Privacy Surveys Privacy Surveys Week 12 - April 6, 8 1 Privacy Policy, Law and Technology • Carnegie Mellon University • Spring 2004 • Lorrie Cranor • http://lorrie.cranor.org/courses/sp04/
Current events Current events n Gmail ´ Good, bad, or ugly… what do you think? ´ Ads are customized based on content of email, but no human ever reads the email … are privacy concerns satisfied? 2 Privacy Policy, Law and Technology • Carnegie Mellon University • Spring 2004 • Lorrie Cranor • http://lorrie.cranor.org/courses/sp04/
GPS GPS n Clarifying a misconception… n GPS is a one-way system ´ GPS receivers listen for radio beacons and triangulate their position ´ If receivers are to report their location back they must use another system, for example cellular phone network n GPS does not work indoors 3 Privacy Policy, Law and Technology • Carnegie Mellon University • Spring 2004 • Lorrie Cranor • http://lorrie.cranor.org/courses/sp04/
Privacy Surveys - Questions to Ask Privacy Surveys - Questions to Ask n Who conducted the survey and why? n What population was sampled and how? ´ What is this sample representative of? n When was this study done? ´ Are there more recent studies? ´ Are results still applicable today? n How was this study done? n How are the results being spun? 4 Privacy Policy, Law and Technology • Carnegie Mellon University • Spring 2004 • Lorrie Cranor • http://lorrie.cranor.org/courses/sp04/
The role of surveys in public policy The role of surveys in public policy n Survey respondents may be ´ Misinformed ´ Confused ´ Not randomly sampled ´ Non-expert ´ Unaware of alternatives, impacts, etc… ´ Biased n But then again, so might other sources of information that feed into policy process 5 Privacy Policy, Law and Technology • Carnegie Mellon University • Spring 2004 • Lorrie Cranor • http://lorrie.cranor.org/courses/sp04/
Beyond Concern Beyond Concern n Study by Cranor, Ackerman, and Reagle, 1999 ´ http://www.research.att.com/projects/ privacystudy/ n Motivation: Design better P3P user agent ´ What data elements are users most sensitive about? ´ What aspects of privacy policies do users care most about? ´ Would users like data to be transferred automatically? 6 Privacy Policy, Law and Technology • Carnegie Mellon University • Spring 2004 • Lorrie Cranor • http://lorrie.cranor.org/courses/sp04/
Privacy surveys Privacy surveys n Many surveys show high levels of concern about privacy, BUT “Despite this wide range of interests in privacy as a topic, we have little idea of the ways in which people in their ordinary lives conceive of privacy and their reactions to the collection and use of personal information” (Hine and Eve 1998). 7 Privacy Policy, Law and Technology • Carnegie Mellon University • Spring 2004 • Lorrie Cranor • http://lorrie.cranor.org/courses/sp04/
Design for a new survey Design for a new survey n How people respond to situations in which personal info is collected online n Sensitivity to particular privacy practices n General attitudes and demographics 8 Privacy Policy, Law and Technology • Carnegie Mellon University • Spring 2004 • Lorrie Cranor • http://lorrie.cranor.org/courses/sp04/
Methodology & sample Methodology & sample n Invitations to complete a Web-based survey emailed to 1,500 DRI Family Panel members n 523 surveys completed - 381 US respondents n Our US sample differed from nationally representative sample: ´ More educated ´ More Internet experience ´ More concerned about Internet privacy issues 9 Privacy Policy, Law and Technology • Carnegie Mellon University • Spring 2004 • Lorrie Cranor • http://lorrie.cranor.org/courses/sp04/
Clusters Clusters n Privacy fundamentalists (17%) ´ extremely concerned about any use of data n Pragmatic majority (56%) ´ often had specific concerns and tactics for addressing them n Marginally concerned (27%) ´ willing to provide data under most circumstances, but had mild general concern and some specific concerns 10 Privacy Policy, Law and Technology • Carnegie Mellon University • Spring 2004 • Lorrie Cranor • http://lorrie.cranor.org/courses/sp04/
Users prefer to be anonymous Users prefer to be anonymous n In two scenarios, respondents were more likely to provide information when they were not identified 11 Privacy Policy, Law and Technology • Carnegie Mellon University • Spring 2004 • Lorrie Cranor • http://lorrie.cranor.org/courses/sp04/
