How NPWR is Built to Keep This Promise

NPWR is designed to balance privacy with the ability to discover insights about Nevada's education, higher education, and workforce policies, initiatives, and programs. NPWR is built on a foundation of multi-layered security. Every aspect of the NPWR system is  designed to ensure the complete anonymity of all Nevadans and the privacy and security of all data within the system.

The key organizational elements that ensure privacy include:

• NPWR does NOT collect data from citizens.

  It is a system that was developed among its participating agencies to allow the merging of data in a highly controlled environment  using technology that strips "exposure data" (i.e., any information that identifies an individual) before merged  information is released to researchers (see below for more details on the technology).

• NPWR is a "hybrid federated" model for longitudinal data.

  Rather than build an entirely new "data warehouse" to collect and store data, which would have required the duplication of private data (and redundant infrastructure to support and protect that data), Nevada chose to implement a system that leaves its  data where it has always been, secure within the participating agencies' databases.

• One cannot simply walk up (login) to NPWR to initiate a data merge.

  Data merges are initiated by vetted researchers who first have completed an application process and whose research questions have  been reviewed and validated. Then they are assigned a committee of agency "sponsors" who guide and oversee the process - all in  the name of accuracy AND privacy. Each step along the way, from access request to publication of results, must be approved by the  sponsoring agency.

• The NPWR Book of Data Governance requires regular meetings, testing, and assessments of NPWR to include ensuring privacy.

How NPWR Merges Data

If exposure data is protected behind privacy firewalls of state agencies, how is useful data merged? The State of Nevada's solution to this  question was developed by the Nevada Department of Education, the Nevada System of Higher Education, and the Nevada Department of Employment,  Training and Rehabilitation.

Using a state of the art probabilistic matching system, NPWR matches data across agencies through a process that de-identifies and cleanses the  data to provide the highest possible match rate while maintaining full privacy. All data within NPWR are fully de-identified to ensure that  no personally identifiable information ever remains within the system. Each time a researcher requests data, the system will generate a completely  new set of unique identifiers for each individual in the data. This feature, developed to comply with state law, ensures that the new data set  cannot be linked to the previously requested data set based on unique identifiers.

Key Privacy Protections are as follows:

• Before agency data sets are delivered to the data hub for merging, an algorithm is applied to the data to render key, private information  (e.g., name and date of birth) into a string of meaningless numbers and letters.

• The newly created data sets expire and are destroyed at the end of two weeks, after the initial database inquiry, effectively making the data unique for every merge.

• It is important to note that "de-identification" is a "one-way" process that is different than encryption, which is "two-way" scrambling, meaning  the data can move back and forth from encrypted to readable data. In simple terms, once de-identified the data cannot be reversed back to identified  data.

• In the application review process and in the algorithm and Data Hub processes, there are minimum data threshold requirements to ensure that NPWR  does not create data sets that could be used, through a process of elimination, to match de-identified data back to individuals or groups of people.

• The merged data is then reviewed by agency staff for both accuracy and privacy before being released to researchers.

Any data sets that are then to be released to the public are scrutinized again for privacy concerns before release.