In 2013, concerns of mass-surveillance capabilities deployed in the interest of security received unprecedented attention across the globe. Ongoing media coverage sparked public debate about the pros and cons of such capabilities. This debate is characterised by lack of robust empirical evidence on the extent to which individuals are in favour or against surveillance in the name of security. Each argument 'against' or 'in favour' of relevant security interventions and technologies is based on sources like key/expert opinions, public opinion-polls and scholarly opinion.This paper reports on insights into citizens' preferences regarding privacy, security and surveillance derived from an innovative large-scale survey. In particular, we explore the apparent dichotomy between how people respond to abstract questions about the importance of privacy and security compared to their preferences in real world contexts where individuals appear to be exhibit a greater degree of comfort with interventions that may adversely affect their privacy.Our research elaborates upon this ‘privacy paradox’ through stated preference discrete choice experiments which enable respondents to express preferences over security, privacy and surveillance within real-world contexts, namely Internet use, travel on train or underground and purchase of a personal health-record storage device/system. Respondents are presented with scenarios described in terms of context dependent features relating to privacy and security implications e.g. the breadth of CCTV-camera information sharing, or the stringency of oversight regarding internet surveillance by the authorities.Systematic combinations - based on principles of experimental design of these features - create each of the options. Respondents are then asked to choose among three options presented across variations of these three different contexts with associated monetary implication. Alongside these data we gather socio-economic and demographic information and attitudinal statements relating to trust, privacy and security.Data was collected across a sample of 26,800 participants across 27 EU through a mix of face-to-face and online methods and was broadly representative of the population within each country.Choices from respondents were subjected to discrete choice analysis. This framework enables us to determine which combination of features comprising each of the alternative options drives (in terms of utility lost/gained) the choices of participants. In order to account for a range of perspectives which may colour these preferences, the analysis takes into account respondents' characteristics and their level of familiarity with each particular context, their attitudes to privacy, security and surveillance and their appetite for risk-taking behaviour in general.Analysis enables us to bring new insights into the passionate debate around privacy and surveillance. It illustrates a number of new insights, not previously explored using quantitative approaches including measurement of the extent people may balance privacy or security. We find that individuals derive markedly different utility for privacy versus security in different contexts, supporting the view that privacy and security are highly context specific. For example, all else being equal, respondents appear to prefer measures to improve online privacy compared to that derived from stand-off surveillance common in transportation. We observe that data sharing is governed by complex trust relationships. We also find that certain socio-demographic characteristics are correlated with people’s preferences.Overall, preference regarding privacy, security and surveillance are context specific. For example, in some cases privacy is considered as the form of control of personal information and others as control of physical space. Our research suggests that security measures and relevant decision-making needs to be matched to different settings in order to provide the appropriate mix of protections and meaningful choice.