ED 615
I have completed several standardized tests during the
course of my life. In elementary school, for several years in a row, I
completed a test called the California Achievement Test. Even though I went to
school in Maryland, we all took the California Achievement Test because
Maryland did not have its own state test at the time. This test was our
equivalent to the MSA or PARCC test that elementary students take now.
In
high school, I took three Advanced Placement tests – one in Calculus, English,
and French. I earned a score of “3” out of “5” on each test. These tests
translated to 14 college credits towards my undergraduate experience at UMBC
and actually allowed me to graduate on time. In high school, I also took the
PSAT and ACT one time each, in addition to the SAT, which I took at least two
times. Because the College Board (makers of the SAT) combined the highest scores
from the English and Math sections- across different test dates- to arrive at the student’s final score, I was
able to maximize my score by taking the test multiple times. I have also taken
the GRE for entrance into graduate school, but I did not fare as well on that
test; I only took it once.
I
do not feel that taking the tests completely demonstrated evidence of learning.
This is especially true with the SAT. In high school after my first time and
before my second attempt at the SAT, I enrolled in an after-school coach class
that my homeroom teacher (who was also a math teacher) provided. What I learned
most from the SAT prep course were strategies for how to approach each type of
test question. It was very effective, but it did not help me display the
content that I had learned during my regular math class.
This example shows the
problem with traditional standardized tests. Teachers can begin to “teach to
the test.” On the opposite spectrum, teachers’ variability in instruction can impose
limitations on a student’s performance. The
instructional sensitivity article shows that not all standardized tests actually
generate the results that they set out to produce. (D'Agostino, Corson, & Welsh, 2007) The study in the
article demonstrated that students whose teachers taught them in the same way
in which the standardized tests were structured scored higher on their tests.
This all stemmed from a case in 1981 (Debra P. v. Turlington) where a student brought
a claim against the state of Florida asserting that students of color were not
taught the material that was included on the state’s minimum basic achievement
test. I have seen a similar example first hand in the middle school where I
work. Two years ago, when I first started working there, I co-proctored a test
for a class of 6th grade students who were taking the Maryland State
Assessment (MSA). Several of the students either asked me to read the word “diagonal”
for them from the test or asked me what the word “diagonal” meant. If so many
of the students had trouble with this one, very important word, how many other
points were they missing? Clearly, these students had not learned everything
that they needed in order to be prepared for the standardized test that they
were taking. The fact that the school had low test scores for several years in
a row made one question how well the classroom instruction was matching up with
the test content. This again occurred last year on a more wide-scale basis when
all of the students in Maryland had to take the MSA again, but the curriculum had
changed to the common core. Thus, the test did not reflect what the students
were learning in their curriculum.
I think what motivated the
development of standardized testing was teachers, parents, policy-makers, and community
members wanting to know how well students from different regions compared, in
terms of acquired knowledge. States could then use that data to make funding decisions
based on how much help the students needed to reach pre-set standards. By allowing
one organization to make a test for everyone, states could eliminate the bias
that each teacher would have imposed had she created her own tests.
The assumption is that giving the same test under the
same conditions to everyone normalizes everything. However, we now know that
various other factors come into play, including the knowledge base of the
teacher giving the instruction, the teaching style of the teacher and prior
knowledge of the students. For these reasons, standardized tests do not always produce
the desired results that their administrators set out to achieve.
References
Balf, T. (2014, November). A Smarter, Fairer SAT. Popular
Science, p. 30.
Black, P., & Wiliam, D. (1998, October). Inside
the Black Box: Raising standards through classroom assessment. Phi Delta
Kappan, 139-148.
D'Agostino, J. V., Corson, N. M., & Welsh, M. E.
(2007). Instructional Sensitvity of a State's Standards-Based Assessment. Educational
Assessment, 12(1), 1-22.
Popham, W. J. (2014). Classroom Assessment.
Upper Saddle River, NJ: Pearson Education, Inc.
I agree I do think that more of a strategy is taught for standardized tests rather than content and application. I remember being told that if you do not know the answer to skip it and that it would not be counted against you in some standardized tests. Do you think it is still valuable to teach strategies and different approaches to take in standardized tests or should we simply stick to the concepts and applications?
ReplyDeleteI have the same conflict with teaching test taking strategies. That in and of itself is a symbol of what is wrong with standardized testing, we use valuable instructional time teaching students how to take a test. In order to better their chance of test-taking success, we gear content delivery to match the style of tests rather than the learning style of students. Further, we are now in this realm of testing higher-order thinking skills through standardized assessment, which hopefully will soon be recognized for being as accurate as the IQ tests of yore. As you described, Juanita, the institution is set on creating these scientific apples-to-apples comparisons of our students, but without accounting for the number of variables or defining valid controls. Do you think this is a worthy goal? In what environment would this model of assessment actually "achieve the desired results"?
ReplyDelete