Lesson Planning

Design backward from the result

Most people plan the way they were taught: open to the next section, find an activity that fits, run it. Grant Wiggins and Jay McTighe named that habit the twin sins of activity-oriented and coverage-oriented planning, and built Understanding by Design as its inverse. Their framework moves through three stages. Stage 1 identifies the desired results. Stage 2 determines the assessment evidence that would show those results were reached. Stage 3 plans the learning experiences and instruction. The engine of the framework is alignment: what is named in Stage 1 must be what is assessed in Stage 2 and taught in Stage 3.

Calling this order backward is a comment on ordinary practice. Wiggins and McTighe report having watched curriculum planning collapse into listing activities, with only a general sense of intended results and little if any attention to assessment evidence. Order matters here because an activity is easy to love and hard to evaluate. Once you write the evidence sentence first - by the end of this period I will have seen every student produce X - an activity either yields that evidence or it does not, and the question stops being whether the room seemed engaged.

One caution comes from the framework's own authors. They chose the unit, not the single lesson, as the focus of design, and say plainly that they do not recommend isolated lesson planning separate from unit planning. Their reason is scale: understandings, essential questions and transfer performance tasks are too large for one class period. An essential question is meant to be revisited across weeks, not answered by the bell. What travels down to the single lesson is the logic - result, then evidence, then activity - and they report that teachers who attend carefully to Stages 1 and 2 find their lesson planning sharpens as a result.

Frames, not validated recipes

Two named frames dominate teacher preparation, and both are worth knowing precisely because they are so often misread.

Madeline Hunter's clinical theory of instruction, developed at UCLA and usually called ITIP, names elements a teacher might build into a lesson. A typical enumeration runs: anticipatory set, purpose, input, modeling, guided practice, checking for understanding, independent practice, closure. Robert Gagne's nine events of instruction run: gain attention, inform learners of the objectives, stimulate recall of prior learning, present the content, provide learning guidance, elicit performance, provide feedback, assess performance, and enhance retention and transfer.

The clearest witness against treating either as a recipe is Hunter herself. In a 1985 article she objected that administrators had come to expect every element in every lesson - she called that a white sauce recipe - and wrote that the elements were designed to help teachers plan, that in no way can a teacher be judged by the inclusion of all of them, that many lessons will incorporate only a few, and that any observer using a checklist to confirm all seven elements does not understand the model. She set her own standard for judging a model: whether it guides behavior, predicts outcomes and stimulates research, not whether it is the final answer. A small tell that these frames are not fixed procedures is that published lists disagree with each other. Hunter refers to seven elements; a widely used university reference page lists eight.

So separate the frame from the practice. Neither Hunter's elements nor Gagne's events carries a formal effectiveness rating as a package. Particular practices inside them do. The What Works Clearinghouse rated systematic instruction as strong evidence on 43 studies, for elementary mathematics intervention, where systematic means reviewing previously learned content, using accessible numbers when a concept is new, sequencing so knowledge builds incrementally, providing visual and verbal supports, and giving immediate supportive feedback that addresses misunderstandings. A different WWC guide rated a Model-Practice-Reflect cycle strong on eleven studies of grades 6-12: the teacher models the strategy, students practice it, students evaluate their own use of it. Ratings attach to specified practices for specified populations. Frames organize a plan; they do not inherit the ratings.

The anatomy of a period

Assume the objective already exists - writing one well is its own subject. What the plan adds is the route from where students are to that objective, and most routes contain the same moves.

Open by activating prior knowledge: a short review or warm-up that pulls the relevant earlier learning into play and, just as importantly, lets you see whether the prerequisite is actually there. State the objective, so the class knows what the period is for. Then input and modeling: the new content in small pieces, with the teacher working an example aloud and narrating the reasoning rather than only the steps. Guided practice follows, with students attempting the same work while you are still in position to intervene. Checks for understanding sit inside the sequence rather than at the end of it. Independent practice comes only once guided practice has gone well. Closure consolidates - a summary, a one-sentence explanation, a final problem - and doubles as a last read on the class.

