(iStock)

Some students unhappy with the College Board’s decision to toss a section of the June 6 SAT because of a printing error in many test booklets are petitioning to be allowed to retake the SAT for free, saying they don’t believe the College Board’s assurances that the final scores will be valid.

Many test booklets around the United States had a printing error that told students they had 25 minutes to complete either Section 8 or Section 9 — either a math or reading section depending on the version of the college admissions test — when they were supposed to only have 20 minutes. How many students had booklets with the error is unknown but all of the students who took it are affected. About 487,000 people registered for the June 6 SAT.

The student petition says in part:

 Many students feel like their score will be severely impacted by the College Board’s decision. This will hardly be an accurate representation of the scores many students work so hard for. This test plays a major role in college applications, and a scoring system that deviates from the norm is not good enough. The College Board needs to offer an optional retest for students to ensure the students get the scores they deserve.

Asked on Thursday whether the College Board was considering allowing students to retake the SAT, Zach Goldberg, the organization’s director of media relations, wrote in an e-mail: “All students will receive valid and reliable scores within the usual time frame. We are communicating directly with students to address any concerns.”

The College Board, which owns the SAT, said in a note on its Web site that the test “is designed to collect enough information to provide valid and reliable scores even with an additional unscored section.” It also said: “We have deliberately constructed both the Reading and the Math Tests to include three equal sections with roughly the same level of difficulty. If one of the three sections is jeopardized, the correlation among sections is sufficient to be able to deliver reliable scores.”

But students who are petitioning on the Web site change.org  say they don’t trust the final scores. Some other students, parents and school counselors have contacted a nonprofit called FairTest, or the National Center for Fair and Open Testing, to discuss “pursuing a range of alternatives from partial refunds (for partial scores) to free retests to lawsuits, according to Bob Schaeffer, FairTest’s education director. He said in an e-mail: “Most do not believe the College Board’s self-serving, unvalidated claim that results from tests with one missing section will, in fact, be comparable to scores from all other SAT exams.  We are currently evaluating the evidence and options.”

The petition on change.org was started by someone named Sara Knoll and now has a few hundred signatures. Here’s the text of the petition:

After learning about the June 2015 SAT mistake, the College Board attempted to fix their error by throwing out the two sections that were compromised. According to the College Board employees, the number of wrong answers from the two previous sections will be averaged together and account for the number of questions wrong on the student’s third section. For example, if a student got an average of 3 questions wrong total on the two previous sections, the College Board would assume that the student must have gotten 3 questions wrong on the section that is going to be thrown out, even if the student got a perfect score on that section. Many students feel like their score will be severely impacted by the College Board’s decision. This will hardly be an accurate representation of the scores many students work so hard for. This test plays a major role in college applications, and a scoring system that deviates from the norm is not good enough. The College Board needs to offer an optional retest for students to ensure the students get the scores they deserve. This test plays a major role in student’s futures, and the College Board’s “solution” fails to fulfill the purpose of showcasing a student’s abilities accurately.

This is the text of the letter the College Board sent to people who took the June 6 SAT:

Dear Test Taker:

We are writing to let you know that your SAT scores from the June 6th SAT Administration will be delivered within the usual timeframe.

As additional background: on Saturday, June 6, we discovered a printing error in the standard test books that we provided to students taking the SAT that day in the United States.

On behalf of Educational Testing Service (ETS) and our partner the College Board, we apologize for this error.

Following each test administration, we go to great lengths to ensure we are able to deliver reliable test scores. After a comprehensive review and statistical analysis, the College Board and ETS determined that the affected sections will not be scored and we will still be able to provide reliable scores for all students who took the SAT on June 6.

To accommodate the wide range of incidents that can impact a testing experience, the SAT is designed to collect enough information to provide valid and reliable scores even with an additional unscored section.

We take our responsibility to students very seriously, and we regret any confusion you may have experienced. For more information, including Frequently Asked Questions, please visit www.collegeboard.org/june-6-sat.

Sincerely,

Test Administration Services

Larry Krieger, a respected test-prep expert who has been preparing thousands of students to take the SAT and PSAT in a career spanning almost 25 years, said the SAT should release the entire June 6 test so people can see the levels of difficulty of each section so that the public can know whether the they are really comparable in difficulty. He wrote in an e-mail:

The College Board does release the October, January, and May SATs. However, in a normal cycle they do NOT release the June SAT. A released test contains the full test. In addition, it includes a chart with the difficulty level of each question and a scale. SAT questions are ranked on a level of difficulty ranging from 1 to 5 with 1 being the easiest and 5 being the hardest. The number of 1s, 2s, 3s, 4s, and 5s form the “geometry” of a given test.

The College Board has unilaterally transformed the June SAT into the June PSAT. They are saying, “trust us.” But as President Reagan often said, “Trust, but verify.” In order to fully trust the CB, the public needs to see the June SAT and the level of difficulty of each question. Given the unprecedented situation, the CB should take the unprecedented measure of releasing the June SAT to the full public. They can do this by posting the test on their website.

Posting the June exam would restore the CB’s credibility. If their experts are correct then there is nothing to hide.

But Schaeffer said that calling for disclosure of the June SAT might help advance the issue but wasn’t enough. He said in an e-mail that “internal data that goes beyond item text and level of difficulty would be needed to evaluate the College Board/ETS claim that it is valid to drop scores from an entire section of the exam.”

Schaeffer also noted that according to College Board publications about the SAT’s characteristics,  a 20-point range on any subsection of the test is not surprising.

The Standard Error of Measurement for the overall Critical Reading and Math sections of the SAT is +/- ~30 points, meaning that on any day a student’s reported score can vary from their “true score” by as much as 60 points. That is why the report sent to test-takers and their schools includes a score band (e.g a CR 650 is reported as a range of about 620 to 680). Moreover, the test-makes tell score users that differences between test-takers of 60 points or less on any section of the test are not statistically meaningful

Standard Error of the Difference (SED): The SED is a tool for assessing how much two test scores must differ before they indicate ability differences. To be confident that two scores indicate a true difference in ability, the scores must differ by at least the SED times 1.5. For example, SAT verbal and math scores must differ by 60 points (40 × 1.5) in order to indicate true differences of ability.

Bottom line: the SAT is not a very precise measurement tool.