|Scrubbing Bubbles: With Colleges Moving Away from Standardized Testing, Will K-12 Follow?|
|News/Features - Feature Stories|
|Wednesday, 14 November 2007 02:29|
When Augustana College announced earlier this month that it was eliminating standardized tests as an application requirement for prospective students, it highlighted a trend in higher education.
"There has been an accelerating surge of selective colleges dropping their admissions-testing requirement," said Robert Schaeffer, public education director for the National Center for Fair & Open Testing, also known as Fair Test. "The pace is such that some folks have suggested we're reaching a critical mass or a tipping point."
Fair Test (http://www.fairtest.org) has counted 762 colleges and universities that are to some extent test-optional. Some schools don't require a standardized-test score for students who meet certain grade or class-rank conditions. Others, such as Augustana, have stopped requiring test scores altogether.
Fair Test has identified 38 colleges and universities that have gone test-optional in the past three years, including Augustana and Galesburg's Knox College.
That might not sound like much in the context of more than 700 institutions, but the larger list includes specialized schools (such as the Berklee College of Music and the Art Institute of Chicago), technical colleges, institutions that have never required test scores, and a lot of caveats.
What's happening now is that a lot of the most competitive liberal-arts colleges are dropping their testing requirements, Schaeffer said.
But kids with their eyes on college shouldn't celebrate the demise of multiple-choice assessment tests yet. The ACT and SAT (both of which evaluate math and language skills; the ACT also has a science section) are deeply entrenched, and nearly 90 percent of selective four-year colleges still require standardized tests for applicants, according to the College Board, which develops and administers the SAT. ("Selective" colleges don't automatically accept all applicants who meet core requirements.)
And smaller institutions such as Augustana are far more likely to drop standardized tests than larger schools.
The movement away from requiring standardized tests stands in contrast to federal policy for K-12 education, specifically the No Child Left Behind act. The legislation, which was signed in 2002, is up for reauthorization, although congressional leaders are skeptical that will happen this year. The law remains in effect even without House and Senate action, and reauthorization/reform proposals are likely to be legislative priorities next year against the backdrop of a presidential campaign.
No Child Left Behind ties federal education funding to state-developed assessments of students at certain grade levels in language, math, and science. Most states have opted to use standardized tests for those evaluations. Schools that don't meet school-improvement standards are subject to sanctions, up to reorganization.
Fair Test's Schaeffer said the divergence between No Child Left Behind and college admissions is telling. "They reflect important differences in how the K-12 and higher-education policymaking worlds are structured in this country at this time," he said. While college-admissions decisions are made by educators, "K-12 testing policy is made by politicians, most of whom haven't been near a classroom since they graduated from college, and are motivated often by the search for a solution that appears quick and simple. Tests offer that false promise."
The question raised by all this is straightforward, even if the answers are complicated: What is the value of standardized testing in education?
Cheap and Easy
Augustana's decision to go test-optional was not a response to a perceived revolt against standardized testing but a choice based on the institution's experience, values, and goals.
"One of the reasons we became test optional was not anger at the standardized-test industry," said W. Kent Barnds, the Rock Island college's vice president of admissions and enrollment, in an interview last week. "We still anticipate that the vast majority of our students will submit standardized-test performance and will expect that that is included in our decision-making process."
The move - approved by a vote of the Augustana faculty - had its roots in one key reality: Standardized tests are simply not very useful for the school. After three decades of institutional research, Barnds said, "it's clear that the best predictor of success for students who attend Augustana is their four-year record of achievement in high school."
Augustana creates a predicted grade point average for its incoming students, and basing that on their high-school records yields an accuracy of roughly 60 percent, Barnds said. "When we added in the results of standardized-test performance, it made that predictor only about half a percent stronger."
In other words, pre-college standardized tests do help refine the college's ability to choose students who will succeed there, but not by much.
That's not just true at Augustana, and even the creators of standardized tests concede the point.
"The best predictor [of first-year grades in college] is high-school grades," said Ed Colby, a spokesperson for Iowa City-based ACT. Among individual factors that predict college success, he said, the ACT is second. And "the best predictor is using ACT scores and high-school grades together, which is one of the reasons that we strongly urge colleges to use ACT as just one factor in the admission process."
St. Ambrose University requires prospective freshman to submit an ACT score, said James P. Loftus, the school's vice president of enrollment management and student services. "What we're looking for [from the ACT] is a certain level of aptitude," he said. While "grades are certainly a greater predictor" of success at St. Ambrose, he added, the ACT scores are still valuable.
That thinking is supported by an annual survey of admissions counselors conducted by the National Association for College Admission Counseling. In its 2006 survey, 76 percent of colleges attached "considerable importance" to grades in college-prep courses, while 60 percent attached "considerable importance" to SAT or ACT scores.
