4.3 Finding Missing Voices Using Search Engines
How they work
Did you know that both academic databases and search engines like Google are owned by for-profit corporations? A key difference between these types of tools, though, is that search engines treat their algorithms as trade secrets, in part to protect their competitive advantage. Academic databases, meanwhile, have clear rules and are typically open about how their search functions operate.
Another key difference is that search engines’ revenue stream is largely built around advertising revenue, where academic databases rely on subscriptions. This means that the design of tools like Google are shaped by what keeps users clicking. Most search engines and social media algorithms also function as black boxes, meaning users (and sometimes employees!) cannot see exactly how they work. We give an example of this when we discuss Google AI search summaries in the next section.
In Chapter 2, we saw examples of how certain voices or perspectives, often from marginalized groups, can be hidden from us because of exclusionary algorithmic practices like shadow banning. Because algorithms are black boxes, we cannot predict when our search results will be impacted by social biases, nor can we know how to fix them. We can see this reflected in the ways that search engine results may:
- Prioritize paid content: Commercial interests can push certain perspectives to the top of the search results.
- Reflect user data: Factors like search history, location, and browsing behaviour, may mean that users get very different results for the same search.
- Reflect corporate politics: Legislation such as the Online News Act
, as well as the corporate response to it, may mean that users in different regions may have certain content restricted or hidden.
How to find diverse voices using search engines
a) Apply Strategy 1: Use a Range of Search Terms
As we saw with academic databases, adding inclusive terms to our Google searches, to specifically name the perspectives of equity-denied groups we wish to include, is an important strategy. It’s here that we’re more likely to see the impacts of searching with community-preferred and colloquial language, more than we would have with academic databases. For example, searching for a phrase like ‘rainbow community’ might lead to community blogs, grassroots organizations, which academic databases don’t index or prioritize.
b) Advanced Search Mechanics
Click the boxes below to reveal tips for searching more efficiently with Google
1. Filetype
This limits your search results to a specific type of file (like PDFs, Word documents, PowerPoint presentations, etc.)
Example: epistemic justice filetype: PDF
2. .site
This restricts your search to a specific website or domain (like .edu, .gov, or a specific organization)
Example: Indigenous education site: cbc.ca
3. NOT (-)
This excludes certain terms from your search results.
Example: If you’re looking for health equity initiatives outside of Canada, you could search: “Health Equity Initiatives – Canada”.
4. Google Advanced Search
Google’s advanced search allows us to filter results by date, region, language, and more. Let’s use our case study as a quick example: if we want to find Kenyan voices, it is likely better to switch our “region” to Kenya rather than use “Kenya” as a keyword.
Stop and Reflect
- How did using these strategies shift or expand what you found compared to a typical Google search? Did they help you locate diverse voices or perspectives?
- Which search strategy had the biggest impact on your results?
- How easy or difficult was it to ‘work around’ the algorithm to bring underrepresented perspectives to the surface?
All activities can also be found in a downloadable workbook. Visit the ‘Using this Resource‘ page to access the workbook in MS Word and PDF formats.
Your chapter authors, share our results for this reflection activity in the video below.
A system where the inner components, logic, algorithms, etc., are invisible or otherwise hidden from users.
The practice of partially blocking, limiting, downranking, or reducing the visibility of content without the awareness of creator or users.
