Skip to main content

Join the Community

or Close


Suggestions for Better Testing

This year, I was on maternity leave during "testing season." For the first time, I had no contact whatsoever with the NY State ELA exam. When I returned to work, I was curious about how things went, and I caught up with some of my NYC colleagues both at my school and in other schools. The consensus seemed to be that this year's Common Core State ELA test was not significantly different than the ones we've seen in the previous two years. What struck me from the conversations, though, was the chorus of common (heh, heh) criticisms of the test itself and the conditions under which we administer it to students.

"I'm vehemently against these standardized tests and all that they are used for," a veteran humanities teacher tells me. "But I'm willing to meet them half way if the test is at least reasonable." A lot of teachers I know feel this way. We know standardized tests don't measure the scope of our teaching or our students' learning in a meaningful way. We see them being used to scare, label and punish children and educators, and we oppose these practices on principle and for the sake of our students and profession.  However, we are not totally against giving a reasonable yearly standardized test in reading and writing. There are a few key changes that might make the tests themselves a bit more reasonable. (I'm not touching how they are used in this post--just talking about the design of the tests.)

"First of all," the same teacher says, "Stop timing students. So many students run out of time, and do poorly on the later portions of the test each day." This means students' scores don't even accurately show what they can do with these passages and questions. Instead they represent a combination of reading and writing skills and speed. Is speed-reading a standard? No, it is not. So either stop timing the test or at least give ample enough time so as to take it off the table as a significant factor in performance in a supposedly common core-aligned test.

Second, let's consider the length and number of reading passages. In the old NY state ELA exams, the seven or so reading passages tended to be about a page long. Now they are often three pages long. There's nothing wrong with longer passages, since students need to be able to read texts much longer than that--but if the passages are longer, we need to limit the number of passages students must read in a single sitting. This isn't just about the amount of time provided--it's about what's reasonable for a child's concentration in the inauthentic context of a test. After reading one or two of those longer passages myself, I'm bored and losing steam. Reading one after another of these long texts without the connection and purpose one has when reading voluntarily is unnecessarily painful. If volume of texts is important, keep them short. If longer texts are needed, pick just a few.  

While we're at it, can we get rid of the extra set of reading passaged with multiple choice questions on day two of the three day test? It's just confusing and unnecessary. Each day should assess students in a coherent way, just like we expect of ourselves in our classroom. Day one can be multiple choice, day two can be writing. I'm not sure why we even need three days... 

Finally, enough with the tricky questions and the beta-testing of tricky questions on students during the actual test. Many of these are so *challenging* that even I cannot determine a correct answer, and I am actually a good test taker. If the testing company can't do a decent job with test creation, just show the new questions to some teachers in advance and we'll tell you if they are fair for our students.  Do not add stress to children with poorly constructed questions.

Educators, what else would make the tests themselves more reasonable and humane for our students?


1 Comment

Christi Carpenter commented on May 22, 2015 at 1:35pm:

Considering two broad types

Considering two broad types of standardized tests:

Norm-Referenced Assessment: A test or other type of assessment designed to provide a measure of performance that is interpretable in terms of an individual's relative standing in some known group. 
Criterion-Referenced Assessment: A test or other type of assessment designed to provide a measure of performance that is interpretable in terms of a clearly defined and delimited domain of learning tasks.

( Linn, R. L., & Gronlund, N. E. (2000). Measurement and assessment in teaching (8th ed.). Upper Saddle River, NJ: Prentice Hall.)

I wish those creating the tests would prioritize criterion-referenced assessments, so the focus would be on creating clear questions that allow students to show what they know, not questions that "trick" and confuse students so they can be ranked against each other.  

Unlike many of my colleagues, I believe in the necessity of standardized tests.  Having taught in Detroit Public Schools for five years, I saw scandalous things happening to students at school: very little teaching and learning was taking place and students were being blamed for the conditions of which they were victims.  Standardized testing, at its best, creates an element of transparency and, ultimately, accountability so conditions can change. 

But, of course, we need to make high-quality tests that seek to illuminate student progress and align with the skills and knowledge they will need post-secondary.

Subscribe to Blogs by Ariel Sacks

Stay Informed

Sign up to receive the latest news and events through email!

Sign Up