When is a Tree Really a Truck? Exploring Mental Models of



Justin Wu and Daniel Zappala, When is a Tree Really a Truck? Exploring Mental Models of Encryption, USENIX Symposium on Usable Privacy and Security (SOUPS), August 2018.


Mental models are a driving force in the way users interact with systems, and thus have important implications for design. This is especially true for encryption because the cost of mistakes can be disastrous. Nevertheless, until now, mental models of encryption have only been tangentially explored as part of more broadly focused studies. In this work, we present the first directed effort at exploring user perceptions of encryption: both mental models of what encryption is and how it works as well as views on its role in everyday life. We performed 19 semi-structured phone interviews with participants across the United States, using both standard interview techniques and a diagramming exercise where participants visually demonstrated their perception of the encryption process. We identified four mental models of encryption which, though varying in detail and complexity, ultimately reduce to a functional abstraction of restrictive access control and naturally coincide with a model of symmetric encryption. Additionally, we find the impersonal use of encryption to be an important part of participants' models of security, with a widespread belief that encryption is frequently employed by service providers to encrypt data at rest. In contrast, the personal use of encryption is viewed as reserved for illicit or immoral activity, or for the paranoid.


Our paper includes an appendix with the semi-structured interview guide.

The following are the transcripts from each of the 19 participants:

The following are the diagrams from each of the 19 participants: