with its member districts to ensure that basic skill curricula are research-based and well-aligned with state standards.

Three of our member districts, the East Central, Hinckley-Finlayson and Rush City school districts used “A Consumer’s Guide to Evaluating a Core Reading Program” in making curriculum adoption decisions. These districts selected the Reading Mastery curriculum because of the extensive empirical research documenting achievement gains for students. In addition, we have established a tiered service delivery model that incorporates research-based interventions at each tier.

Schoolwide organization: Five elements of school organization are promoted to ensure that effective instruction can be provided to every student.

First, student progress is measured on three schedules as described above. Teams of grade-level teachers meet regularly to examine the achievement data of students in their grade levels. Flexible grouping by achievement levels is used to promote student learning at the appropriate instructional level. Principals schedule reading groups at the same time within grade levels and at different times across grade levels to concentrate additional resources (Title 1, special education) at each grade level. Finally, each building has a problem-solving team that is trained on a systematic five-step model for designing individualized interventions.

Outcome Evidence
The St. Croix River districts have been collecting general outcome measurement data (AIMSweb or DIBELS) in basic skill areas since 1996. In reading, student performance is evaluated three times per year on simple, one-minute timed reading passages. The percentage of students reaching benchmark target scores has increased significantly over the past decade from 35 percent to 70 percent. In the Chisago Lakes School District 2144 in Lindstrom, Minn., the percentage of 2nd grade students reaching benchmark target scores has increased from 38 percent to 82 percent over the past decade.

Over the same period, our district has been tracking the performance of students performing at the 10th percentile. The results indicate that at every grade level, student growth rates in reading have at least doubled and in some instances tripled. For example, the median score of 1st graders at the 10th percentile was 15 words correctly in one minute in 1996 but had risen to 39 words in 2007.

In addition, the percentage of students reaching the grade-level standard on the statewide assessment increased from 51 percent at the model’s inception to 80 percent in 2005. This is a slightly faster increase than that of the state overall. Finally, the percentage of students identified as learning disabled has dropped dramatically over the past decade, by 50 percent.

We believe these data trends provide strong evidence of the preventive nature of the RTI framework. Moreover, with the implementation of a multitiered service delivery model, teachers realize they are able to get effective interventions in place for students without having to request an evaluation for special education services.

Some concern has been raised in the education community that the use of RTI models will cause a rapid increase in the rate of students being identified with specific learning disabilities. However, because of the increase in student achievement over the years, we believe we have prevented many disability classifications.

Kimberly Gibbons is executive director of the St. Croix River Education District in Rush City, Minn. E-mail: