Human Subjects Protection

Recommendations for State Public Utility Commissions to Assess the
Sensitivity of Tabular Data Revealing Identifiable Energy Consumption Information

The American Statistical Association’s Privacy & Confidentiality Committee and Energy Subcommittee drafted the following Recommendations for State Public Utility Commissions to assess the sensitivity of tabular data revealing identifiable energy consumption information.

For selecting an appropriate sensitivity rule for assessing the risk of disclosing identifiable information:

• The (15, 15) rule currently used by public utility commissions is overly restrictive.This current specification requires at least 15 respondents in a table cell and no single respondent may constitute more than 15% of the total cell value.
• The p-percent rule is a reasonable sensitivity rule to apply to tabular data.

For selecting an appropriate parameter value in a sensitivity rule:

• Setting the parameter value depends on the skewness of the population characteristics and special combinations of characteristics unique to a given table;
• Parameter values should vary depending upon the customer class;
• Consider the use and purpose of the local government requesters in applying sensitivity rules; and
• Consider different aggregation threshold levels for applying the Threshold rule for standardized reports to local governments based on the customer class.

As a performance measure to evaluate sensitivity rules, before and after suppression, review:

• Distributions and population counts in each customer class for the utilities’ entire state service territory; and
• The moments of the multivariate distribution as a performance measure to evaluate sensitivity rules.

Consider developing restricted access programs for local governments to access sensitive data:

• Data access should be provided to individuals employed by city and state government agencies for purposes that are designed to serve the general public and contribute significantly to understanding of energy consumption in the residential, commercial, and industrial sectors.

State Public Utility Commissions should consider developing strong staff expertise in statistical methodology that is relevant to their mission: 

• State Public Utility Commissions should stay informed and use modern statistical theory and sound statistical practices in developing privacy rules that apply to public utility companies.

Human Subjects Protection, Ethical Research, and Institutional Review Boards (IRBs)

Protecting privacy and confidentiality is one component of a larger mandate to individuals who collect data: protect the well-being of data subjects. This page provides links to resources on protecting human subjects from harm, on ethical guidenlines for research, and on the functioning of IRBs, which are committees based at research institutions that are responsible for the protection of human subjects in research. These are vast and important topics; we only introduce them here. For more information, see the Committee on Professional Ethics of the American Statistical Association.

For researchers planning to collect data, we advise you to work with the IRB of your institution to ensure compliance with protocols for protection of study subjects. Additionally, federal agencies can have different regulations for protecting human subjects, so that researchers should consult the relevant agency to ensure compliance.