Data sensitivity varies Data sensitivity varies 0% Social security # 1% Credit card # N/A 3% Comfort level 1% Phone number for child 11% Comfort level Income N/A 17% for self Medical info 4% 18% 6% Postal address 44% 9% Full name 54% Computer info 28% 63% 14% Age 69% 16% Email address 76% Favorite snack 50% 80% 52% Favorite TV show 82% 0% 20% 40% 60% 80% 100% Respondents who are always or usually comfortable providing information 12 Privacy Policy, Law and Technology • Carnegie Mellon University • Spring 2004 • Lorrie Cranor • http://lorrie.cranor.org/courses/sp04/
Many factors important Many factors important 68% sharing of information 82% 95% 59% identifiable use 77% 97% 68% purpose of information collection 73% 95% 76% mailing list removal upon request 71% 87% 59% kind of information 70% 92% 51% access to stored information 65% Marginally 93% concerned 52% site run by trusted company 62% 80% Pragmatic 35% privacy policy 50% majority 72% 24% privacy seal of approval 39% Privacy 69% fundamentalis 22% disclosure of data retention policy 29% ts 68% 0% 20% 40% 60% 80% 100% Respondents who consider factor very important 13 Privacy Policy, Law and Technology • Carnegie Mellon University • Spring 2004 • Lorrie Cranor • http://lorrie.cranor.org/courses/sp04/
Acceptance of persistent IDs Acceptance of persistent IDs 96% Marginally 100% 90% concerned 80% 77% 80% 60% 55% Pragmatic 43% majority 37% 40% 27% 20% 14% Privacy fundamentalists 0% for customized for customized for customized service advertising advertising across many Web sites Respondents who would probably or definitely agree to a site assigning persistent identifier 14 Privacy Policy, Law and Technology • Carnegie Mellon University • Spring 2004 • Lorrie Cranor • http://lorrie.cranor.org/courses/sp04/
Dislike of auto data transfer Dislike of auto data transfer 70% 61% 60% 51% 50% 39% 40% 30% 20% 14% 6% 10% 0% "Auto-fill" button Automatic "auto- Automatic Automatic data Automatic data for filling in form fill" - no button transfer of data transfer to sites transfer to sites fields - user click required to already provided with acceptable with acceptable intervention fill form, user back to site on privacy policies - privacy policies required to submit intervention return visit accompanied by form required to submit notice to user form Respondents who would use proposed browser features 15 Privacy Policy, Law and Technology • Carnegie Mellon University • Spring 2004 • Lorrie Cranor • http://lorrie.cranor.org/courses/sp04/
Technical implications Technical implications n Extremely simplified interfaces might work for some users but most will need more sophisticated interfaces n Automatic data transfer unlikely to be of interest n Need for different views of information (treat phone number different from postal address) 16 Privacy Policy, Law and Technology • Carnegie Mellon University • Spring 2004 • Lorrie Cranor • http://lorrie.cranor.org/courses/sp04/
Policy & business implications Policy & business implications n Need flexibility n Need trust-enhancement approach n Expressed concerns can help policy- makers prioritize n BUT don’t make policy based solely on survey results 17 Privacy Policy, Law and Technology • Carnegie Mellon University • Spring 2004 • Lorrie Cranor • http://lorrie.cranor.org/courses/sp04/
Privacy Bird User Survey Privacy Bird User Survey n Cranor, Guduru, Arjula 2002 ´ http://lorrie.cranor.org/pubs/wpes02/ n About 20,000 downloads in first six months of public beta trial n Users asked whether they were willing to participate in survey when they downloaded software n We randomly selected 2000 email addresses from those willing to participate in surveys and sent invitation to fill out online 35-question questionnaire n 17% response rate 18 Privacy Policy, Law and Technology • Carnegie Mellon University • Spring 2004 • Lorrie Cranor • http://lorrie.cranor.org/courses/sp04/
Demographics and Internet use Demographics and Internet use n Compared to random sample surveys of Internet users, our sample was older, more predominantly male, better educated, and had more Internet experience n Most of our respondents from English speaking countries – 70% from US, 14% from Australia, 6% from Canada n US respondents had more Internet experience than other respondents and were more likely to have made purchases from web sites n Are our skewed survey respondent demographics representative of Privacy Bird users? n Are our demographics similar to demographics of users of other privacy software? 19 Privacy Policy, Law and Technology • Carnegie Mellon University • Spring 2004 • Lorrie Cranor • http://lorrie.cranor.org/courses/sp04/
Recommend
More recommend