Barak Rosenshine's principles of instruction describe this shape from the research side, and it is worth being exact about what kind of evidence they are. Rosenshine assembled ten principles from three separate literatures: research in cognitive science, observational studies of teachers whose classrooms made the highest achievement gains, and research on cognitive supports such as thinking aloud, scaffolds and models. He argues that the agreement among three unrelated bodies of work is what should give us confidence. That is a synthesis, not an experiment, and it should be cited as one.

The observational strand speaks directly to planning. In the studies Rosenshine summarizes, less successful teachers gave much shorter presentations and explanations and then handed out worksheets; the most successful spent more time in guided practice, more time asking questions, more time checking for understanding and more time correcting errors, and their students were more engaged during independent work. Where guided practice was too short, students were not prepared for seatwork and made more errors. That is a planning finding, not a style preference: the minutes you spend guiding are what make the independent minutes productive.

Budgeting the minutes

A plan without times on it is a wish list. Write a minute figure beside every segment before the lesson and make the figures sum to the length of the period. The value is not accuracy - novice estimates are routinely wrong - but that the arithmetic forces a decision while you can still make it. Fifty minutes will not hold a twelve-minute warm-up, a twenty-minute explanation and meaningful practice, and the budget makes that visible before the class does.

Two rules keep a budget honest. First, budget the parts you tend to lose, not only the parts you control. When a lesson runs long the casualties are guided practice, the check and closure, which are precisely the segments the evidence supports most directly. Protect them by entering them in the budget first and letting the explanation absorb the overrun. Second, decide the cut in advance. 'Problems five through eight become homework and closure stays' is a plan; deciding at 2:12 with thirty people watching is not.

Be careful with one popular pacing rule. Teaching-center guidance often recommends organizing a period into roughly twenty-minute episodes so that no single mode runs too long, and often justifies it by claiming that student attention falls off ten to fifteen minutes into a lecture. Wilson and Korn reviewed the note-taking, observational, self-report and physiological studies behind that number and concluded that the research provides little support for the belief, adding that the studies failed to account for individual differences in attention. Segmenting remains a sensible planning habit - it stops you from talking for thirty-five minutes without noticing. The specific attention-span figure is not a finding you should repeat.

Checks, misconceptions and the fallback

A check for understanding is a planned moment at which every student produces something you can read, and it belongs in the plan with its wording already written. Write the exact prompt. Decide how you will see every response - whiteboards held up, a two-item slip, a fixed cold-call rotation. Then write the decision rule: which result lets you move on, which result stops you. Without the rule a check is data you collect and then ignore.

Rosenshine's contrast is worth keeping in mind. Effective teachers checked by asking students to summarize the presentation so far, repeat a procedure, or say whether they agreed with another student's answer. Less effective teachers simply asked whether there were any questions and, hearing none, assumed the material had landed and passed out worksheets. Silence is not evidence. He also offers a gauge: across studies the optimal success rate for fostering achievement appears to be about 80 percent during practice, and in one fourth-grade mathematics study the most successful teachers' classes answered 82 percent of items correctly while the least successful reached 73 percent. Eighty percent is a planning target, not a grade - high enough to show students are learning, low enough to show they are being challenged.

Checks earn their place twice over, because retrieval is itself instruction. The IES guide Organizing Instruction and Study to Improve Student Learning rated using quizzes to re-expose students to key content as strong evidence, on nine experimental studies with K-12 students, more than thirty with college students, and a large controlled laboratory literature on the testing effect.

Anticipating misconceptions is the other half of the work. The first key finding of How People Learn is that students arrive with preconceptions about how the world works, and that if their initial understanding is not engaged they may fail to grasp what is taught, or learn it for the test and revert to their prior belief afterwards. So name the wrong answers before the lesson. Write the two or three errors you expect for this specific content, write the sentence or representation you will use against each, and design the check so it can actually surface them - a check whose likely wrong answers do not include the error you predicted cannot detect it.

Then write the contingency in one line: if the check goes badly, what gets cut, what gets retaught now, and where tomorrow picks up. A plan that has already answered that question survives contact with a real class.