Level 2 Costa Questions Explained

Most people conflate Level 2 and Level 3 questioning and wonder why their lesson plans fall flat. The distinction matters because students hit a wall when you ask them to evaluate something without first being able to analyze it. I ran into this repeatedly when training new teachers last year. Level 2 Costa Questions sit between basic recall and high-level judgment. They require students to process, interpret, and connect information rather than simply retrieve it or pass final verdicts on it. Think of them as the bridge between comprehension and critical thinking.

What Level 2 Costa Questions Actually Are

Level 1 questions ask what happened or what is said. Level 2 questions ask why it happened, how it connects to something else, or what pattern emerges. Level 3 asks whether it was right or wrong, good or bad. The most common mistake I see is teachers writing prompts like "What do you think about this character?" That looks like analysis but it's actually Level 3 disguised as Level 2. "What do you think" requires a value judgment. Level 2 asks students to explain meaning, not judge it. Valid Level 2 prompts include: Explain the relationship between X and Y. What can you infer about...? How does this event connect to...? Describe the pattern in... These require students to work with the text, not pass final judgment on it.

Why Level 2 Gets Skipped (and Why That Hurts Students)

I had a colleague who wrote all her discussion questions at Level 3 because she wanted "deeper" conversations. Her students shut down. They hadn't been asked to actually interpret the material before being told to evaluate it. It's like asking someone to critique a recipe without ever having tasted it. Level 2 questions force students to slow down and read carefully enough to notice cause and effect, contradictions, or underlying assumptions in the text. Skipping this step means students produce surface-level opinions dressed up as deep thinking.

Get the Full Details

Level Questions Examples at Alan Matheny blog
Level Questions Examples at Alan Matheny blog

How to Write Level 2 Costa Questions

Start by identifying the key relationships, causes, effects, and patterns in the material. A question that targets any one of these is likely Level 2. Avoid question stems that begin with "Do you think," "Is it fair," or "Should." Those push straight into Level 3 territory. Here are stems that consistently produce Level 2 responses: Why did the author include this detail? How does section three relate to section one? What evidence supports the claim that...? What is the main difference between...? How would you describe the tone in paragraph four?

These all require inference, comparison, or explanation. None require moral or aesthetic judgment.

A Real Case Where This Got Complicated

Recently I was reviewing a unit plan for a middle school science class. The teacher had written: "Is it right to use animals in testing?" This is clearly Level 3. I suggested she reframe it as: "What are the possible benefits and limitations of using animal models in medical research?" Now students have to analyze tradeoffs rather than announce a position. The second question produces more substantive written work because students need to cite specific examples from their reading. That single rewrite cut the amount of shallow argument from my grading pile roughly in half. Not every rewrite does that, but it's a useful benchmark.

Costa's Level of Questioning Poster Set with Student Activities!
Costa's Level of Questioning Poster Set with Student Activities!

Level 2 Costa Questions in Practice

Here are a few more examples across subjects: English: What can you infer about the protagonist's motivation from her actions in chapter five? Science: How does the data in table two support or contradict the hypothesis stated at the beginning? History: What factors contributed to the outcome of the treaty? Math: Explain how the formula in section 4.2 relates to the problem on page 37. The trick is consistency. If you mix Level 1 and Level 2 questions too heavily, students default to looking for the single correct answer. If you go straight to Level 3 without scaffolding, they guess. The ratio that tends to work is roughly three Level 2 questions for every one Level 3, with Level 1 sprinkled in only where needed for context.

The one scenario where Level 2 completely fails is when the source material has no ambiguity or complexity. If you hand students a straightforward manual and ask them to infer meaning, they'll just restate instructions in different words. Level 2 questions depend on the text having enough depth to interpret. Pick better source material or move to a different questioning level.