It’s testing season. Cue the prep rallies aimed at hyping up
students for the standardized test they have been preparing all year for. This
isn't one of those anti-testing pieces. On the contrary, I believe testing is a
good thing. It is a necessary tool to fairly evaluate student knowledge and
progress, compare students in a school, district, state, or across the nation,
assist teachers and school and district leadership in identifying areas of
strength and needs for growth, and, albeit controversial, can be helpful in
identifying effective and ineffective teachers. The problem is that the way we
currently use assessments doesn't effectively do any of this. We don’t need to
do away with testing (and let’s be real, tests aren't going anywhere). We need
to be serious about testing reform. And while there are many places to start,
one of the biggest priorities should be differentiation.
Imagine a fourth grade student reading a passage written for
what should be his independent reading level. After the passage, he is given
six questions meant to demonstrate his understanding of the text. One question
may ask him to identify the main idea, draw a conclusion or make an inference
using information the author has given, or to identify a factual piece of
information from the text. That fourth grade student only answers one question
right. The next passage he reads is on a third grade reading level, but tests
the same skills. This time, that fourth grade student gets 3 questions right.
The next passage is on a second grade reading level, testing the same skills,
and he gets them all correct. From this scenario, we learn a lot about this
fourth grade student. First, we learn he is not reading on grade level. We can
ascertain he reads about a third grade reading level. Second, we can see which types of questions
he gets right (his strengths) and the types he gets wrong (areas for growth).
This is differentiated testing. From this, we can gain
useful data that can be used to provide targeted instruction to this student
and help him progress. Computer adaptive testing uses an algorithm to tailor
assessment test questions to each individual student to truly assess their
needs. It is the first attempt to differentiate standardized testing and it should
be a norm across the country.
It is standard practice for teachers to differentiate
instruction for all students. Some examples of this differentiation include
grouping students based on their instructional reading levels and assigning
different types of books and activities with respect to those levels, assigning
different classwork and homework based on ability, and to provide different accommodations
such as the use of manipulatives to help students be more successful with a given
task. All year long, teachers differentiate to meet learners where they are.
Yet, states give the standardized one size fits all test to students, which
fail to really account for what students may or may not know.
Take the example of the fourth grade student. By simply
giving that student a series of fourth grade reading passages, we aren't
assessing if he can analyze text in a complex way, because he can’t read it. Instead,
we are frustrating students by giving them a series of reading passages above
their comprehension. At some point the child is likely to start guessing and
randomly selecting answers just to be finished. As educators, we gain no real data
from it other than the fact that the child can’t read (which we probably
already knew). This “non-data” is then used to rate teachers and schools and to
determine allocation of funding and resources by the state education agency and
the federal government.
The idea of computer adaptive testing already exists with
the Smarter Balanced Assessment. It is used in 18 states, with much push back
from teacher unions who argue there is too much testing and that this is simply
about data collection. I certainly agree that there is too much testing and
part of testing reform should include limiting the number of times a student is
tested to three (beginning, middle, and end of the year). I also agree that
this is about data collection, but it doesn't have to be a bad thing. Our
profession should, in part, be data driven. The problem is that data is only
useful when used appropriately. Differentiated standardized assessments provide
useful data on the whole student that can be used to inform instruction and
help that student be successful.