PA Coalition for World Class Math

                             Common Core Assessments to imitate defunct KIRIS?

 Remarks by Richard Innes of the Bluegrass Institute for Public Policy solutions:

Beyond the Common Core State Standards   Beyond the issue of Common Core Standards State Standards looms the equally critical issue of what the new state tests will look like.  

Educational Testing Service recently released some one-page summaries of the test proposals from two separate consortia that are working on this. One is known as “The Smarter Balanced Assessment Consortium” (SBAC) and the other is “The Partnership for the Assessment of Readiness for College and Careers” (PARCC)

 As I read these, I was struck by their similarity to the description of the performance based assessment program that Kentucky launched way back in 1992. That program was known as the Kentucky Instructional Results Information System (KIRIS).  

KIRIS was supposed to be a performance based assessment. It featured complete reliance on open response written questions (multiple choice questions were administered intermittently, but never counted), performance events, and portfolios.  

The old Kentucky performance events in particular sound chillingly similar to the proposals from PARCC and SBAC.  

Kentucky's performance events worked this way. A facilitator was sent out the school to administer the event to a small group of students. A typical fourth grade example involved handing out a standard 8-1/2 by 11 inch piece of paper covered with many approximately life-size images of the lady bug insect. Students also received rulers, protractors, etc. Student groups were to determine how many images were on the paper. Students then wrote up individual reports on how they had solved the problem. Those reports were collected by the facilitator for grading at a central site.  

The question creators believed students would come up with an estimation process such as cutting the paper into fourths, counting the images in one of the smaller areas, and then multiplying by four to estimate the overall total.  

Of course, by fourth grade it should have been a trivial task for the students to simply count up all the images on the paper, but that apparently never occurred to the KIRIS performance event creation team.  

In any event, the important point is that these sorts of events proved impossible to manage in Kentucky’s high stakes state assessment program. The questions were so vivid and easily remembered that they had to be changed out every year to avoid test gaming. That created new and very difficult challenges for linking and equating events of equal difficulty. Those issues were never solved.  

Ultimately, the performance events collapsed in 1996 when the middle school performance event generated totally unusable results. The fallout led to the legislature completely removing performance events from the testing program along with a concerted, but unsuccessful, effort to obtain damages from the testing company that created the performance events.  

Eventually, the rest of KIRIS, and yet another follow-on Kentucky test effort, known as the Commonwealth Accountability Testing System (CATS), both collapsed. CATS attempted to maintain the same, heavy emphasis on open response written answer questions (always expensive, time-consuming and subjective to grade, and always resulting in tests that could not sample individual student knowledge accurately due to a matrixed test design required to keep testing time reasonable) and writing portfolios (always inflated in grading as teachers did the scoring).  

Flash forward to today.  

What in the current proposals looks much different from what has already been tried, and failed, sometimes twice, in Kentucky?  

Will the new consortia come up with a way to create different performance events from year to year that can be linked and equated with high accuracy? This is absolutely essential for a valid and reliable longitudinal assessment. Can the process support changing questions frequently to preclude cheating?  

Will the consortia come up with a way, never found in Kentucky, to insure enough questions are on the test to create valid and reliable scores for individual students (never achieved with either KIRIS or CATS) if those tests include time-consuming to administer (and grade) open response questions?  

How will the open response questions and performance events be graded? If by outside scorers, can states afford to hire scorers with adequate subject knowledge and grading skills (problems not solved with KIRIS or CATS)? If teachers do the scoring, how will inflation be avoided if results are to be used for accountability (problem never solved with Kentucky's writing portfolios, which were teacher scored and always found inflated in multiple scoring audits)?  

With many states ultimately gambling a huge investment on the new Common Core based testing program, these issues need to be resolved in a careful, thoughtful manner.

 

Richard Innes

Education Analyst

Bluegrass Institute for Public Policy Solutions

400 E. Main Street Bowling Green, KY 42101