Washington state education officials are seeking a waiver from federal requirements to conduct wide-scale standardized testing of public school students this spring, proposing instead to test a representative sample of about 50,000 students across the state.
In a normal year, about 700,000 students would sit for exams — 14 times as many as is proposed for this year.
As many of most populous school districts in Western Washington respond to Gov. Jay Inslee's order requiring in-person instruction, officials are pitching the model — developed with University of Washington researchers — as a way to preserve teaching time while still collecting broad data about students' knowledge of math, language arts and science during the pandemic. Neither students nor teachers will be held accountable for the scores, which will be shared publicly only at a statewide level. Scores for individual schools or school districts will not be available.
The tests will be administered between April 5 and the end of June. Results will be analyzed in conjunction with several other measurement tools, including student attendance, the results of a state survey on student mental health, and student scores on local assessments such as literacy tests.
The tests aim to provide some answers to a hotly debated and complicated question: How did students fare during the pandemic?
The average scores will be broken up by student demographics and grade level, and the state plans to look at what impact an in-person or remote instructional model may have had on the results.
The proposal has won the favor of some groups that are normally critical of standardized testing, including the statewide teachers union, the Washington Education Association. Earlier this month, the association's president, Larry Delaney, wrote a letter to the U.S. Department of Education expressing his support for the state's waiver request.
"Testing a representative sample of students this year with a shorter test will allow for a much faster administration of the required tests and allow our educators to return more quickly to what matters most for our students: meaningful time for engagement, teaching and learning," Delaney wrote.
Last school year, when the start of the pandemic collided with the spring testing season, the federal government waived requirements to test students in grades three through eight, and high school. This year, the requirement is back, but states can ask the feds for permission to shorten the hourslong tests, give the assessments to fewer students, and not hold school districts to account for the scores. Washington state's waiver request mentions all three adjustments, and several states such as Colorado and Maryland are also planning to scale back their testing programs.
"We think it's a good balance of reducing the impact, and the time spent testing, and yet giving us a statewide picture of what has happened during this unusual school year," said Deb Came, assistant superintendent of assessment and student information for the Office of Superintendent of Public Instruction (OSPI.)
The state is proposing to test fewer grades on each subject than typical: students in third and seventh grades will be tested on language arts, fifth-graders and seventh-graders on math, and eighth-graders on science. Normally, the state administers the language arts and math tests to grades three through 10, and science tests to fifth-, eighth- and 11th-graders. Each test will be about two-thirds of its normal length. School districts that still want to opt into testing all their students in all grades can choose to do so, state officials said.
As always, students or families can choose to opt out of the test, and the state won't require testing for students who aren't attending school in-person. Officials considered offering the tests remotely, but ruled it out, saying the results wouldn't be comparable to previous years if the test-taker and proctor were in different locations.
For each grade level tested, the state will select schools to participate at random, accounting for geographic diversity.
Here's how it works: Let's say the state is selecting schools for the eighth-grade science test. From each of the state's nine Educational Service Districts — regional groupings of school districts — the state will randomly select 50% of districts to participate. From each district, the state will then randomly select a quarter of schools that enroll eighth-graders to participate. The schools selected will then administer the tests to all their eighth-graders.
University of Washington researchers Chun Wang and Elizabeth Sanders, who collaborated with OSPI on the sampling model, said around half of school districts in the state will have a school that is participating.
There are some limitations to what this model can provide compared to previous years, the researchers said.
For one, the smaller the sample size, the larger the margin of error in the results. Drawing conclusions about grade-level scores for very small student groups, such as homeless students, Native American students, Pacific Islander students, and students in foster care is going to be difficult, the researchers said. To provide more rounded data, the state plans to also combine grade levels for these student groups where applicable. The model is also going to be representative only of students who are opting-in and attending school in-person.
One group of students that won't be impacted by these changes, if they are approved: English learners. The state is still requiring all English learners going to school in-person to take an exam this year gauging their English proficiency.
Veronica Gallardo, OSPI's assistant superintendent overseeing the state's Migrant & Bilingual Education Program, said that's because the test provides an important data point about whether students should stay or exit an English learner program.
"It also provides teachers information about the proficiency level, and specially designed instruction that needs to happen," said Gallardo.