Delivering Quality Data to Campus

Many institutions have created three- to five-tier data sensitivity frameworks. Rather than assign a blanket privacy level to all the data from an entire system (e.g., the learning management system), data governance committees typically assign a sensitivity tier to each term—either during the definition process, or in batches at a later date. This determination later helps data stewards respond to data access requests and is vital to any automated role-based access scheme.

View an Example Framework

Role-based access at the University of Washington is used for human resources, financial, student, and research data. The data governance committee created a privilege level map for each of these data domains, keying categories of data elements (or individual data elements) to four potential privilege levels—baseline, expanded, high, or full. The data governance working committee then mapped each of the 14 roles to a privilege level for each of the four business domains.

See the Privilege Level Map

When a user runs a report, the University of Washington’s security administration system checks the platform for each column to determine whether the particular user may view all column-level data or has more restrictive access rights. If the latter is the case, certain columns will be unpopulated in the pulled report.

Users thus have different visibility into native reports based on their assigned access role (and, ultimately, the relevancy of different data to their job duties).

Read More

Understanding that automated access may not always fulfil users’ needs, the University of Washington’s online data request form is simple and user-friendly. Users review the data access role options and compare them to their data needs to identify the role most appropriate for their specific data requirements. Users also complete a free text field to provide supplemental information about how they intend to use data to which they would gain access.

Learn More

We’ve provided guidelines for improving unit-level data quality. Use the checklist to keep track of the steps you’ve completed, from identifying data elements that have recurring quality issues to creating accountability mechanisms to ensure high data quality over time.

Get the Checklist

There are many data quality challenges that occur in student information systems, HR systems, finance systems, and even across all college and university systems. We created a chart of common data quality challenges, their definitions, and their prevalence in higher ed institutions.

See the Challenges

An automated data quality testing system lacks utility unless functional staff are involved in the creation of logic as well as in error correction within source systems.

Once Institutional Research staff identify fields with recurring data quality problems, they communicate with functional staff to isolate information necessary to properly code an error check for that issue. IT staff then write a SQL statement that is incorporated into nightly quality checks run during the process of transferring data into the institutional data warehouse or data hub.

Daily emails containing error report information are automatically sent to appropriate end users (e.g., human resources for HR data problems) who can fix the identified problems in the source systems.

Read More

To begin tracking longitudinal data, and move towards unit-level accountability for persistent data quality issues, governance groups should partner with another office (e.g., internal audit), to set standards for data quality, and begin tracking departmental performance.

View a Scorecard

Data governance groups should leverage the data strategy committee’s authority to partner with the internal audit office to create an escalation process for units that consistently fail to meet stated data quality goals. Units whose data quality does not improve over time should be turned over to the internal audit function for a business process audit.

Learn More

The University of Washington recommends reports to users through their business intelligence portal, leveraging usage information to improve navigation of analytical resources. To do this, an algorithm identifies and recommends reports that the user has never run (or hasn’t run recently) based on other reports the user frequently views as well as reports that other campus members in a similar role frequently view.

See the Example

Once users are viewing the correct reports for their needs, the University of Washington realized that the work doesn’t end there. Even within the correct report, users often struggle to identify how to use the report well. To address this issue, they have created a virtual business intelligence user group where users can post comments about how they use the report, identifying usage opportunities for others, and providing a culture of collaboration around institutional data and problem solving.

Learn More

The analytics team at the University of Maryland University College has done targeted consolidation of their academic program-related reports. Based on the ad hoc requests received by their team, they created a master dashboard for academic leaders to review program performance.

Read More

To understand the needs of campus, and identify important new areas for investment, the data strategy committee should deploy surveys to members of the campus community. Surveys can be broad and inclusive, or targeted to specific groups and units. Questions should focus on decisionmaking culture, satisfaction regarding data availability and data services, and potential use cases and opportunities for specific data elements.

See Example Questions