Testimony by Cassie Creswell at State Assessment Review Committee Hearing on PARCC | 7.7.16

PARCC is a deeply problematic test. The organizations I am speaking on behalf of today have concerns about (1) the validity and reliability of the PARCC test and the automated algorithms used to score it, (2) the interpretation and use of PARCC scores and (3) the cost of PARCC, especially given the proposed expansion of the PARCC testing program.

Due to the rushed implementation of the Common Core State Standards, the tests created to measure students in alignment with those standards for accountability purposes were also rushed out. We have only seen a single technical report on PARCC. The validity and reliability of PARCC scores, whether there has been an effort to reduce bias in PARCC questions---whether based on race, gender or socioeconomic status, and the role of mode effect (i.e. device differences) on scores are all unresolved issues.

There has been only one (not peer-reviewed) test of PARCC’s predictive validity as a measure of college readiness. It was not longitudinal and did not administer the test to school-aged children, but to college students. The results showed that PARCC was no better than the existing state test in Massachusetts or the SAT in predicting a college student’s grades. We know based on extensive research (and have known for many years) that high school report card grades are a better predictor than standardized test scores of college performance.

By next year, all PARCC test questions including essays and extended response items will be scored by computers, with a small number given a second pass by a human reader. PARCC has been marketed to legislators and the public as a “next generation” assessment—one that can evaluate higher-order cognitive skills in our children (e.g. critical thinking, problem solving, etc) and provide high-quality instructional feedback for teachers and students.

However, there is no evidence that the shallow features used by the statistical algorithms to assign scores to PARCC questions are measuring anything like these skills. Current automated scoring algorithms cannot provide feedback more meaningful than that already available for free from commercially available grammar and spell-checking products.

PARCC score levels were set in order to label 70% of students as non-proficient, a label which has been misinterpreted widely in the media and elsewhere as “failing”.

Setting score levels is a subjective, political decision. PARCC’s proficiency rate was chosen in order to mimic the scores of the National Assessment of Educational Progress (NAEP). And as the National Center for Education Statistics itself says very clearly: “Proficient is not synonymous with grade-level performance.”

The achievement levels of the NAEP have long been controversial; ‘proficient’ on the NAEP is actually a very high level of academic achievement, described as “competency on challenging subject matter”---roughly the equivalent of all As.

It is not reasonable to expect more than half of children to achieve “proficient”; and it is nonsensical to expect more than 90% to perform at that level.

It is even more nonsensical for the state to expect high performance from all students in all schools in the state with the most inequitable funding formula in the nation and where the state contributes the least to local K-12 budgets.

Given the educational spending situation in Illinois, we are deeply troubled that in January ISBE proposed an assessment budget for FY2017 that requests an additional $18 million dollars for more PARCC testing---none of it required by federal statute, including non-summative K-8 testing and two more years of summative testing for high school.

The overall ISBE assessment budget request for FY2017 is an increase of 150% over three years ago, $40 million dollars in a state that has not provided even foundation-level funding to every district over the past five years.

We acknowledge that under federal law, the state must administer annually a standardized test in math and reading to grades 3-8th and one year of high school.

We condemn vehemently the plan to pay for more such testing, in particular PARCC testing for K-2 given that there is no evidence that standardized testing in early childhood is reliable.

We urge this committee to recommend to ISBE that (1) the state end the use of PARCC in high school altogether as a redundant and resource-wasting exercise and (2) for third through eighth grades, the state find a less expensive and shorter alternative to fulfill federal testing requirements.