2013 Research in Review

This post summarizes the research I published in 2013. The work divides roughly into three components covering: 1) the role of time in search, 2) the use of human computation in search, and 3) computer mediated communication.

Time in Search
For many years a consistent theme in my research has been to look at the role that time plays in people’s information interactions. Our past information interactions shape what we know and provide insights into how we will use information in the future. This year in particular, I explored:

• Search trends
• Interaction with change
• Slow search

Search trends: Several of the papers I published in 2013 focus on understanding, modeling, and predicting trends in what people are interested in. For example, right now people around the world are searching for cocktail recipes and New Year’s resolution ideas. Understanding and predicting that behavior allows us to support it better.

• Behavioral Dynamics on the Web: Learning, Modeling and Prediction
• Towards Supporting Search over Trending Events with Social Media

Interaction with change: Not only do people’s interests change over time, but the information they interact with also changes. For example, the search results returned for a query you issued before may be very different when you issue it again. This is often a good thing – the new results are likely to contain more personalized, relevant, fresh content. But it can also be confusing, because consistency provides context. My thesis research looked at how people interact with search results as they change, and I revisited this theme this year as search engines are getting better equipped to use change intentionally to benefit their users. Expect to see more on this front in 2014.

• Understanding How People Interact with Web Search Results that Change in Real Time Using Implicit Feedback
• Benevolent Deception in HCI

Slow search: On top of trying to understand and model the impact of time in search, I started to actively exploit it. Search engines are optimized to try to return results as quickly as possible, but may be able to provide a higher quality experience by taking more time. I will be writing and posting more about slow search in the coming year.

• Search Engines’ Quest for Speed
• The Case for Slow Search
• Slow Search: Information Retrieval without Time Constraints
• Slow Search: Improving Information Retrieval by Including People in the Search Process

Human Computation in Search
I also devoted a substantial amount of time to exploring the use of human computation in search. This builds on a standing interest in the role other people play in supporting our searches. In previous posts, I have discussed online question asking, the difference between asking and searching, and how to ask an effective question. This evolved in 2013 into an interest in exploring the use paid crowd workers. Human computation not only helps us understand question asking, but also allows us to try out (or “Wizard of Oz”) slow search experiences.

• A Crowd-Powered Socially Embedded Search Engine
• Crowdsourcing for Robustness in Web Search

In addition to looking at how paid crowd workers can augment algorithmic search and question asking, I have also looked at what makes using human computation in search different from using computer computation. One opportunity is that using people may make personalization easier. Personalized human computation has the potential to go beyond existing techniques like collaborative filtering to provide personalized results on demand, over personal data, and for complex tasks. But human computation also presents a challenge, in that human workers are, well, people. The search engine optimization market is estimated at $20 to $30 billion dollars in the United States alone, and poses a real challenge for search engines. If crowd systems are used in search, they will certainly become targets for new types of manipulative attacks.

• Personalized Human Computation
• Raising an Army: Attacking Crowd Systems

Computer Mediated Communication
My remaining publications for the year center around computer mediated communication.

• Remote communication
• Co-located communication

Remote communication: Several projects looked at how people negotiate communications over distance. Working with my husband, I explored how people use availability information to establish communication. We showed that a recipient’s communication decisions are impacted by the fact that they know that their availability state is visible to others. I also looked at how people establish common ground when communicating over IM.

• I Know You Know I’m Busy
• Understanding How the Projection of Availability State Impacts the Reception of Incoming Communication
• Clarifications and Question Specificity in Synchronous Social Q&A

Co-located communication: Another running theme in my research over the past few years relates to co-located mobile device use. Smartphone use is often seen as rude or distracting. We get annoyed, for example, if our dinner companion takes out their phone and places it on their table. However, this doesn’t have to be the case. Phones can encourage people participate more fully in what is going on around them and build stronger ties with their companions. I will post more on this in the coming months.

• Smartphone Use Does Not Have to Be Rude: Making Phones a Collaborative Presence in Meetings