How we assess what students have learned is one of the most hotly debated topics in education today. High-stakes testing determines whether students get into the college of their choice and helps school districts judge the effectiveness of their teaching staff.
But all this focus on testing raises concerns that the urge to test is overwhelming what really matters: whether children are actually getting the education they need to thrive in an increasingly sophisticated, knowledge-driven economy.
We asked Joseph McDonald, professor of teaching and learning at NYU’s Steinhardt School of Culture, Education, and Human Development, about the best ways to assess the success of teaching. McDonald notes that the best teachers are always assessing – that is, doing things deliberately to figure out whether they are getting through to their students – even when they are not handing out weekly quizzes and final exams.
In this Q&A, McDonald talks about:
- The most effective assessment methods
- The advantages of doing summative and formative assessments
- Scheduling tests to avoid over- or undertesting
- Analyzing the data teachers gather from assessments
- Adapting teaching to diverse classrooms
Let’s go to Professor McDonald’s answers:
What are the most effective ways to assess a child’s learning?
When people think of assessment today, they often think first of standardized tests – that is, ones developed by testing companies and used by states, schools, and districts in standardized ways to measure what students have learned with respect to some criteria. These tests are an important part of U.S. education, and are likely to remain important for the foreseeable future. But these tests are also likely to change in format and prevalence.
For example, they are likely to be increasingly administered online, and to be adaptive – that is, adapt the level of prompts to an estimated level of a particular test taker’s understanding or skill level. This change will save time in both testing itself and in test prep. But the time saved won’t be enough by itself to deal with widespread concerns today about overtesting.
As the standardized testing opt-out movement has proved powerful in states such as New York and Illinois, some policymakers are ready to cut back on testing demands. Meanwhile, there are signs too that a new kind of standardized testing system may emerge – at least for teenagers and young adults. This system would be one to support – and provide legitimacy and validity for – a growing interest in digital badging, which use digital credentials to convey core academic knowledge and other competencies that can’t be measured by traditional assessments.
The individualized pursuit of badging is likely to emerge as an important design element in initiatives to reimagine 21st-century high schooling – so long as it can be undergirded with an authentic and valid system of assessment.
For now, however, the most adventurous area of assessment is assessment in teaching. This interest in assessing teaching is encouraged by research on learning that reveals that learning is never simply additive. Learning nearly always involves unlearning too – or subtracting some previous understanding.
So even as they teach toward the goal of new understanding, good teachers probe continuously for current understanding and emerging understanding. Many do such probing through questioning, but this use of questioning is still rarer than it should be.
One culprit here is that many teachers – even those who are well versed in the content they teach – are untutored in the typical patterns of misunderstanding that are characteristic of their content area. As most of us know from our own experiences as students, content expertise by itself is not enough to make a good teacher.
But there are other ways besides questioning to assess knowledge while teaching. A simple technique is to pause in the middle of a lesson, and to ask everyone to write down or tweet or otherwise share briefly their understanding of [such and such] at this instant.
A similar and favorite device of new teachers, because it’s relatively easy to use, is called entry and exit tickets. Students have to “pay” their way into class by quickly writing down – or otherwise sharing – accounts of their current understanding of something on the day’s learning agenda. This is the entry ticket, which has the advantage of previewing the learning agenda.
At the end of the class, the students pay their way out by accounting for changes in their understanding: “Take no more than one minute to write down your current understanding of [such and such], and don’t sign your name.” This is the exit ticket, which has the advantage of introducing metacognition into the learning process – something that research on learning suggests plays a big role in learning. After class, the teacher can quickly read and compare entry and exit tickets to estimate the range of cognitive change, and the relative need to reteach, review, or move on in the curriculum.
How do you define, interpret, and strike the right balance between formal and informal means of assessment?
Instead of formal and informal, I prefer summative assessment – at the end of some unit of instruction, or some “gate” like the end of fourth grade, or the completion of the unit on gasses in the chemistry course – and formative assessment.
The summative ones should be much less frequent in a student’s education than the formative ones. Students should always feel that a summative assessment is an appropriate capstone of some kind, and have time to prepare for it – and summon up metacognition for the purpose.
The best summative assessments ask students to construct a response or a set of responses – rather than select right answers from a list as in multiple-choice exams. The best summative assessments are also ones that seem authentic – that is, true in some fashion to how the assessed knowledge is actually used in the world. It’s much better, for example, to ask students to write a movie review than to ask them to list the major requirements of a movie review.
Formative assessments are really a kind of teaching. What students understand is not, after all, confined to how they cognitively enter and exit a particular instructional period. It evolves during the period, too, and nearly every one of a teacher’s moves can help it evolve in the right direction.
This learning progression only happens, however, if the teacher teaches in ways that continually inquire about the evolution – not just en masse as in exit tickets (What do my students understand now?) but also in individualized ways (What does Jose understand at this moment? How about Mirabelle?). Good teachers are alert to signs of learning – Mirabelle’s look, Jose’s posture – and, of course, they literally ask individual students to explain their understanding.
These are the kind of teachers, however, who do not ask questions to fish for right answers, and who do not discount wrong answers. They explore whatever answers they get in order to unearth misunderstandings – so that these can be cleared away, and so that scaffolds can be communally erected to reach higher levels of understanding across the class.
And they teach students to ask questions, too, as a good way to put them in touch with each other’s emerging understandings. Sadly, however, too few teachers know how to do all of these methods well. The gap may be in part an ironic artifact of the fixation of much assessment on right answers.
How often should assessments occur? Can you easily over- or undertest and/or assess, whether it’s formal or informal?
How often? Summative assessments should be infrequent, while formative ones should be highly frequent. Undertest? Yes. A college professor who lectures incessantly and assigns only one long paper is guilty of this.
Overtesting is much more prevalent, however – at least P-12. Overtesting with standardized tests, including test prep, is particularly dangerous. Standardized tests use a relatively small sample of items drawn from a large domain.
Overtesting can invalidate the sample by displacing attention to the larger domain. Educators may think, “This test doesn’t ask questions about Asian history so we’re not going to cover it.” Overtesting can also displace its practice in it, by encouraging fourth-graders to spend so much time reading short passages and answering multiple-choice questions about them – in preparation for a reading test – that they don’t read any actual books.
What should teachers do with the data they get back from these assessments to modify their teaching approach to enhance student performance?
I’m studying data use in schools right now, and one of the things my team and I are learning is the phrasing of your question is misguided when it comes to the heart of the process of using data in teaching.
What I mean is that only a small proportion of the data that’s useful in teaching is data that teachers “get back.” This is standardized test data – whether from annual testing, or so-called benchmark testing (periodic tests that claim to predict outcomes on annual tests). But good teachers understand that the data they derive from teaching itself – from pressing for understanding, unearthing misunderstanding, and so on – is much richer. And so is data that they gather themselves in formative assessments of various kinds and in collecting other purposeful samples of student work – and that in the best schools, they examine together with their teaching colleagues.
Indeed, what we’re learning from our study is that opportunities to look together at student work with colleagues, and regularly talk together about teaching and learning – is one of four key things that teachers need in order to use data to benefit their students’ learning. The other three things are:
- An inquiry-guided teaching habit
- A good data management system, including tech-savvy colleagues, at the school level
- A smart policy environment that doesn’t oversell the power of standardized testing or downplay the measurement error inherent in it
In a diverse, inclusive classroom, how do you adapt and/or tailor assessment methods to address differing needs and ways of learning?
Lots of formative assessment helps, because it makes diverse levels of understanding and skill development discernible and thus more addressable. It helps a lot, too, to build a classroom culture that acknowledges diversity as a learning asset for the whole group, and that encourages and creates opportunities for all students to work on “hard stuff” with the understanding that one person’s hard is nearly always another person’s easy – but any group of humans is always “smarter” as a whole than any of its parts.
And it helps to create summative assessments that to the furthest extent possible permit a range of ways to express one’s understanding of what they cover.
Finally, it helps to incorporate a range of technology supports into teaching, and formative and summative assessment, too – for example, smartphones and apps that can turn text into sound and vice versa, post questions that others can read and respond to, and respond in an instant to teacher inquiries about what students know – even graph the results. And all students should have access to such tools – not just students whose Individualized Education Plans say they should. Most classrooms are far more diverse and inclusive than most people imagine.