When I responded to Students Should Be Tested More, Not Less by Jessica Lahey and the related study by Henry L. Roediger III and Jeffrey D. Karpicke in the blog post Students Should Be Tested Less, Then Not at All, resulting comments and Tweets suggest that the topic of moving toward fewer and even no tests needs further discussion and clarification.
One aspect of debating the role of tests in education revolves around the term “test.” For the general public, Lahey’s headline, I am certain, triggers a relatively basic view of tests—students answering questions created by a teacher or a standardized testing company. For the general public, distinguishing between teacher-made tests and high-stakes standardized tests or between summative and formative assessments will likely not change that basic perception.
And thus, Lahey’s headline is certain to cause more problems than good in the public debate about accountability, education reform, teacher effectiveness, and student achievement.
Many have noted the headline problem, but quickly argue that Lahey’s article, and Roediger and Karpicke’s research make a valuable case for formative assessment, adding that the study also raises concerns about high-stakes standardized testing and seeks to encourage more in-class formative assessments.
As I noted in my initial post, however, Roediger and Karpicke’s study is flawed—in their narrow defining of learning as retention and recall as well as their idealizing of testing (they raise concerns, but argue the positives outweigh those negatives).
Here, then, I want to clarify that calling for fewer and then no tests is not hyperbole on my part and not some idealized goal unfit for the real world of public school. As a co-editor with Joe Bower and building off the work of Alfie Kohn, I have detailed how to de-grade and de-test the writing classroom—practices I began as a public high school English teacher for 18 years and then expanded as a writing teacher in first-year seminars.
In terms of magnitude, yes, high-stakes standardized tests are by far the most corrosive types of tests impacting negatively teaching and learning. Standardized tests remain significantly biased by race, class, and gender, and their high-stakes status encourages the worst characterizations of schools, teachers, and students while also draining valuable resources and time from teaching and learning.
Despite the tradition of using standardized tests, U.S. education should end all high-stakes standardized testing—with a reasonable compromise being the use of randomized samplings of NAEP periodically to monitor large trends in measurable student outcomes (recognizing the limitations of measurable outcomes).
While ending standardized testing, or even lessening its frequency and impact, would be a huge move forward, continuing in-class testing would remain a misguided practice. Let me offer a few reasons and then an alternative.
Even the best in-class and teacher-made tests are reductive and only partial representations of learning because testing by its nature is artificial.
For example, consider testing any courses or student activities outside the so-called core curriculum, such as visual art, music, or athletics.
A course in painting that seeks students who can create their own original paintings does not begin with paint-by-numbers, and art teachers would never rely on traditional in-class tests of any kind to represent a student’s ability as a visual artist.
High school football teams, as well, line up each Friday night and the high school players actually play football; they don’t sit in desks and take tests to decide the best team (see Childress for an elaboration on this idea).
In other words, education has conceded the least accurate process, testing, to the core courses that we deem essential, while allowing in the so-called non-essential courses and activities the most authentic demonstrations of learning and teaching practices.
If tests are inadequate for determining a student’s ability in chorus, art, or soccer (where we allow and require students and players to perform the real task), I suggest that they are also inadequate for English, math, science, and social studies.
Now, before offering a brief consideration of what should replace testing, let me also explain that testing fails because it occupies time better spent doing real activities and receiving authentic feedback from teachers. This is the same issue with isolated grammar instruction as it fails the teaching of writing.
Isolated grammar instruction does not transfer to student writing and the time spent on that futile grammar instruction would have been better spent asking students to write. Such is the case with testing—as it wastes time better spent doing whole and authentic activities.
A transition to whole and authentic activities by students in class must begin by reconsidering the place of content acquisition and retention. Most commitments to testing see content as fixed and assume that memorization of that content must come before application, evaluation, or synthesis.
This is the distorted traditional view of Bloom’s taxonomy applied both to instruction and assessment in U.S. education—a view that reduces Bloom’s work on assessment to a linear and sequential model of teaching and learning.
To embrace students engaging in whole and authentic activities instead of tests, the acquisition of knowledge must be re-imagined as the result of that engagement, not a prerequisite to that engagement.
We own and know facts, knowledge, and details because and once we have used those facts in whole and authentic ways. Again, consider how we have learned to paint a work of art, play an instrument, or participate in an athletic event. All of these require some basics, some practice, some artificial preparation, but the real learning comes from the doing, the feedback while performing as a novice, and then the re-doing, and re-doing.
About 60 years ago, Lou LaBrant (1953) lamented:
It ought to be unnecessary to say that writing is learned by writing; unfortunately there is need. Again and again teachers or schools are accused of failing to teach students to write decent English, and again and again investigations show that students have been taught about punctuation, the function of a paragraph, parts of speech, selection of “vivid” words, spelling – that students have done everything but the writing of many complete papers. Again and again college freshmen report that never in either high school or grammar school have they been asked to select a topic for writing, and write their own ideas about that subject. Some have been given topics for writing; others have been asked to summarize what someone else has said; numbers have been given work on revising sentences, filling in blanks, punctuating sentences, and analyzing what others have written….Knowing facts about language does not necessarily result in ability to use it. (p. 417)
And this is essentially my argument about testing.
If we want students to be better at taking tests, then more testing will certainly accomplish that goal (again, that is basically what Roediger and Karpicke show).
But if we redefine learning and frame our teaching goals toward whole and authentic behaviors by students, we must recognize that students learn by doing those whole and authentic things.
Instead of tests, then, and grades, students need extended blocks of time in school to perform in whole and authentic ways (ways that occur in the real world outside of school; ways that occur in art class, chorus, and band, and on athletic fields and courts) along with having teachers observing and offering rich and detailed feedback that contributes to those students trying those performances again and again.
Not tests, whether we call them formative or summative, of the artificial kind, but whole and authentic performances and rich feedback leading to more and more performances.
Again, if you seek examples of what should replace the inordinate amount of time spent testing in schools, visit an art class, chorus, an athletic event—or consider that a central aspect of science courses are labs.
Commitments to testing are commitments to the static classroom where teachers are active, students are passive, and content is central. These commitments are asking very little of students.
I am calling for de-testing and de-grading the classroom in order to increase student activity, engagement, and thus learning in ways that are whole and authentic.
As Childress concludes in his argument that football is better than high school:
What I am saying is that we have a model for learning difficult skills — a model that appears in sports, in theater, in student clubs, in music, in hobbies — and it’s a model that works, that transmits both skills and joy from adult to teenager and from one teenager to another.
For Further Reading