Understanding Search Behaviour On Mobile Devices

Web search on hand-held devices has become enormously common and popular. Although a number of studies have revealed how users interact with search engine result pages (SERPs) on desktop monitors, there are still only few studies related to user interaction in mobile web search, and search results are shown in a similar way whether on a mobile phone or a desktop. Therefore, it is still difficult to know what happens between users and SERPs while searching on small screens, and this means that the current presentation of SERPs on mobile devices may not be the best. According to the findings from previous studies, including our earlier work, we can confirm that search behaviour on touch-enabled mobile devices is different from behaviour with desktop screens, and so we need to consider a different SERP presentation design for mobile devices. In this thesis, we explore several user interactions during search with the aim of improving search experience on smartphones. First, one remarkable trend of mobile devices is their enlargement of screen sizes during the last few years. This leads us to look for differences in search behaviour on different sized small screens, and if there are any, to suggest better presentation of search results for each screen size. In the first study, we investigated search performance, behaviour, and user satisfaction on three small screens (3.6 inches for early smartphones, 4.7 inches for recent smart-phones and 5.5 inches for phablets). We found no significant differences with respect to the efficiency of carrying out tasks. However, participants exhibited different search behaviours on the small, medium, and large sizes of small screens, respectively: a higher chance of scrolling with the worst user satisfaction on the smallest screen; fast information extraction with some hesitation before selecting a link on the medium screen; and less eye movements on top links on the largest screen. These results suggest that the presentation of web search results for each screen size needs to take into account differences in search behaviour. Second, although people are familiar with turning pages horizontally while reading books, vertical scrolling is the standard option that people have available while searching on mobile devices. So following a suggestion from the first study, in the second study we explored the effect of horizontal and vertical viewport control types (pagination versus scrolling) with various positions of a correct answer in mobile web search. Our findings suggest that although users are more familiar with scrolling, participants spent less time to find the correct answer with pagination, especially when the relevant result is located beyond the page fold. In addition, participants using scrolling exhibited less interest in lower-ranked results even if the documents were relevant. The overall result indicates that it is worthwhile providing different viewport controls for better search experiences in mobile web search. Third, snippets occupy the biggest space in each search result. Results from a previous study suggested that snippet length affects search performance on a desktop monitor. Due to the smaller screen, the effect seems to be much larger on smartphones. As one possible idea for a SERP presentation design from the first study, we investigated appropriate snippet lengths on mobile devices in the third study. We compared search behaviour with three different snippet lengths, that is, one line, two to three lines, and six or more lines of snippets on mobile SERPs. We found that with long snippets, participants needed longer search time for a particular task type, and the longer time consumption provided no better search accuracy. Our findings suggest that this search performance is related to viewport movements and user attention. We expect that our proposed approaches provide ways to understand mobile web search behaviour, and that the findings can be applied to a wide range of research areas such as human-computer integration, information retrieval, and even social science for a better presentation design of SERP on mobile devices.