Editor’s note: This commentary is by William J. Mathis, who is the managing director of the National Educational Policy Center, a member of the Vermont State Board of Education and a former superintendent of schools. The views expressed are his own.

Measuring the effects of education is like the apocryphal group of blind people describing an elephant based on the part they feel. People with assorted predispositions touch different parts of the elephant and shout how the truth of their dearest theory is now confirmed. Alas, for the educational elephant, the proclamations are almost invariably of the “The scores are too low and the costs too high” genre.

The past weeks have provided the oracles of numerology with two totems to brandish: the release of the SAT and the SBAC testing consortium results. The SAT is one of the two major college entrance examinations while the Smarter Balanced Assessment Consortium (SBAC) is one of the tests adopted by a group of states to meet federal requirements. Despite a strong public backlash against too much testing, such “college and career readiness” tests are bally-hooed by the federal government as essential.

Most prominent among the critics are media lamentations about the low student “proficiency” rates on the SBACs and the “10-year low” on the SAT tests. Quickly recognizing a media opportunity, pro-privatization think tanks also seized the initiative. Thirteen years into the No Child Left Behind era, Michael Petrilli of the right-leaning Fordham Institute was quick to double-down on this demonstrably failed strategy, calling for more standards-based reform initiatives in high schools.

But, let’s take a closer look.

It is true that the SAT scores are at a 10-year low. They re-normed the test 10 years ago. (Vermont scores remain basically even with last year.) However, what is missing from the story is that scores were consistently increasing from 1980 to 2000. They are going to re-norm it again next year so, like the state tests, long-term trend comparisons will be impossible. Yet, this is not the important point.

The SAT is not a required test, it is voluntary and the participation rate varies. Nationally, there are more test-takers than there were before. Also, the composition of test-takers has changed. To encourage economically stressed children and children of color to go on to higher education, many states are paying for all students to take the tests. As more of these students with fewer educational opportunities take the tests, the overall average comes down – even as the average scores of students in that subgroup go up. The irony is that as more students of color graduate and aspire to higher education, the overall effect of these gains is to push SAT scores down.

The SBAC results have also just been released. These tests were built to measure the controversial Common Core. For each of the grade levels (3-8, 11), a “proficiency” cut-score was established, and each student was scored as proficient or non-proficient depending on whether they were above or below the cut-score. For Vermont, the percent of students who scored proficient in math ranged from 37 percent to 52 percent depending on grade level. That’s not very high. For English Language Arts, the proficiency percentages were in the 50s. What this means is that according to the cut-off criteria, almost half of the students were declared to be not “college and career ready.” (Most of the 17 other states that administered the SBAC scored lower than Vermont.)

Are these scary numbers accurate? We really don’t know. The cut-scores have not been validated. The scores of those who were successful in colleges and careers have not been compared with those who were not. To be sure, the tests were subjected to a variety of rigorous examinations by “experts.” Unfortunately, this is just a collection of opinions. Setting the cut-scores was a political decision unconnected to the readiness claims. The effect is that commercial textbook and testing corporations who have a huge financial interest in scary numbers publish a lot of scary numbers.

The National Assessment of Educational Progress (NAEP), which is given to a random sample of students, gives us a steadier long-term benchmark. It tells us that for reading and mathematics, the nation has improved at a relatively consistent rate over the past 30 years. The good news is that lower performing sub-groups have increased at the same pace as higher-performing ones. The bad news is that, they are not catching up. As a nation we continue to systematically underprovide for our neediest children. While we don’t know if the tests measure “college and career ready,” we do know they measure poverty quite well. One in three children in the U.S. lives in poverty and the needy are increasingly clustered together. And as the income gap increases, the achievement gap increases.

Sadly, test-based reforms provide us with the illusion and the excuse that giving tests, publishing scary numbers and talking-tough about “accountability” will close achievement gaps and educational opportunity gaps. They won’t and that’s the elephant in the room.