6.4 Peer Review Essentials
Peer review is a valuable way to obtain second opinions on your work. In the workplace, the review can serve many purposes:
- Gain insight on the expressed ideas and approach
- Check for accuracy and support
- Double check technical data and information
- Obtain additional opinions and perspectives
- Assist with proofreading
- Obtain approval prior to final submission
Peer reviews can be conducted by intra-departmental and extra-departmental colleagues, team members, and even AI-assistants like LLMs. While nothing can replace the wealth of knowledge, experience, and contemporaneous contextual expertise that a colleague can bring to the review process, AI-assisted peer review is becoming the most convenient and speedy way for writers to obtain that review.
Note: Unless you as a reviewer obtain permission from the author, it is a violation of copyright to upload another author’s work to a LLM.
Peer review entails having a peer—a fellow student who is familiar with your assignment, a colleague who understands your purpose in writing, or a LLM—read and offer feedback on the effectiveness of your document and how it might be improved. While peer review ultimately benefits the reader as a form of “quality control” ensuring an effective final product, it also has clear benefits for both the author and the reviewer. In the process of peer review, authors can get helpful feedback on how they can improve their drafts, and reviewers can learn from the document they are reviewing, reinforce the document’s requirements, and, in college situations, help to align with assignment and grading criteria. Writers then come back to their own draft with a fresh perspective. For both, it is a helpful way to review content, structure, style and formatting before submitting your final document to its intended reader.
It is common in technical workplaces for peer review to occur, especially when complex documents with critical operational and financial implications are created. Asking a co-worker, manager, or subject matter expert to review your document demonstrates due diligence and a commitment to quality work.
Whether you are in an educational setting or a workplace, the review will be governed by the request set out by the author.
Asking a Human for a Review
Keep the following strategies in mind to obtain the maximum benefits of peer review:
Author Strategies
- Be prepared: Submit the most polished and complete draft possible in order to fully benefit from peer review. A partial or very rough draft can benefit from some preliminary review, but a completed draft can elicit more helpful feedback.
- Give the reviewer some guidance: Describe the concerns you have about your document at this point. Alert your reviewer to areas where you would particularly like some feedback. If the reviewer is not familiar with your purpose and audience, fill them in.
- Be open to new ideas: Avoid becoming defensive about the feedback; remember, your reviewer is trying help you improve your draft. Consider alternative viewpoints you may not have thought of before; listen before deciding what you will do.
Consider advice carefully: Think critically about whether the reviewer’s suggestions will help you to improve your draft and if they are appropriate for your purpose and audience. What and how you revise is ultimately up to you; seek additional advice if you are unsure what to do.
Reviewer Strategies
- Know the document’s purpose: If the document is created for a college course, review the project or assignment description and/or grading rubric before reviewing the draft; know what the document is trying to achieve before you assess whether it needs revision to achieve it.
- Focus on the positive: Be honest and critical, but also point out strong areas within the writing where things are working well. Use a positive, constructive tone to discuss areas that could use improvement. Don’t gush (“Your paper is so awesome!”) and don’t trash (“This totally sucks! Rewrite the whole thing!”). Remember your purpose is to help the author find ways to improve the draft; recognize what is already good, and suggest what needs further work.
- Be specific: Explain why a sentence, paragraph, or image needs improvement; explain why you as the reader are confused or bothered by a specific phrase or passage, or why the logic does not flow for you (be “reader centred”).
- Be courteous: Be aware of your language use and tone when addressing peers. Avoid patronizing or “talking down” to your peers when giving advice.
- Don’t edit: Your job is reviewer, not editor. Don’t fix errors or phrasing problems; just point out areas that need improvement. You might offer ONE sample correction to demonstrate what you mean, but do not engage in wholesale editing. The purpose of peer review is for each person to learn how to edit their own document based on reviewer feedback.
- Be efficient: Don’t overwhelm your author with too much detailed feedback. A page that has more feedback notes than content will be very difficult to process. Focus on a handful (three-to-six?) of the most important revisions that are needed to help improve the draft. For example, if there are numerous spelling errors, don’t point them all out; highlight one or two and then comment that the draft contains numerous spelling errors that will need fixing before it can be submitted.
Using a LLM to Obtain a Review
Authors are now often making routine use of LLMs to obtain feedback on their work, be it at early or later stages of the document creation process. The LLM can give you rapid insights that often will include ideas that you may not have yet considered. You can obtain quick, informal feedback by simply asking: “Give me some feedback on this document” or if you want more context-specific input, your request could be more developed. See the example below for how you can go about this:
Asking a LLM for a Directed Review
In addition to applying many of the principles described in “Asking a Human for a Review” you can follow this process to obtain specific feedback from a LLM, which can save you time since you are providing direction. Here’s how to go about the process:
- Upload the document to be reviewed.
- In the context window, include this prompt (which you can modify for your own needs):
Prompt:
Please read and review the uploaded document. The document is created to address the following situation [describe the context]. It is created primarily for [identify the audience] who [has/does not have] technical knowledge in the [subject area]. With this document I want to achieve [state the purpose]. I would like a thorough review based on the following criteria along with explanations for your suggestions:
- Document context and requirements: [Offer details relating to contextual requirements or criteria that the LLM would need to narrow its focus.] Read through the document and suggest areas where I could explain the context of the situation better.
- Organization of ideas: Read through the document and suggest areas where ideas could be better organized. Ensure that the introduction is complete and that I have a clear purpose statement and overview and that the contents of the document align with the stated purpose.
- Argument: Read through the document and check the argumentation for any gaps in logic and use of fallacies. [Add details about the argument you are trying to prove.]
- Support and evidence check: Read through the document and check to ensure that key ideas are supported with verifiable and accurate evidence. Suggest areas where I could develop ideas more. Identify gaps in detail. [Add specific details pertaining to the context and topic.]
- Writing skills: Read through the document and suggest ways that I can create a more friendly yet professional tone. Also suggest how I can improve the grammar and mechanics.
- Readable page design: Verify that the document has been designed in a manner that will promote readability and useability. Suggest ways that I can make it easier for the audience to find information.
Using both a human reviewer and an AI-assisted review would help you obtain thorough feedback on your work that would include context specific and more generalized suggestions on improvement.
