This paper presents a user survey-based analysis of the correlation between the users' willingness to personalize Web search and their social network usage patterns. The participants' responses to the survey questions enabled us to use a regression model for identifying the relationship between SNS variables and willingness to personalize Web search; the obtained results show that there is a strong relationship between willingness for personalized Web search and social network usage patterns. Finally, based on the ndings of our survey we present some implications and directions for future work.
For over a decade Web search personalization has been viewed as an e ective
solution for user-tailored information-seeking and its e ectiveness depends on
the quality of user-related information, which is generally implicitly gathered.
The implicit collection of user information however often raises serious privacy
concerns, and it represents one of the major barriers in deploying personalized
search applications [
To the aim of nding a characterisation of users who would prefer personalized search results more than non-personalized search results we conducted a user survey. We designed the user survey so that we can investigate the social network usage patterns of users along with their preferences to opt for personalized Web search. The collected data were then used in a regression model to analyse the correlation between SNS variables and willingness to personalize Web search. In doing this analysis we discovered a number of useful patterns and behaviours about users' openness to Web search personalization; these outcomes can be of help in the future developments of personalized Web search and social search systems. 2
For over a decade Web search personalization has been viewed as an e ective
solution for user-tailored information-seeking. A huge amount of research
proposes the use of implicitly-gathered user information such as browser history,
query and clickthrough history and desktop history to improve search results
ranking on a per-user basis [
The literature closest to our identi ed research questions addresses the use of
social information in Web search [
In order to understand how SNS usage patterns a ect people's willingness to personalize Web search results, we designed a survey and dispatched it to a wide range of people in various countries (i.e. Ireland, Italy, Spain, France, United Kingdom, Finland, United States, Canada, Pakistan, India, South Korea). The survey comprised 16 close-ended questions with ve questions of a general nature (collecting demographic information about the participants), four questions related to various aspects of Web search personalization3 and ten questions related 3 Respondents were asked through a binary variable a \Yes" or \No" response whether or not they considered personalized search results to be of any bene t to them. 49.5% of the respondents considered personalized Web search as useful while 50.5% considered it as not useful. to SNS usage4. The survey was completed by 380 people with 61.8% males and 38.2% females. Statistics about the use of various SNS tools are shown in Table 1 and Table 2. Participants were recruited via university distribution lists (both online and o ine) and social networking sites (chie y, Facebook and Twitter) and we recruited a diverse range of people; distribution lists were employed to avoid recruiting only those participants who have a social network presence and thereby avoiding high skew in the results. 4
The results from Table 3 show that males are more likely to consider Web search personalization as bene cial. Furthermore, the presence of a user on Twitter and/or Google+ is a strong indicator that he/she will consider Web search personalization as bene cial and similar is the case for his/her high usage of Twitter and/or Google+ (row 7-8 corresponding to WP and row 12-13 corresponding to WP ); the increase is more signi cant for user presence on Google+ and for high usage of Google+. Additionally, as expected an increase in posting frequency on Facebook increases the odds of willingness to personalize Web search and contrary to expectation, an increase in the number of users' friends on Facebook decreases the odds of willingness to personalize Web search (row 16 and row 18 corresponding to WP ). Lastly, users who use SNS more frequently for Q & A activities are more likely to be willing to opt for Web search personalization in addition to users who frequently consider that responses coming from SNS as more reliable than responses from search engines (row 23-26 corresponding to WP ). 5
We will now discuss the theoretical and practical implications of our study. The correlations (and their results) investigated here may seem intuitively obvious 4 These were related to various SNS tools (more speci cally Facebook, Twitter, LinkedIn etc.) as well as at the characteristics of SNS usage (such as frequency of SNS usage, frequency of posting SNS updates, number of friends on SNS, frequency of asking questions on SNS) that relate more closely to users' willingness for Web search personalization.
WP
Male 1.635** American 0.001
Asian 0.001 European 0.001
Age 1.069 Facebook Presence 0.982 Twitter Presence 1.544* Google+ Presence 1.816***
LinkedIn Presence 0.940 Bookmarking Sites Presence 1.289
High Usage of Facebook 1.599 High Usage of Twitter 1.166* High Usage of Google+ 3.042*** High Usage of LinkedIn 1.193 Facebook Usage Frequency 0.898 Facebook Posting Frequency 1.637*** Facebook Liking Frequency 0.920 No. of Facebook Friends 0.873*
Twitter Mentions 1.000
Twitter Retweets 1.001 No. of Topics in Tweets 0.997
No. of Tweets 0.999
Prefers Q & A Activity on SNS 1.821*** Considers Q & A Activity on SNS as Useful 1.771***
Frequency of Q & A Activity on SNS 1.366**
Frequency of Considering Responses from SNS More Useful than Search Engines 1.374***
Note *p<.05, **p<.01, ***p<.001
in terms of high engagement in social networks signifying a higher degree of
readiness to accept the (at least partial) loss of privacy that is inevitably
involved in search personalization. However, recent research indicates that this is
not the case as users of SNS have shown growing privacy concerns [
In this paper, we utilized a survey methodology to investigate the correlations between users' social network usage patterns and their openness to opt for Web search personalization. We believe that the ndings of our study may have significant implications for the design of future personalized search and social search applications. Proper exploitation of SNS data can lead to advancements in personalized Web search systems by providing a more acceptable method for derivation of user interests.