Peer Review in This Class

1. Why peer review, and why it's not a warm-up for the teacher

There are two reasons we do peer review, not just one. The first is the obvious one: you get feedback that makes your work stronger before the final version is submitted. That alone justifies the time. The second reason is less obvious and matters more in the long run: giving good feedback to someone else is the most direct way to develop your own judgment about what good work looks like. When you have to articulate why a peer's claim is unsupported, or why their evidence doesn't quite match their argument, you build the muscle of seeing those problems in your own work later.

Peer review is not a warm-up for the teacher's review. It is its own valid practice, with its own outcomes. If you treat it like a courtesy step on the way to the "real" feedback from the teacher, you will get out of it exactly what you put in: nothing. The teacher can see things a peer cannot; a peer can see things the teacher cannot. Both rounds matter.

A specific case: when you sit across from a classmate and try to defend your argument while they push back, you discover the weak points in your reasoning faster than any teacher comment can show you. The discovery is in the conversation, not in the markup. That conversation is what every defense in this course is preparing you for. Peer review is daily practice for the high-stakes defenses.

2. Two skills, not one

People talk about peer review as if it's one skill. It's actually two, and the second is harder than the first.

Giving feedback

Reading someone's work carefully enough to see what's strong, what's underdeveloped, what's missing, and what's wrong. Then saying so in a way that is honest, specific, and useful. The skill is selection + precision. Vague positive feedback ("nice job, looks great") is not feedback; it's a courtesy. Vague critical feedback ("this is unclear") is not feedback; it's a complaint. Specific, evidence-anchored feedback ("your claim in paragraph 2 is supported by data from Source A, but Source B contradicts it and you don't address that contradiction") is feedback.

Receiving feedback

Hearing someone else's read on your work without either dismissing it ("they didn't get it") or collapsing under it ("they're right, my whole draft is bad"). The skill is processing what's useful, judging what's mistaken, and revising deliberately. This is harder than giving feedback because your draft feels like an extension of you, and criticism of it feels like criticism of you. It is not. Good revision starts with separating the two.

3. What makes peer review fail

Three failure modes show up over and over. Recognize them in your own behavior so you can name them and stop.

3a. The reciprocity trap

You give your partner soft feedback because you want them to give you soft feedback back. They do. Now neither of you has improved. The reciprocity trap is the single most common reason peer review produces no learning. It feels polite and it is the opposite of polite. You have wasted your partner's revision time and they have wasted yours.

3b. The surface trap

You spend the review fixing your partner's grammar, formatting, citation style, and word choice. You feel productive. You have not engaged the argument, the evidence, the structure, or the reasoning. Proofreading is not peer review. Save grammar comments for the last 5 minutes if at all.

3c. The performance trap

You give long, sophisticated-sounding feedback that demonstrates your own knowledge but doesn't tell your partner what to actually do. "Your discussion of riparian rights is interesting but could engage more deeply with the broader hydrological-political-economy literature" is performance feedback. "Your paragraph 3 cites two riparian-rights cases, but neither one tests your stronger claim about Indigenous water access: can you find one that does?" is useful feedback.

4. What makes peer review work

1. Know the criteria before you read. Pull up the rubric or the assessment prompt. You cannot give useful feedback against criteria you haven't internalized. Re-read the criteria first; then read the draft.
2. Read the whole draft once before commenting. The first paragraph often makes sense only after you've seen where the argument lands. Don't comment in real time on the first read; just read.
3. Comment specifically and locate precisely. "Paragraph 4, the sentence starting with 'Carbon sequestration…' doesn't connect to the claim in paragraph 2 about ecosystem services. Either add a connecting sentence or move it." Page-and-line-locatable. Action-specifiable.
4. Aim for 3–5 substantive comments, not 30 minor ones. Pick the comments that will produce the biggest improvement. A draft with 30 minor markups looks heavily reviewed and is harder to revise than a draft with 4 sharp comments.

5. Four protocols we use in this class

Pick the protocol that fits the moment. Default to Warm-Cool for early drafts; default to Tuning Protocol for late drafts or high-stakes deliverables. The strand-coded review (Section 6) is layered on top.

Protocol 1: Warm-Cool feedback

Reviewer offers two kinds of comments. Warm = specific strengths ("your introduction sets up the stakeholder map clearly"). Cool = specific stretches ("your conclusion drops the equity argument you raised in paragraph 2: bring it back"). Both must be specific and locate to a part of the draft. Useful for early drafts; pairs evenly. Don't skip the warm half: knowing what's working is also information for revision.

Protocol 2: I notice / I wonder

Reviewer makes observations ("I notice…") and asks questions ("I wonder…"). No prescriptions. "I notice that you cite four sources, and three of them are from the same year. I wonder whether the dispute looked different in earlier or later periods." This protocol is good when you want the writer to do the thinking, not just take orders. Best for substantive arguments where there's a gap but you're not sure what the right move is.

Protocol 3: By the rubric

Reviewer rates the draft against the actual rubric (K/U, T/T, C) AND defends the rating with specific evidence from the draft. "I'd rate this at T/T 5–6 because the analysis is strong (you compare three sources well) but you don't push to the transfer step (what should the reader DO with this argument)." Useful for drafts approaching submission: gives the writer a concrete sense of where the work currently sits on the actual scale.

