Because this study involves understanding social networks, UC Berkeley has agreed to share with the research team four months of “de-identified” email metadata (i.e., records without names or email addresses of who sent a message to whom and when) and some data about our staff composition (e.g., campus department and demographic information such as race/gender) for eligible staff (see criteria above) who do not opt out of Phase 1.
The “de-identified” email metadata will include messages sent to and from all full-time, non-academic staff on campus over the four-month period leading up to the professional development experience that is a core component of this study. Email addresses will be hashed (transformed into an unrecognizable code), and only hashed to and from email addresses, and the message timestamp will be shared with the researchers. Any other message data such as subject line, the content of the email, and any attachments will not be accessed by anyone and will not be part of the metadata at any time. Following the professional development experience, the same data set for a new time period will be shared with the research team for the four-month period following the professional development experience. This will enable researchers to understand how participation in the professional development experience might have influenced participants’ social networks.
The figure below illustrates how Berkeley IT will provide hashed email metadata, anonymized before sharing with the research team for further analysis. As you will see in the figure below, researchers will not have access to any message content, email addresses, or identifying information about campus staff. Neither the researchers nor Berkeley IT staff will have access to staff members’ actual email messages as part of this procedure.A similar process will be followed by People & Culture staff to de-identify the HR data that will be used for this study: demographic information such as gender and race, as well as department affiliation. To ensure that these data are not inadvertently identifiable given that some departments are small and have only a few people in a given demographic category, the department affiliation field will be removed for any department in which there are fewer than 10 individuals in a demographic “cell.” For example, if there are fewer than 10 Asian females in a department, no department information will be provided for these individuals.