The SIFT Framework

SIFT is a framework developed by Mike Caulfield that can help you evaluate information within the broader information ecosystem, rather than simply accepting a source at face value. SIFT stands for: Stop. Investigate the source. Find better coverage. And trace claims back to their original context. Let's take a look at each step. 

Stop. When you first encounter a resource, or an answer from an AI tool, stop before you start reading deeply, citing it, or sharing it. Ask yourself: Do I know where this information came from? Do I know whether the claim and the resource are reputable? And perhaps most importantly: What am I trying to accomplish? Are you looking for a scholarly source to cite? Background information? A news article explaining research? Knowing your purpose helps you decide whether a particular resource is appropriate. And if you're not confident about the source, that's a signal to move beyond the page and investigate further. 

Next, investigate the information. Rather than staying on the page you're looking at, practice lateral reading. Open another browser tab and look beyond the original source. One way to do this is by asking the 5 Ws: Who created the information? When was it published or updated? Where did the information come from? What kind of resource is it? And why was it created? For example, who is the author? What expertise do they have? When was the information published, and is it still relevant? Where did the author get the information? Are there citations you can verify? What kind of source are you looking at - an article, book chapter, website, blog, or something else? And why was the resource created? Is it meant to educate, entertain, persuade, or sell something? You don't necessarily need to answer every question in every situation. The goal is to gather enough context to understand what you're looking at. 

Next, find better coverage. Don't rely on a single source - or a single answer generated by AI - when you can look at what other credible sources have to say. Look for reliable databases, scholarly sources, credible news organizations, and other trusted references. Then compare what you find. Are multiple researchers reaching similar conclusions? Are there important disagreements? What does the broader conversation look like? The goal isn't simply to find sources that agree with you. It's to understand the evidence and where there is consensus or disagreement. 

Finally, trace claims, quotations, or other evidence back to their original context. If a source refers to a study, quotation, statistic, or other resource, follow that trail. Ask: Does the original source actually say what this source says it says? This is particularly important when evaluating AI-generated information. AI tools can provide references that look legitimate, but the claims made by the AI may not fully match the source material. Tracing information back to the original source lets you check for yourself. 

So remember: Stop. Pause before accepting or sharing information. Investigate. Look beyond the original source and gather context. Find better coverage. Compare the claim with other credible sources. Trace. Follow important claims back to their original context. SIFT gives you a practical way to move from simply finding information to actually evaluating it. In the next video, we'll put all four moves into practice.

[Questions? Contact us. lib.uoguelph.ca] [Music by Google Flow] [This work is licensed under  CC-BY-NC-SA 4.0 , except as otherwise noted.]