Sidebar
Table of Contents
Governing Policy
Questions?
Please use the contact section in the governing policy.
This procedure provides guidance on how and when University public or private unit record or summary data can be shared with audiences external to the University. This procedure applies to external requests that go through the Data Request Center or directly to individuals and units (e.g. central units, colleges, departments) across the University campuses for any purpose outside of internal University business.
External audiences can be provided with University data from a variety of platforms (e.g. official reports from the Offices of Institutional Research, ad hoc reports created by central units, colleges and departments) either electronically or in printed form. When sharing data with external audiences, providers should limit the breadth of the information to meet only the scope of the request, and not provide any additional information.
Individuals or units providing data in any form are responsible for the application of this procedure and its related policy (see Administrative Policy: Public Access to University Information).
Definitions
- Data Security Classification
- A simple and high level means of identifying the level of security and privacy protection to be applied to a Data Type or Data Set and the scope in which it can be shared. For a list of public and private data elements see the appendix: Examples of Public, Private and Confidential Information provided through Administrative Policy: Public Access to University Information. University of Minnesota classifications are listed in the Data Security Classification policy.
- Unit Record Data
- Non-aggregated data at the lowest level of detail (e.g., individual student or employee level data)
- Data Suppression
- The rules that should be applied when presenting summary data or considering whether a data set should be de-identified; applied if any of the data provided are classified as private-restricted or private highly-restricted by regulation or policy.
- Rule of 10
- If an individual number in any cell of a summary table of private data elements is less than 10 it is too small to be displayed in the table if the summary is below the college/administrative unit level.
- Note: The Rule of 10 only applies to counts, not ranges or percentages
- Cell Value
- The display value of an individual element in a summary table.
Out of Scope
Private data (e.g., HIPAA, social security numbers, PCI DSS) that is classified as private-highly restricted as defined in Administrative Policy: Data Security Classification will not be shared in this manner and are out of scope for this procedure.
Those receiving requests for data from internal audiences should be directed to the following procedures:
- Sharing Data with University Educational and Administrative Audiences.
- Sharing Data with University Faculty and Researchers.
Procedural Guidelines for Sharing Data with External Audiences
When an individual receives a request for data from external audiences that is not covered in an existing data sharing agreement it is recommended to send the request through the Data Request Center. If sharing directly with external audiences, follow these steps:
Steps:
- Determine if the request is for public, private-restricted, and/or private highly-restricted data
- Apply appropriate steps based on the data security classification and level of summary (see table below)
- Communicate with the appropriate University Office (e.g., University Relations News Service, Government & Community Relations) prior to releasing any data, regardless of data classification
- Keep a record of the request and data shared
Data Classification/Summary Level | Sharing with external audiences |
---|---|
Public Data - Summary | Can be shared (subject to the procedural steps 3 & 4 above) |
Public Data - Unit Record | Can be shared (subject to the procedural steps 3 & 4 above) |
Private-Restricted Data - Summary | Can be shared with suppression rules applied (see below) |
Private-Restricted Data - Unit Record | Not to be shared: private unit record data will not be shared without prior approval from the Office of the General Counsel |
Private-Highly Restricted | Not to be shared at summary or unit record level. Direct all requests to the Office of the General Counsel |
Data Suppression
Data suppression are the rules that should be applied when presenting summary data or considering whether a data set should be de-identified if any of the data provided are classified as private-restricted or private highly-restricted by regulation or policy. When a cell value is less than 10 the data should be suppressed as a user could reasonably identify an individual, known as the “Rule of 10”.
Number of Private Elements | Summary Level: ZDept, Department | Summary Level: College/Administrative Unit, RRC, Campus, University |
---|---|---|
1 Private Element | Suppression | No Suppression** |
2 or more Private Elements | Suppression | Suppression |
**No suppression is required when the summary level is broad enough that individuals are not reasonably identifiable
Addressing Suppression Requirements:
Choosing the approach for suppression depends on the data set.
Approach | Benefits | Drawbacks |
---|---|---|
Remove data from every cell that has fewer than 10, and remove any column totals that would expose the cell number | Most straight-forward way to meet suppression requirements that does not require groupings or data manipulation | Does not provide all of the numbers requested, and may become less useful without total numbers as context |
Change from numbers to ranges/percentages for all cell values and totals | Presents information for every cell | Detailed divisions of elements are not available |
Roll-up the concept to larger cell values (e.g. individual ethnicities to Persons of Color) | Provides all of the numbers requested | Can’t do detailed analysis |
All questions about this procedure or how to apply it should be routed to Data Governance by sending an email to [email protected].