The fact that 60 percent of colleges do value standardized tests - and that the percentage is up from less than 50 percent prior to 1997 - suggests that the makers of the SAT and the ACT don't have much to fret about.
"Most of the colleges that have gone to the type of admissions that don't require an ACT or SAT score are those that are relatively small and have a little more luxury to examine each individual student applicant thoroughly," Colby said. "The larger institutions are not likely to have that kind of a luxury with the size of their application pools. That's where the standardized test scores really help them to make determinations about each student."
Barnds concurred that a test-optional admissions policy requires work. "We've added the features of a selective admissions program that allow us to be a test-optional school," he said. "Letters of recommendation from a guidance counselor and from a teacher; a writing sample required for all of our applicants; encouraging students to participate in the admissions interview, where we have the opportunity to get to know a student personally and find out what he or she is passionate about."
Furthermore, students that choose not to submit standardized test scores are required to participate in an admissions interview and submit a graded paper from their junior or senior year of high school.
"There are lots of places that must use some standard ... to cut the chaff from the wheat," Barnds said.
"Test optional is not a one-size-fits-all policy," Schaeffer said. "It does require a commitment by the institution's leadership of sufficient resources to carefully look at each applicant. The most resistance to test-optional comes from large, public universities that have severely understaffed admissions offices, in which they use test scores as a cheap, easy initial sorting tool."
He added that standardized tests provide useful demographic and income information that can help colleges "build a class. ... It gives them another piece of data they can use."
And perhaps most importantly: "It's data they get for free," Schaeffer said.
The number of ACT tests taken grew 9 percent last year compared to the 2005-6 school year, Colby said, and the number of individuals taking the test jumped 8 percent.
He called the value of the ACT an "apples-to-apples comparison of student achievement" and said that grades are imperfect measures, as well: "We have documented grade inflation. We're seeing higher and higher grades. So grades are kind of a tenuous measurement as well." Colleges can use standardized tests "knowing that students took the same exam measured on the same scale."
But being standard doesn't mean it's meaningful, Schaeffer argued. The subjectivity of standardized tests is in the design - what's included, what's emphasized, how the questions are phrased. "The objectivity is a bit mythical," he said. The standardization "doesn't make a subjective process objective just because it goes through a machine."
A larger problem with standardized tests, however, is that they're standardized tests. By trying to measure student knowledge or skills acquired over years in the space of a few hours, there are going to be problems.
"No three-hour test on a Saturday morning captures more than a tiny sliver of what it takes to succeed in college," said Fair Test's Schaeffer.
Barnds said eliminating that issue was one component of Augustana's test-optional decision, "to be more inviting and accessible to those candidates who have performed strongly in high school but have experienced modest success on standardized tests.
"We think there are a lot of students ... who have self-selected not to apply to a place like Augustana because we are a selective place, and we enroll students with higher test scores and higher class performance," he said. "This focus on the test score has prevented some students from applying."
Poor test performance also has demographic components - what you'll often hear called "racial bias" - that make standardized tests less attractive to institutions committed to improving their diversity.
As Barnds said: "We also want to be more inviting to those students who have historically performed not quite as well on standardized tests. First-generation college-bound students, those students for whom English is not their first language, ... [and] many students from African-American, Latina/Latino, and Asian-American communities do not perform, at least consistently, as well on standardized tests."
"Eliminating test-score requirements for admissions and scholarships enhances both equity and excellence," Schaeffer said.
Loftus said that St. Ambrose is sensitive to questions about racial bias in standardized tests, and noted that the school corrects for that problem by carefully considering each application.
Colby didn't dispute that there are racial differences in standardized-test performance. "We've been working very hard for a long time to eliminate bias in our test questions, to ensure that every student has an equal opportunity to do well on the ACT based on their knowledge and skills," he said. Still, "we're still seeing significant differences in terms of racial and ethnic background. Students in some racial/ethnic groups don't perform as well as students in others on the whole."
"My guess is that they've done what they can" to eliminate test bias, Schaeffer said.
In an e-mail, Colby added that racial differences can be explained by looking at other factors: "In terms of the impact of race on test scores, our research indicates that when coursework taken, grades earned, and high school attended are taken into account, race/ethnicity is associated with only 1 percent of the variability in ACT scores.
"We notice two factors that cut across all racial/ethnic lines in terms of ACT scores: level of academic preparation and family income. Those students who have taken ACT's recommended core curriculum (four years of English, three or more years each of math, science, and social studies) achieve higher average scores than those students who haven't, no matter what their gender or race. In addition, the higher the average family income, the higher the average ACT score, regardless of students' gender or race."
That might be true, but it points out a critical flaw in standardized tests. Yes, they provide a uniform yardstick, but that measure doesn't and can't account for differences among schools and in test-taking ability.
"They've made changes, but they do not alter the fundamental problems with the test," Schaeffer said.
And with an emphasis on diversity in higher education and an increasing proportion of nontraditional students, he said, "the tests become ever-weaker predictors, because they can't adjust for those factors. They work best for kids from mainstream cultures who are going directly to college from high school." Of course, that's still a vast majority of students applying to selective colleges.
Augustana's strategic plan includes increasing its overall enrollment and the proportion of American people of color. The college's enrollment was 2,200 students when Barnds arrived in 2005, and reached its fall 2009 goal of 2,500 students two years early.
Furthermore, Augustana has a long-term goal of having 15 percent of its student body being people of color. The percentage was 10 percent for this year's freshman class, and less than 9 percent the year before.
Barnds emphasized that Augustana is not becoming less selective. It had its two largest applicant pools the past two admission seasons, he said. "We're coming off the strongest year we've ever had," he added. "We're at full enrollment. This is really a philosophical decision for the institution."
He also noted that the college measures the success of its student-selection process by monitoring the percentage of students who return to the school for a second year. The retention rate for the entering class of 2006 was 87 percent, he said, and "that is one of the highest retention rates we've had" in the past 30 years. "We are selecting and enrolling those students who are a very good fit for Augustana."
Standards of Excellence
It's not logical to say that because more colleges and universities are dropping standardized-test requirements, therefore the federal government should eliminate mandatory testing to measure student achievement at the grade-, middle-, and high-school levels.
But one critical point by people on both sides of the standardized-test debate is that while these tests can have a value, it's primarily in concert with other measurement tools.
"The ACT score should be used as only one of a number of different factors they use" in college admissions, ACT's Colby said.
But that's not how No Child Left Behind has been implemented in the states that have opted for standardized tests.
Schaeffer said that the law's assessment requirements "have the same core problems [as other standardized tests], and more so."
He noted that the ACT and SAT aren't high-stakes tests - especially as their importance is diminished in higher-education admissions. Students can take them multiple times. And they're considered against the backdrop of grades, curriculum, activities, and other factors.
"One kid getting one more item wrong on the test [under No Child Left Behind] is the difference between that school making adequate yearly progress and not," he said, "and if it's not, it faces sanctions, which ultimately can lead up to school restructuring - the staff being fired. The stakes are very, very different."
In summary, he said, "A mediocre test used poorly is worse than a bad test that doesn't have money stakes attached to it. No Child Left Behind uses tests - be they good or bad - in the worst possible manner."
Even though No Child Left Behind calls for assessment of "higher-order thinking skills and understanding," a paper co-written by Fair Test Co-Director Monty Neill (http://fairtest.org/nattest/RefocusingAccountability.html) claimed that "most assessment tools used for federal reporting focus on lower-level skill that can be measured on standardized, mostly multiple-choice tests. High stakes attached to them have led schools to not engage in more challenging and engaging curriculum but to limit school experiences to those that focus on test preparation."
As Barnds said, referring to the National Association for College Admission Counseling survey that found that only 6 percent of colleges find state exams of "considerable importance": "State examinations are virtually meaningless to colleges and universities in the admission process."
No Child Left Behind as it is currently implemented could also lead to problems similar to those that have cropped up around the SAT and ACT, with well-to-do families able to help their children score better through preparation courses.
"I wonder sometimes if K through 12 is becoming part of the test-prep industry by teaching to the tests, and by focusing so much on teaching to the test that they're not individually developing students in the way that is necessary at that level," Barnds said. "As long as there are techniques that can be applied to do better, and more importantly, as long as one can buy those techniques, we're going to have problems with this."
Most people agree that it's critical to assess schools and students. "The goal of No Child Left Behind is laudable," Schaeffer said. "The mechanism of No Child Left Behind is counter-productive."
But what could replace No Child Left Behind as a national policy for assessing student achievement?
"There is no consensus on how to change it because there's so much criticism," Schaeffer said.
But he pointed to Nebraska's assessments. As described in the above-referenced paper: "Nebraska utilizes a system of assessments created and scored by local educators. These systems are peer-reviewed in a system supported by assessment experts and include a check on the validity of such assessments through the use of a statewide writing examination and the administration of one norm-referenced test."
And that sounds an awful lot like what's happening in higher education. Institutions that can are making a commitment to a labor-intensive process that goes beyond number-two pencils, "where you're looking not just at how well kids fill in bubbles on a test one day, but at the work they do over time," Schaeffer said. "Excellence isn't just one thing. ... It's just not the same absolute criteria across the board."
Tags See All Tags