Skip to Main Content
JMU Libraries logo .jmulib-logo-purple{fill:#450084;}
Loading

Physician Assistant

Best Evidence for Type of Question

Evidence pyramid

Evidence Pyramid” by Tufts University can be reused under the BY-NC-SA license 4.0

evaluating

IF I APPLY(D)

The IF I APPLY(D) framework below is a particularly helpful one as it asks us as searchers to critically reflect on our own biases in how we are searching for information. Are we being open and curious to what possible answers exist, or are we confirming what we already believe?

As you search and review search results, asking yourself a few quick questions about the sources you are finding can help you quickly determine if it's worth taking it to the next step. And by reflecting on our own biases we can better adjust our search strategies going forward to ensure we are finding the best possible information for the question at hand.

The Personal Steps

I - Identify
Do you have any emotions around this topic? What identities are affected by this topic and should be represented in the research?

F - Find
What sources will provide overviews of the topic? Find various sources to see different perspectives on the issue

 

I - Intellectual Courage
Am I looking outside my comfort zone to find the best information? If not, what can I do to help myself find those sources?

Are any new or unexpected aspects emerging related to my topic that may need to be explored?

 

The Source Steps

A - Authority
What authority or expertise does this author have to write about this topic?

P - Purpose/POV
What agenda might the author or publisher have? Who was it written for? Who published it?

P - Pertinence
How relevant is this source to your information needs?

L - List of Sources
Do they have citations? Are they accessible and reliable?

Y - Year
How does the year affect the information?

D - Data

If data was presented, does it support the author's claim? Are there any noticeable methodological issues?

 

Framework adapted from Phillips, K. (2019). IF I APPLY. Retrieved from https://guides.libraries.psu.edu/IFIAPPLY

Deep Reading

Once we have started to collect some information that looks reliable and relevant to our needs, we can conduct more robust evaluations of the content itself.

As we discussed in the "How to Read Scholarly Articles" section, each part of a research article serves a distinct purpose in contextualizing and detailing the author's objectives. Asking yourself these questions in each section can better help you understand, evaluate, and later summarize the information you read!

A quick note about peer review

**There are plenty of credible sources that are not peer-reviewed (i.e. books, newspapers, government studies, trade journals, etc). Peer-review can be a great indicator of a valid scholarly source, but it is not the only one. Use your own critical thinking and evaluative skills to determine if a source is accurate and relevant to your research. You can also talk with your librarian or your professor if you are still unsure.**

SIFT

When searching for information outside of scholarly databases, we should use the same principles as in IF I APPLY(D). Examining the source itself is an important step. But when evaluating claims we find through Google, the news, or social media, it is also important to fact check those sources. Here is where "Lateral Reading" comes into play. This is a skill employed by most fact checkers and involves looking outside the source itself. The SIFT method, developed by Mike Caulfield, is a useful framework to consider when reading laterally.

Stop: Stop and think about the source and your reactions to it. This is where a lot of the first questions (IF I) from IF I APPLY(D) come into play.

Investigate: Again, if you have critically evaluated the source using the APPLY(D) steps, you have already done this!

Find better coverage: Now is when you start searching outside of the source. Are there other sources reporting on the same issue? Are their conclusions supportive or do they deviate from what your source has found? There may not be consensus, and that's ok! This is about looking at the context to help you more critically evaluate.

Trace claims: Does your source cite any of its claims? Trace those citations back as far as you can to help determine the credibility and accuracy of them. Oftentimes information on the internet is stripped of context, these last two steps help you re-determine that.

Image above is licensed under CC-BY 4.0 License by creator Mike Caulfield

Evaluating information generated by AI

GenAI has become a go-to for many people seeking answers to questions they have. While these tools can be a valuable part of the research process, there are some drawbacks that must be considered, including the risk of the AI generating inaccurate, false, and/or biased information. Evaluating the information you get from AI follows much of the same advice as with other information. It's important to recognize any biases you may be bringing to your query and critically analyze the generated information for potential bias. And as with any online information, lateral reading--like the SIFT method--is your friend! Identify the claims the AI has generated and do quick Google searches to see if other sources (even Wikipedia!) support these claims as well. This video from University of Maryland Libraries shows what this process should look like.

But ChatGPT gave citations for its claims!

ChatGPT in particular has garnered quite a reputation for "hallucinating" information and creating false citations. These citations often look real and use real author names and real journals, but when you dig just a little deeper, it's impossible to find. When you do find citations through a GenAI query, it is crucial that you then search for source itself. This can be done using Google Scholar or the Libraries Search. If you can't find the source, it is likely to have been hallucinated. Reach out to your librarian if you need help with this or finding new sources!

Matrix Template & Example

A matrix is a useful tool that can help organize all of your research articles into one table for your research papers and assignments.