This PhD project is part of the broader Australian Search Experience 2.0 project at the ARC Centre of Excellence for Automated Decision-Making and Society (ADM+S) which will generate significant new knowledge about the breadth and quality of information returned in response to diverse search queries, and offers a significantly more realistic perspective than past research for the impact of variations in users’ queries on the results recommended to them. Search Experience 2.0 project aims to significantly advance the state of the art in the field by developing novel methodologies for the study of search and recommendation in conventional (text-based) and emerging (voice-based and AI assisted) search interfaces, producing new insights into their impact on users and content creators. You will get the opportunity to interact with an interdisciplinary team of researchers who are working on different aspects of this project across social and computer science.
Potential proposals could encompass one of the following areas:
Developing and Evaluating Novel Methods to fulfil Users’ Information Needs: Building on existing or proposed findings, we can prototype novel interfaces that use contextual information (e.g., device type, location) and account for query variation to provide meaningful search outcomes. The approaches could focus on building responsible search systems or interfaces that aim to make the search systems deliver unbiased results to users regardless of their personal attributes and search contexts. We are also interested in developing interfaces that provide seamless access to a range of underlying search systems, such as traditional search or AI-assisted search.
Understanding How Search Interfaces Drive Query Variation: Users’ personal preferences influence the query formulation process and can subsequently impact the diversity and quality of search results they obtain. However, traditional search interfaces have diversified, and users increasingly seek information using a range of devices and interfaces, such as conversational interfaces (e.g. voice search using Google Assistant, Generative AI-powered Bing search), which may impact how they formulate queries. We argue that a better understanding of the impact of this additional contextual information is needed to study query variation and isolate the impact of personal/demographic factors. Proposed approaches could include conducting observational and longitudinal data collection studies.
Collecting and Analysing User Behaviour: One viable approach is crowdsourcing search history data for a specific period, which includes search queries and subsequent visits to specific websites (which indicates the original information need). Alternatively, we could conduct lab-based or in-the-wild user studies, where we provide users with a prototype search implementation that works on a range of interfaces such as smartphones, smart speakers, wearable devices or mixed-reality devices. We could use diary studies or experience sampling method to complement the limited understanding we obtain from analysing interactions.
Understanding the impact of diverse search queries on public interest knowledge platforms: Given the importance of search for critical social, health and public interest information, how do different queries relate to the results generated and what are the implications for major information producers such as Wikimedia, news and research sources? Wikipedia plays a significant role in search results and knowledge panels, but what paths lead to Wikipedia vs alternative sources? This topic could build on existing or proposed findings from the Search Experience 2.0 project and/or generate new queries using digital ethnography, crowdsourcing, generative AI or other methods. The research could also be targeted towards specific topics, languages or interest groups.
More details
https://www.admscentre.org.au/australian-search-experience-2-0
