Skip to Main Content

International Social Work (EN): 3. Critically Assess

Libguide for students of the course International Social Work.

Critically assess

In the previous step we used search terms in combination with several operators to find articles that might be useful for a project. This step aims towards critically assessing the usefulness of various sources. This is especially important for information that is not peer-reviewed: inspected by another specialist before publishing. 

Types of information sources

We can discern the following types of information sources. Each types comes with its own characteristics to look for when assessing reliability.

  • Websites: anyone can build a website and present information on its pages. When using information on websites it is therefore especially important to assess the reliability of the pages. Information from renowned organisations in a particular field are often more reliable and therefore usable.
  • Books: The barrier to write and publish a book is often higher than a barrier to start a website. It often requires a large investment in time and money from an author. Books can nevertheless still contain unreliable information, depending on for example the purpose for which the book was written.
  • Scientific articles: The main purpose of scientific articles is sharing new knowledge. Especially peer-reviewed articles (articles that are reviewed by another expert in the same field of study) are therefore often seen as reliable. Other factors to look for are the reputation of the author, the quality of the journal in which is published, and the study design.  
  • Organisation reports: Organisations may publish reports about various topics and for various purposes, often related to specific projects. Important aspects to look for are the reputation of the organisation and the purpose of the project.

Artificial Intelligence (AI) and reliability

Artificial intelligence can be a useful tool for brainstorming and adjusting the formulation of a research question. It is however not suited for directly finding information because of the following reasons:

  • AI can only access the surface web and therefore has a limited view on the information available.
  • AI looks for trends that occur in information regardless of the source. It can therefore base information on unreliable sources, for example reddit pages. Answers can also be based on outdated information.
  • AI tends to hallucinate (make up) information when it cannot give an answer to a question. 
  • AI is trained on a dataset containing a lot of information up to a certain point in time. It cannot give information past the most recent point at which the training is updated.
  • AI is known to fabricate source references based on actual papers and authors as well as give real-sounding titles that do not actually exist. It often also presents information that is not actually derived from the source to which is referenced. This means that there is a high risk of plagiarism.
  • The process in which AI chooses words in an answer is untransparent, meaning the information may contain a bias towards certain topics, possibly presenting a very narrow view on a topic.

The SHB/UKB Information Literacy Working Group (2025) developed a checklist for assessing Generative AI content to overcome these challenges. This document can be found below. 

Assessment criteria

The CRAAP test (Rotterdam University of Applied Sciences, n.d.) can be used to assess the reliability of information sources for your project. You can download the checklist below.

Currency: how up to date is the information?

  • When was the information published?
  • Has the information been updated or revised?
  • Is the information current enough for your topic?

Relevance: is the information relevant for your subject?

  • Does the information relate to your subject, or answer a research question?
  • Who is the intended audience of the information (e.g. general public/ experts/ researchers)?

Authority: is the creator an authority on this subject?

  • Who is the author/publisher/source/sponsor?
  • What do other sources say about the creator? What information do you find about this online, is the creator perceived as reliable, or controversial?
  • What makes the author an expert in this field? (e.g. education, background, connected to a particular organization)

Accuracy: is the information true, correct and reliable?

  • Is the information supported by evidence?
  • Are there references to used sources?
  • Has the information been reviewed by other experts (e.g. peer review)?
  • Do other sources confirm the information?

Assessing articles

Microsoft. (2025). Copilot. [AI assistent]. Retrieved at 6 november 2025, from copilot.microsoft.com