Complete guide to Webb's DOK Level 3 (Strategic Thinking). Definition, action verbs, 40+ question stems by subject, common mistakes, and how to write items that require reasoning with evidence.
Webb's Depth of Knowledge (DOK) Level 3 is the strategic thinking level. Students at DOK 3 reason, plan, use evidence, and justify their thinking — there's no single correct answer, and different students may arrive at different (defensible) responses. This is where teacher-made assessments most often stop short: writing a DOK 3 item takes deliberate effort, but it's the level where "did they really learn it?" gets answered.
DOK 3 is the workhorse of a rigorous unit test, and the whole point of a Socratic seminar or performance task. This guide gives you 40+ DOK 3 question stems by subject plus the "reasoning with evidence" rule that separates DOK 3 from every level below it.
What DOK Level 3 Actually Measures
Norman Webb defined DOK 3 as requiring reasoning, planning, using evidence, and thinking at a higher level. Key markers: multiple valid approaches, decisions at multiple points, evidence required to defend the answer, and no single "right" response. Students must construct a chain of reasoning, not follow a procedure.
The tell for DOK 3 is that if you asked two students to solve the same item, they might arrive at different (both defensible) answers. If the answer is unambiguous, it's DOK 2 or lower — no matter how many steps it takes.
Action Verbs at DOK 3
Analyze (a relationship, argument, or system)
Justify, Defend, Argue
Assess, Evaluate, Critique
Compare (with novel criteria)
Draw conclusions with evidence
Formulate a hypothesis
Develop a logical argument
Investigate (a claim)
Support or refute
Cite evidence
Interpret (complex data)
Predict (with justification)
40+ DOK 3 Question Stems by Subject
Science
"Analyze the data. What is the relationship between fertilizer amount and plant growth? Use at least two data points to justify."
"Argue whether the experimental design is strong enough to support the conclusion. Point to specific weaknesses."
"Draw a conclusion about which factor most affected the reaction rate. Cite two pieces of evidence."
"Formulate a hypothesis for why the population is declining. Explain the reasoning behind it."
"Assess the claim that this study proves the vaccine is unsafe. What in the design supports or undermines the claim?"
"Predict how modifying the setup would change the results. Justify with what you know about the underlying process."
"Investigate why plant A grew twice as fast as plant B despite identical conditions. Propose a testable explanation."
"Support or refute the claim that this is a healthy ecosystem, using at least three ecosystem indicators."
Math
"Analyze the two graphs — which representation better communicates the trend, and why?"
"Justify your choice of solution method. Compare it to an alternative and explain the trade-off."
"Draw a conclusion about the meaning of the mean, median, and mode of this data set — what do they suggest about the distribution?"
"Formulate a real-world problem that would require solving a system of equations. Explain your reasoning."
"Assess whether the sample size is large enough to justify the study's conclusion. What would you need to know?"
"Argue whether the graph is misleading. Point to specific design choices."
"Investigate whether these three data points lie on a linear, exponential, or quadratic function. Show your reasoning."
"Predict which model best fits the data. Use residuals to justify."
ELA
"Analyze the author's use of imagery in Chapter 7. What specific words and images build the theme?"
"Justify whether the character's decision at the end of the novel was ethical. Use at least three pieces of textual evidence."
"Compare how the two authors treat the theme of friendship. Which treatment is more nuanced, and why?"
"Assess the credibility of the two sources on the same event. Which is more trustworthy, and how did you decide?"
"Argue whether the poem's central metaphor is effective. Use specific lines."
"Formulate an alternate interpretation of the ending. Support with textual evidence."
"Investigate whether the author's use of dialect strengthens or weakens the narrative voice. Cite specific passages."
"Support or refute the claim that the protagonist is a hero. Use at least three examples."
Social Studies
"Analyze the causes of the French Revolution. Rank them by impact and justify your ranking."
"Assess whether the New Deal succeeded in its stated goals. Use economic data from at least two sources."
"Argue whether Truman's decision to use the atomic bomb was justified. Address the strongest counterargument."
"Compare the reliability of two primary sources on the same event. Explain your assessment of each."
"Investigate why voter turnout differed so much between the two elections. Propose an explanation."
"Draw a conclusion about the long-term impact of the Industrial Revolution using at least three data points."
"Justify whether the Constitution's original design accounted for modern political parties. Cite two provisions."
"Predict how the Cold War might have unfolded differently if the Cuban Missile Crisis had escalated. Ground in specific events."
How Much DOK 3 Belongs in Each Assessment
Common Mistakes at DOK 3
Grading position instead of reasoning. A confident "yes" with no evidence shouldn't score higher than a hedged "it depends" with strong evidence. Weight reasoning highest.
Vague rubric criteria. "Quality of argument" gets 30 different interpretations. Fix: score three parts explicitly — claim (2 pts), evidence (4 pts), reasoning connecting them (4 pts).
Confusing DOK 3 with DOK 2. If the answer is unambiguous once you've done the analysis, it's DOK 2. DOK 3 requires that multiple defensible answers exist.
Not requiring evidence. DOK 3 rubrics should require at least two pieces of evidence or sourced quotations. Otherwise students opine instead of reason.
Providing the analysis in the stem. "The author's metaphor critiques war — analyze how" hands the student the analysis. Rewrite: "How does the author use figurative language in the third stanza, and what does it reveal about her view of war?"
Generate DOK 3 Items Instantly
AssessmentWiz's Test Generator lets you set a DOK 3+ target — the AI produces items with sourced evidence prompts, rubrics that require multiple pieces of textual support, and defensible-alternative marking guidance.
Webb's DOK Level 3 is strategic thinking. Students reason, plan, use evidence, and justify their thinking. Multiple valid approaches exist and different students may arrive at different (defensible) answers. The tell: if the answer is unambiguous, it's DOK 2 — no matter how many steps.
What are DOK 3 examples?
"Analyze the data — what is the relationship between fertilizer amount and plant growth? Use two data points." "Argue whether Truman's decision to use the atomic bomb was justified — address the strongest counterargument." "Assess the credibility of the two sources — which is more trustworthy?"
What's the difference between DOK 3 and DOK 4?
Both require reasoning with evidence. The difference is time and complexity: DOK 3 is typically a single-class-period question (essay, extended response), while DOK 4 requires extended investigation over multiple days or weeks with multiple sources of data. DOK 4 is closer to a research project than a test question.
Can multiple choice items be DOK 3?
Yes, if the item requires reasoning between options that all have some plausibility. Example: "Which of the following BEST explains the pattern in the data?" where students must weigh multiple defensible answers. Most multiple choice items land at DOK 1 or 2, but well-designed ones can reach DOK 3.
How much DOK 3 should be on a rigorous unit test?
About 35% — the heart of the second half of the test. One extended-response item + 3-5 short-answer items where students defend a claim with evidence.
Try These Related Tools
Test Generator — Build comprehensive unit tests with sections, DOK levels, and answer keys
Benchmark Test Generator — Generate Regents, STAAR, FAST, SBAC, PARCC, NJSLA, SOL & NWEA MAP-style practice benchmarks in seconds