Backward design and alignment
The exam tests what is easy to grade, not what you said the course was for
Your stated goal is that students can critique a published design, and your point distribution is dominated by computing measures from two-by-two tables.
What it is
Designing in the reverse of the usual order. Content-first design runs content, lecture, assessment, objective, goal, so the goal gets written last and describes what you had already decided to cover. Backward design runs goal, objective, assessment, activity, content.
The workshop’s own diagram makes one further substitution worth noticing. At step four it strikes out “lecture” and replaces it with “activity.”
Run it
- Run the exam backwards first. Take last year’s final, write next to each question the goal it tests, and total the points.
- Anything with no goal attached, and any goal with no points attached, is the finding.
- Reverse the chain for next time: goal, objective, assessment, activity, content.
- Run the audit below over the finished design.
Say this
The alignment audit, as given at the workshop, minus its last question
Do your assignments and activities gradually build to the summative assessments?
Are all of the elements of the assignments clearly articulated to students?
Is there first, last, repetition, uniqueness?
Do you have too much information?
Do you see your stated purpose reflected?
Are your activities student centered?
The original had a seventh question, asking whether your lectures run more than fifteen minutes. Delete it. That ceiling has no source, and no controlled trial manipulating live lecture segment length exists (Bradbury, 2016; Rey et al., 2019). Ask instead how long students go without doing anything, which is what the rule was reaching for. See chunking.
In my courses
EPI 601 is being rebuilt this way for Fall 2026, having previously been taught by someone else.
The assessment redesign came directly out of running the exam backwards. Exams drop from 45 percent to 30 percent, split into three equal pieces. Scaffolded study critiques enter at 15 percent as formative lead-ups to the final paper, which keeps its 25 percent but gains a draft feedback cycle. Student-led reading discussion and the oral presentation stay as they were.
Evidence
The alignment principle traces to teaching-centre web pages and practitioner books, none reporting data (Fink, 2003; Wiggins & McTighe, 2005).
The workshop justified it by saying alignment produces meaningful deliberate practice. That link is weak and worth dropping. Deliberate practice accounts for about 4 percent of variance in education, and a direct replication of the founding study failed (Macnamara et al., 2014; Macnamara & Maitra, 2019). The alignment argument survives without it: say that students should get practice matched to what they will be assessed on, which is a claim about your own course.
Note what this cannot tell you. A perfectly aligned course can be perfectly aligned on the wrong goals, and alignment gives you no way to detect that.