What website should I use to search regarding a specific topic?
If you're ever not sure where to start, I recommend reviewing the Article Databases page for an overview of some key resources. Their descriptions should give a general indication of what is/is not a good fit for your research topic. The Subjects A-Z list is another great tool if you want to see a full list of U of T databases organized by subject area.
How do you tell if a database is scientific or not?
The library has pre-organized key databases by subject area (e.g. Environmental Sciences) to help streamline this process for you, but if you're ever interested in judging the "scientific vigour" of a database, it can be helpful to look online to see how transparent their indexing and decision-making process is (e.g. see Web of Science's Journal Evaluation Process and Selection Criteria web page). That can give you a good sense of coverage and vetting that goes into their content. Likewise, it can also be helpful to review what their peer review policies are for indexed journals (Web of Science example).
Which databases are most preferred by profs? How do I know if an article is trustworthy?
If a professor has a preferred database for students to use, that will usually be expressed in the assignment instructions. Other times, they may give guidance about the type of journals that you're pulling literature from (e.g. peer reviewed journals, ISI journals in the case of this course, etc.). Other times, students may be given free reign to determine where to pull literature from. The key step in this case is to make sure you're using quality references; check out the Reading Articles page for tips on evaluating article quality.
How do I tell if an article is a primary source?
Database filters such as Web of Science's "Article" document type can be a helpful first step to scope your results to empirical/primary studies. If such a filter is not available, then you will want to look for clues that the paper is reporting on original research, rather than summarizing existing literature like in a review/secondary article. This can include things like a hypothesis statement and accompanying methods/results sections within the paper.
What happens if your keywords don't yield results, or your search is too restrictive? What if I don't find what I'm looking for?
Retrieving low numbers or topic-irrelevant results is generally linked to one of three things:
Are techniques used in Web of Science applicable to other search databases? Where do you find out how a database registers certain characters, like the hyphen?
Many databases use similar functions (e.g. an asterisk * to truncate), which helps streamline the process of searching from database to database. if you're ever unsure, check out the database's help documentation, where they give a specific breakdown of their search operators and modifiers. You can usually find these documents directly linked within the database interface (look for things like 'search help,' 'search tips,' etc.), or they are also a quick Google away (e.g. searching 'Web of Science truncation' in Google should bring up the database's help page).
Please see the links in this guide's menu for pages on database-specific search tips, as well as the page on Building a Search Strategy for Boolean operators (AND, OR, NOT) and modifiers (e.g. phrase searching with "quotation marks," wildcards such as *) commonly used across most databases. If you have any questions about using a specific resource, I'm also happy to advise!
How can the literature searching process be streamlined so it's easier for more people to use?
The honest answer is that literature searching, especially when done well, takes time and is often iterative (i.e. your first search strategy is not likely to be your last). There's certainly more work required to set up a good search using the methods described in class, but I would argue that the pay-off at the end will save you a lot of time and energy when you actually have to weed through the results. If you have any questions about specific resources, I encourage you to peruse the search tip pages linked in this guide or reach out to me to set up a consultation.
Is there a way to have all the search engines work together so that you don't have to search between them?
Many of the databases highlighted today are offered through unique vendors that have distinct interfaces, so unfortunately they need to be searched separately. Even in cases where databases have the same vendor (e.g. both MEDLINE and EMBASE are offered through the OVID platform), I still recommend searching them separately because a) they may have different search tools unique to a specific database (e.g. BIOSIS offers a "Taxonomic Data" search category that the Web of Science: Core Collection does not), and b) they may have unique controlled vocabulary that varies from database to database (e.g. MeSH for PubMed and MEDLINE vs. Emtree for EMBASE).
What's the best search engine, Google Scholar?
Different search engines have various strengths/weaknesses, so there isn't one that I'd designate as "the best" in that sense. It can be helpful to learn about how they retrieve information and what factors go into their search algorithm. For example, Google uses geographical location and your search history to influence what results are retrieved. A search engine like Duck Duck Go, by contrast, does not track personal search history and can be considered more "neutral" in terms of the results it retrieves. If you're aiming to be comprehensive, try searching across a variety of search engines to look for unique results.
How can AI devices be interconnected to article databases to help obtain more relevant primary research articles?
The key thing I'll stress here is that the AI tools in circulation today (e.g. ChatGPT) are language processing interfaces, not search interfaces, and as such are not recommended for use as a research/literature searching tool.
Another facet of this is that most research databases are licensed products that require payment to access (which in our case is covered by institutional affiliation, via library negotiations with database vendors). Major copyright concerns come into play when this licensed content gets ingested into AI tools, and they are also otherwise limited to whatever openly available content they are trained on (e.g. the web). So my advice here would be to stick to the research databases, which can help avoid some known key issues with AI tools such as ChatGPT:
How can I access an article that is seemingly unavailable?
Please see the page on Locating Article Full Text for tips on how to find and / or request articles.
What is an effective method to choose between different results? How do I choose between similar articles?
Not all papers are created equal, and some will align closer with the scope of your research topic than others. Check out the Reading Articles page for resources on critically evaluating journal articles, as well as ideas on possible "research red flags" to keep an eye open for. If you need additional help on this particular part of the literature review, I also recommend reaching out to your TA or instructor for additional tips and resources.
How do I show previous literature that I've opened or used?
Most databases will let you create a free account where you can save / email / export specific articles that you're interested in or save a search strategy to rerun at a later date (rather than having to build it again from scratch). Alternatively, citation managers are a great way to keep track of the literature you've searched. Most research databases allow you to export citations to these management tools, and many citation managers even have built-in annotation tools for you to "attach" notes to papers you've read. I'm particularly fond of Zotero, but feel free to reach out if you have any questions around specific citation managers!
Is there a digital way of requesting help (or access) from the Library without coming to campus?
Definitely! For immediate help, I recommend using the Ask a Librarian chat service (also linked on this guide's "Get Help!" page). We also have an email research help form, though turn-around time may take a bit longer depending on how in-depth your question is.