Photo by Robert Bye on Unsplash
We are heavily connected these days. There aren’t many articles circulating that haven’t already been scanned and uploaded to a database somewhere.
I’m consistently reminded how lucky my generation is by older academics on the limitless access to journal databases and their advanced search methods we have now.
We get it, we are blessed by the times. In a sea of articles, how do we find what we truly need for a project? This post will cover two search methods I recently employed in my latest papers: the Snowball method and the Search Query method.
The Snowball method is where one finds the one article that may or may not be exactly what they were searching for and then moving forward, “forward snowball”, to see where that paper was cited or backward, “backward snowball”, to see what citations that paper used to build their case. This is also known as “forward channeling” and “backward channeling”. A resource I found most beneficial during my first attempt a paper called Guidelines for snowballing in systematic literature studies and replication in software engineering by Claes Wholin.
With the ACM Digital Library, we have the flexibility to filter through articles by just about anything. During an assignment, I found an article by Stigall, Waycott, Baker, and Caine (2019) that had an interesting query to obtain the results they sought.
The researchers used a query for ACM’s digital library to search for articles containing the various terms for Voice User Interface (VUI) and their respective topic of interest, which was aging/elderly adults. By using the OR operator between the variations of terms they sought and the AND operator for the conjunction of the two groups. They excluded anything from the search results that were prototype focused or didn’t have older adult use. In the case of my paper, papers that were focused on prototype usability were included. I added strict parameters to my search, restricting articles to have the criteria of having been published within the past two years, associated with the Special Interest Group on Human-Computer Interaction (SIGCHI), and containing certain keywords related to voice user interfaces, usability, and frustration. The following was the result of my query:
[[Keywords: “voice user interface”] OR [Keywords: or “vui” or] OR [Keywords: “ai speaker”] OR [Keywords: or “speech” or] OR [Keywords: “natural language interface”] OR [Keywords: or “frustration” or “behavior” or “affect” or “emotion” or “design” or] OR [Keywords: “user experience”]] AND [[Publication Title: “voice user interface”] OR [Publication Title: or “vui” or] OR [Publication Title: “ai speaker”] OR [Publication Title: or] OR [Publication Title: “interactive voice system”]] AND [Publication Date: (01/01/2018 TO 12/31/2020)]
Granted, this query could use some adjustment, but it provided me with my initial set of papers and I proceeded to whittle down the set until a handful of relevant ones remained.
The good news is that you really don’t have to unless you are writing about the effectiveness of the method itself. Researchers may have a preference towards one or the other but it really depends on you and what your end-goal is. Snowballing is heavily used in systematic reviews but finding the initial article set starts with a simple(?) query. I used the query to narrow down the list of articles I was to sift through for relevance and then began snowballing from there.
Some searches will still render thousands of results. Don’t feel discouraged by this, just keep refining your search until a manageable set is returned. The journal database you use may also have additional filtering features that can help with this.
ACM Access: Accessing the full PDF for the ACM resources I provided will require either a subscription or for the user to access the site via a library proxy. University students will be able to go through their libraries to access ACM via their proxy. The abstracts and respective citations are openly accessible to the public, however.