Texas Holds Students With Disabilities to a Higher Standard on State Exams

Let’s just call it like it is. 2016-2017 was not a great year for student assessment in Texas. Between the scandal of mis-graded STAAR tests, graders hired for peanuts on Craigslist, and large numbers of parents and students conscientiously objecting to even taking part in our state’s annual testing rituals, many Texans have been left wondering where, exactly, IS the accountability in our state accountability system.   

But if there is one person who is keenly aware of where the accountability is, it’s the high school senior who’s taking a STAAR End of Course test. That senior needs to pass that test in order to graduate in that same month. That student is painfully aware that the STAAR test doesn’t care where he grew up, or who her parents are, or whether he is dyslexic or learning disabled or gifted and talented. Regardless of what challenges she faced or what advantages he received, everyone all sits down and takes the exact same test. That’s what TEA promises. All students take the exact same test. All students meet the same standards. Because that is what is fair. 

This is a key feature of a standardized accountability system after all. Everyone receives the same chance. Of course, it is an undeniable reality that we do not all really receive the same chance. Some of us attend private schools. Some of us attend urban Title 1 schools. Some of us are dyslexic and words slide around on the page. And so some of us receive our tests with some types of approved accommodation. 

Here’s a slide from a TEA presentation about STAAR test accommodations, and how these accommodations do NOT modify the content of the exams.

The STAAR online allows students access to the test on a computer rather than on paper. The online exam can scale up or down the number and type of accommodations a student receives. There may be a text-to-speech function for dyslexic students or students with visual impairments. For students who qualify as Limited English Proficient (LEP) the test may provide clarification on non-content specific vocabulary.

So although the accommodations the test provides may be different depending on which mode a student is using to take the test, the content of the test remains the same. 

After all, if you can answer a textual analysis question on a reading passage, then does it matter whether the passage was on a window  or on paper? If students had the same reading selections, the same test questions, and the same level of difficulty? In a standardized test system you don’t give students an unfair advantage by changing the way they access a test. 

Except wait.

Pictured below are screen shots from TEA showing the conversion tables from raw score to scale score on the English II paper administration - and the online administration.

In direct contradiction to TEA's statements, these two tests- despite being exactly the same in content - have different passing standards.

For a student to pass the paper version of the STAAR English II EOC a raw score of 41 was needed. To pass that exact same test online? A 42. 

A difference of one point. One wrong answer on a multiple choice question. Something so small and insignificant probably no one would have noticed. 

No one that is except one senior. One senior in a North Texas high school who got a raw score of 41 on the online administration. And he failed. Many of his cohorts who were lucky enough to NOT need the accommodations ONLY offered by the online version of the test passed with that same 41 he received.

 He noticed. And we’re talking about it.

Now TEA doesn’t admit, that we’ve seen, to “bell curving” these tests. Not that there would be anything inherently shady in adjusting scale scores on an exam every year to better fit those scores into a normal distribution. This is common standardized testing practice, and is, in fact, a useful tool to counteract tests with subjective grading such as an English test. Think of it this way - if one year all of the graders TEA uses are just unusually harsh, re-norming the scale scores will counterbalance their low marks. There’s nothing intrinsically unfair about norming tests.

But using different normalizing scales on a version of the test where students read the material on the computer rather than read it on paper? 

If thats what TEA did here … and, let’s be clear, we believe that it is. 

Look at the third column in the conversion table - the percentile. Notice that the raw score of 41 conforms to a percentile of 38 for the paper-based, and 42 to 38 for the online. All the students whose score was below the 38th percentile state-wide failed the test - for paper-based that was students who scored below 41. For online it was below 42. This is a dead giveaway for norming.

So if that is indeed what TEA did, let’s consider the implication behind choosing to fit online scale scores into a different normal curve than a paper-based test. The students who take the online version of the STAAR exam are those students who meet the state’s stringent requirements for receiving assessment accommodations. These are students who, most usually, meet medical requirements for receiving accommodations. Or they are students who have been speaking English usually for less than three years. Historically speaking scores on the online version of the STAAR exam have averaged lower than the scores of the general population who takes the unaccommodated paper based version.

So why would TEA choose to normalize the online version of the STAAR test to itself rather than include it with the paper based test?

There’s one logical prediction. The online test should skew lower since the average raw score most likely will be lower. Students should have an “easier” time passing when taking the online test - easier in that the raw score required to pass should be lower.

That’s not what happened here. Instead, these results imply, our most at-risk student population actually out performed our general population on the STAAR test. And at least one of them failed because of it.

TEA, we need answers. We need clarity. No one we’ve spoken to has been able to provide us any comprehensible response as to what has happened here. We’ve been told the test was “easier”. But you said the test was not easier; that it covered the exact same content. We’ve been told this is just responding to the “mode” of the test. Are you implying that students who are receiving medically necessary accommodations - accommodations such as reading less than 50% of the question and answer choices aloud, and NONE of the reading passages - confers an educational ADVANTAGE on the students taking the test?

And if it is that you are just “bell curving” the exams, and perhaps felt that doing these exams separately would help our most vulnerable students, and this was a strange and unpredictable backfire … why are you doing that? Why assume that these students need you to modify the passing standard for them in any way whatsoever - higher or lower? Our students with disabilities may need additional support taking your test, Texas. That does not mean they are incapable of meeting your expectations.

What is going on here? Texas students, parents, and teachers NEED to know.