Home

ROSE is a software tool to observe usage practices of Facebook users for a limit period of time during an empirical field study. ROSE collects data on technical events, i.e., when a user triggers specific functions of the online social network, and enables study participants to comment in situ on events that caught their attention.

Advantages for Study Participants

ROSE has been designed to fully respect the privacy of study participants. Since ROSE is implemented as browser plugin, data collection only takes place on the personal computer of the study participant. Thus, ROSE differs from other research tools implemented, e.g., as Facebook applications that collect data in the background on a server controlled by the researchers.

ROSE informs the study participant about all data collected. Each participant can decide which data she wants to transmit to the researchers; no data transfer takes place without the explicit consent of the participant.

ROSE does not collect any content shared with other users of the online social network including but not limited to chat messages, images, wall postings or targets of like or unlike actions.

For each shared data object, ROSE creates an identifier with very limited uniqueness. This ROSE identifiers enable researchers, on the one hand, to understand the relationships between user actions without knowing the shared content. On the other hand, identifiers can not be mapped back to Facebook data objects with reasonable precision or computing effort, so participants' privacy is preserved.

Because ROSE is open source, the software's code can be audited by any interested party.

The german manual for study participants can be found here

Advantages for Researchers

ROSE builds upon the Kango cross-browser framework so that it can be compiled for different web browsers. We provide builds for Google Chrome, Mozilla Firefox and Opera.

Through it's privacy preserving design ROSE aims on minimizing interferences with the field to increase the validity of collected data. On the other hand, ROSE monitors occurrences of a couple of events and actions distinctive for the usage practices of online social network users. ROSE identifiers enable researchers to assign users' in-situ comments to events and actions, for single participants and across all study participants, without knowing the actual content the users have shared.

ROSE has been developed by field researchers for field researchers. We use ROSE in our own studies and continuously improve the software. We currently prepare a tool, to analyze data collected by ROSE. However, since ROSE uses common text based data formats (JSON or YAML), researchers can easily build there own analysis workflow.

Advantages for Developers

ROSE's software architecture is split into two major parts:

code to collect data on actions and events in Facebook that occur in participant's browser, and code to generate the user interface enabling study participants to check collected data and to transmit it to the researchers. The data collection code is designed to be adoptable to changes in Facebook's user interface. However, it yet sticks to structural features of Facebook's user interface that are unlikely to change. The code can be easily extended to monitor other online social networks if this is required for a study.

The user interface code can be completely changed to fit other purposes, e.g., privacy wizards or monitors, or prototypes of newly developed privacy features.

Wiki Structure

This Wiki has three major topics:

ROSE user manual describes to study participants how ROSE works ROSE researchers manual contains the information necessary to use ROSE in empirical field studies Ivy manual describes the software to analyze collected ROSE data (Note that this software is currently under development and not released yet) Please choose a wiki page from the table of contents shown at the left sidebar.