Protocol 4: Tuning Protocol (presenter-led)

Used for high-stakes drafts and the rehearsal sessions before Defense Panel / Stakeholder Roundtable / Policy Panel. Structure:

1. The writer presents the draft + states the specific question they want feedback on (3 min).
2. Peers respond in two rounds: warm round, then cool round (5–10 min each).
3. The writer listens silently during the rounds: no defending, no clarifying.
4. Writer reflects out loud on what they heard (3 min).

The silent-listening rule is what makes this protocol powerful: writers who normally jump in to explain are forced to hear the feedback fully before responding. Adapted from the National School Reform Faculty Critical Friends Group tradition.

6. Strand-coded feedback (K/U + T/T + C)

Your assessments are graded across three strands separately. Your peer review should be too. Layer the strand frame on top of whichever protocol you're using.

• K/U review questions: Is the content accurate? Does the writer demonstrate solid understanding of the relevant concepts? Are facts cited correctly? Are misconceptions present?
• T/T review questions: Is the analysis going beyond what's in the sources? Is the writer making connections, comparisons, judgments? Is there real transfer to new contexts, or just summary?
• C review questions: Is the writing (or the oral defense) clear for the intended audience? Is the structure logical? Is the precision of scientific language present?

Give 1–2 comments per strand. A single strand-coded comment is more useful than five general ones because it tells the writer exactly which dimension of their work the comment is about.

A specific note on "this is unclear": sometimes "unclear" is actually a K/U problem (the writer doesn't fully understand and is masking it with vagueness) or a T/T problem (the writer hasn't done the analytic work and there isn't anything to be clear about yet). Diagnose which strand the unclarity is coming from before commenting.

7. AI in peer review

AI can read a draft and give feedback faster and often more comprehensively than any peer can. It is tempting to substitute AI feedback for peer feedback entirely. Do not do this. The peer review pact in this course is built around three benefits: (a) feedback ON your work, (b) practice GIVING feedback, (c) practice being read and challenged by another human. AI substitution gives you (a) and removes (b) and (c). Two thirds of the value disappears silently.

Legitimate AI uses in peer review

• After your peer has reviewed your draft, ask AI to check whether you missed any obvious issues. "My peer flagged X and Y. What else might a careful reader catch?" Disclose this use.
• When you're stuck giving useful feedback because you don't understand the topic well enough, ask AI for help understanding the topic FIRST. Then give the feedback yourself.
• Use AI to clarify what a comment from your peer means if you didn't follow it.
• Use AI to test how the revised draft sounds to a different audience.

AI uses that cross the line

• Pasting a peer's draft into AI and forwarding the AI feedback to them as if it were yours. This is dishonest twice over.
• Asking AI to review your own draft and skipping the peer round entirely.
• Pasting the peer's draft and asking AI "is this good?" for a one-line verdict you then transmit. There is no review here, only a transmission.
• Using AI to generate generic-sounding praise to fulfill the warm-feedback requirement without doing the reading.

The peer review reciprocity trap (Section 3a) and the AI-substitution trap are first cousins. Both produce frictionless, harmless-feeling exchanges that leave neither party smarter. Both are forms of doing the appearance of the work.

8. How to receive feedback (the harder skill)

Most students get more training in giving feedback than in receiving it. Receiving well is harder.

1. Read all the feedback before responding to any of it. Don't jump to defend the first point you disagree with. See the whole set first.
2. Separate the feedback from the person giving it. A bad feedback giver can still raise a real issue. A good feedback giver can still be wrong. Judge the feedback on its merits, not the relationship.
3. Look for the same critique twice. If two reviewers flag the same thing, that's a strong signal. If only one does, weigh it more carefully. They may be right, or they may be the outlier.
4. Decide what to act on, document what you rejected. You don't have to act on everything; you should act on what's most likely to improve the work. When you reject feedback, note why.
5. Thank the reviewer specifically. "Thanks for the comment on my counter-argument paragraph: I see what you mean and I'm going to add the equity dimension you flagged." Specific thanks rewards specific feedback. Generic thanks rewards generic feedback.

9. One last thing: about the social cost

Giving honest critical feedback to a friend feels worse than giving vague praise. That asymmetry is real and it is why most peer review is bad. There is no clever trick that removes the social cost. What changes is what you decide to pay it for.

If you give your partner vague praise, you have spent zero social cost AND zero useful information has changed hands. If you give your partner specific honest feedback, you have spent some social cost AND your partner's draft is now better than it would have been. The transaction goes from "free, nothing happens" to "small cost, real improvement." Most adults eventually learn that the second transaction is the one worth doing.

It helps to remember: the partner who gives you honest feedback respects you more than the one who gives you empty praise. Empty praise treats you as someone who can't handle the truth. Honest feedback treats you as someone whose work matters enough to push on. Be that second kind of partner. The friendships built on honest exchange are stronger than the ones built on agreement that nothing is ever wrong.