Monday, 1:30pm


Looting Hoards of Gold and Poaching Spotted Owls: Data Confidentiality Among Archaeologists & Zoologists

Rebecca D. Frank¹, Adam Kriesberg¹, Elizabeth Yakel¹, Ixchel Faniel²
¹University of Michigan, United States of America; ²OCLC Research, United States of America


Researchers in the social and health sciences are used to dealing with confidential data, and repositories in these areas have developed mechanisms to prevent unethical or illegal disclosure of this data. However, other scientific communities also collect data whose disclosure may cause harm to communities, cultures, or the environment. This paper presents results from 62 interviews and observations with archaeologists and zoologists. It focuses on how researchers’ perceptions of potential harm influence attitudes about data confidentiality, and how these, in turn, influence opinions about who should be responsible for managing access to data. This is particularly problematic in archaeology when harm is not to a living individual but is targeted at a community or culture that may or may not have living representatives, and in zoology when an environment or a species may be at risk. We find that while both archaeologists and zoologists view location information as highly important and valuable in facilitating use and reuse of data, they also acknowledge that location should at times be considered confidential information since it can be used to facilitate the destruction of cultural property through looting or decimation of endangered species through poaching. While researchers in both disciplines understand the potential dangers of allowing disclosure of this information, they disagree about who should take responsibility for access decisions and conditions.

Private: Information Quality, Privacy Policies, and Data Safety Practices in Online Social Networks for Health: A Longitudinal Analysis

Kaitlin Light Costello
School of Communication & Information Rutgers, the State University of New Jersey, United States of America


This paper describes a longitudinal analysis of online social networks designed for patients diagnosed with chronic kidney disease. Observational analysis of 20 indicators in three domains – practices for auditing and moderating the quality of content provided by users, accessibility of privacy policies, and data sharing policies and member control over data sharing – was conducted on 10 sites in 2013 and again on 12 sites in 2015, with 7 of the same sites included in both samples. The sites were identified using Google. Indicators for each domain were scored dichotomously. These scores were compared among sites in order to analyze their general practices and policies. Total composite scores were also analyzed to determine whether individual sites had significantly different practices and policies in comparison with the group. Finally, scores for each domain were compared across years in order to assess whether practices and policies had changed over time.

Differences in site practices and policies between 2013 and 2015 were not significant, although there is much room for improvement in all domains. Quality was variable across all sites, with gaps in medical disclaimers, a lack of external review of privacy policies and data safety audits, and missing information about internal quality control in the form of moderators. Although most sites employ moderators, their credentials are not often reported. Privacy policies are inaccessible across the board, with none readable at below a twelfth-grade level. Data safety practices are also problematic, with most sites sharing user data with third parties.

The quality and safety of social networks for CKD is variable, and improvement is feasible. Suggested improvements include auditing privacy policies and data safety practices, making information about moderators easily available, and third-party audits of information posted by users to ensure the removal of misinformation.

Password Creation Strategies Across High- and Low-Literacy Web Users

Caitlin Rinn¹, Kathryn Summers¹, Emily Rhodes¹, Joel Virothaisakun¹, Dana Chisnell²
¹University of Baltimore, United States of America; ²Center for Civic Design, United States of America


Research has shown that password security practices typically conflict with general usability principles. Though the challenges faced by low-literacy users when creating and managing passwords are likely to extend beyond those experienced by the general public, little research has been done to explore password usability in this at-risk group. This survey of 20 low-literacy participants aims to examine password behaviors within this population, including password creation, recall strategies, and perceptions of password strength and security. It expands on the work of Chisnell and Newby (2015) based on a nationwide survey exploring password use, password creation strategies, and perceptions regarding password security. Thus, this study allows for comparison between password use and perceptions among the broader population and those of users with low literacy skills.