Twitter's Search Functions
Twitter's search function, like many of its companions and competitors, allows users to search a variety of information contained on the site like user handles, topics, and keywords. Twitter's search function may look different depending on what type of device is used. Here is a link to some basic directions for Desktop and Mobile authored by Twitter. Below, we have provided further guidance, with screenshots, investigating searches on Twitter.
How to Search
Twitter's search function can be accessed on both mobile and desktop versions of the site, as well as an application available for both iOS and Android devices. On the mobile application, the search icon (i.e., a magnifying glass) is located on the bottom of the screen. On the desktop site, the search bar is located in the top right of the screen. On both versions, users can tap or click in the search bar to search content they want to find on Twitter.
On the mobile application, the search page can be accessed through tapping the magnifying glass icon.
Once opening the site on desktop, the search bar can be found in the top right.
Users can specify and limit their search using filters. Filters include arranging results from most relevant (“Top”), most recently posted (“Latest”), searching for users that share the search term (“People”), and searching for posts that include specific media along with the search term (“Photos” and “Videos”). Users can tweak these filters more. For example, a user can limit the categories of users in the results (“From anyone” versus “People you follow”) and location of the tweet “Anywhere” versus “Near you”).
On the mobile application, users can select filters to limit their results.
Users can further specify their results by selecting other filters.
On the desktop site, users can refine their search even further by using Advanced Search. Advanced Search allows users to add more specific ranges and categories to their results, similar to a scholarly database.
After searching a term on desktop, users can select Advanced Search to refine their results.
Filters in Advanced Search provide users with the most specific result limitations.
How Search Works
Twitter's content algorithm, similar to Instagram, appears most often in a reverse chronological order on the home screen. However, when searching on Twitter, the "Top" filter appears to have the most algorithmic influence compared to other filters.
Top allows users to see only the “most relevant” posts in a search term. Twitter admits that the relevancy of a Tweet is determined by an algorithm, but they only disclose that the algorithm uses popularity, keywords, and “many other factors.” What could those factors be?
Interestingly enough, Twitter also claims that the “Top” filter will provide users with “Tweets [they] are likely to care about most first.” So, this particular algorithm seems to take into account data about users, such as their behavior and engagement (e.g., favorites, retweets, replies, etc.). Twitter's proprietery algorithm uses three different components to calculate revelvancy of content involved in the search process (Twitter 2011):
Static signals, added at indexing time
Resonance signals, dynamically updated over time
Information about the searcher, provided at search time
All of these signals are fed into the algorithm to produce search results for a user. Although Twitter seemingly provides the inputs, it's unclear what exact information is being harvested from the generalized categories. However, users have relatively more flexibility in toggling the impact of the algorithm during the search process.
Search Complications
Like most social media companies, Twitter is somewhat shy on exposing the mechanisms that drive their search algorithm(s). On the surface, Twitter’s hidden algorithm could seem like an attempt to preserve the uniqueness of their functionality, like that of a secret family recipe passed down through generations. However, this secrecy introduces an opportunity for Twitter to add unsavory ingredients to their algorithmic recipe without the knowledge of its dinner guests.
Unlike most social media companies, Twitter has recently announced a new in-house program that serves to indentify problems and bias in their algorithm. The Responsible Machine Learning Initiative has already exposed some issues in their algorithm and have since put measures in place to fix them:
The image-cropping algorithm showed a racial bias leaning toward images of white people that could affect search results; and
Tweets about right-leaning politics had added algorithmic amplification in 6 out of 7 countries tested.
Users can interact with the Responsible Machine Learning Initiative by adding "#AskTwitterMeta" in their tweet. This could be useful if a user spots a problem with a Twitter search.
Takeaways
When searching on Twitter, be aware of how their algorithm might affect search results.
Search results can be skewed based on how a user behaves on the site.
To limit the impact of the algorithm's influence, use filters to narrow results.
Using Advanced Search on Twitter's desktop site or avoiding the "Top" filter will produce the most pure results.
Identify and collect any issues with a Twitter search and use #AskTwitterMeta in a Tweet to relay the message.