The “how it started, how it’s going” meme started on social media in 2020 (according to information on the internet), the same year that the Office of the State Superintendent of Education (OSSE) adopted its Student Privacy and Data Suppression Policy. Coincidence? Probably. But five years later seems like an opportunity to put that meme to good use.

What is suppression? In education data, “suppression” refers to methods used to protect personally identifiable information (PII) about children and students, part of OSSE’s responsibilities under the Family Educational Rights and Privacy Act (FERPA). This can also be called “disclosure avoidance.”

In action, this means that when releasing aggregate data (such as at the state, local education agency [LEA], school or child development facility level [CDF]), OSSE obscures certain information to prevent the identification of individual children and students. Typically, this is reflected in notations (such as n<10, data suppressed [DS], <5%, >=90%) in OSSE data sets.

How it Started

All aggregate data releases – whether at the state, LEA, school or CDF level – carry some risk of accidental exposure of individual identities. In other words, no suppression method can completely and perfectly eliminate all risk.

But with a focus on continuous improvement, OSSE identified the need for and developed a policy that would:

• Enhance the balance between transparency to the public and OSSE’s legal responsibility to protect student privacy;
• Improve equity in showing growth, even small growth, in LEAs, schools and student groups;
• Improve consistency in how OSSE suppresses data across projects and releases;
• Increase the efficiency and speed of producing and reviewing data through process standardization; and
• Improve OSSE's ability to analyze and share data and explain data and data practices to stakeholders and the public.

Specifically, the policy:

• Reports denominators of fewer than 10 children or students (and any corresponding percentages) as n<10;
• Requires “blurred” or inexact percentages (such as <=10% and >=90%) for very high and very low percentages, based on the denominator;
• Implements complementary suppression (noted as “DS”) to mask data that would be able to be calculated by simple math using other values in the data set;
• Prohibits the publication of values of 0, 0% or 100%, which would reveal data about all children and students at a school, LEA, or CDF; and
• Recognizes exceptions for low-risk data, such as enrollment data by race, ethnicity, gender, and grade.

How it’s going

In looking at implementation over the last five years, it’s clear that OSSE’s suppression policy has met many of its main goals, including:

• Improving child and student privacy in OSSE aggregate data releases;
• Aligning expectations across the agency about how OSSE suppresses agency data sets released to the public;
• Increasing process standardization, which has improved the efficiency of producing and reviewing data sets for public release; and
• Strengthening documentation of OSSE data practices and data governance and privacy at OSSE.

It also sparked the development of a companion policy that applies to the release of data about the educational workforce.

Other learnings include:

• While mostly a technical practice, suppression is also part art, relying on human judgment, especially when weighing the risks of specific data releases.
• A policy needs continuous care and feeding; no policy can anticipate all of the use cases (or thorny edge cases) that require additional analysis.

For More Information

For more information, email [email protected] or visit Student Privacy and Data Suppression Policy At A Glance or Data Governance and Privacy at OSSE.