Description of PRISM® (Advanced Version 1.0)

Margaret F. Fels, Kelly Kissock, Michelle A. Marean, and Cathy Reynolds

A standardized tool for estimating energy savings from billing data

The Advanced Version of PRISM® (PRInceton Scorekeeping Method)* makes it easier than ever to transform run-of-the-mill billing data into statistically sound savings estimates. With PRISM, utilities and energy analysts can systematically estimate total savings from a conservation or demand-side management program, for large samples of houses or buildings participating in the program, and for comparison groups as well. Enhanced model tuning and data pruning in the new version of PRISM produce even more reliable savings estimates and expanded statistical capabilities, from the same readily available data.

Summary of Method

PRISM® is a statistical procedure that processes a year of monthly billing data from a house or building to produce a weather-adjusted Normalized Annual Consumption (NAC) index, along with other physically meaningful parameters and extensive reliability statistics. A key feature of the method is its estimation of best reference temperature to which heating and cooling degree-days in the model are computed. PRISM® is generally
Advanced Version of PRISM® run on the pre- and post-weatherization periods for all buildings in a sample to produce distributions of savings across the sample. With PRISM*, participant and control groups are easily compared, in graphical and tabular forms.


  • Heating-Only (HO) PRISM: Regular and Robust
  • Cooling-Only (CO) PRISM: Regular and Robust
  • Heating-and-Cooling (HC) PRISM: Regular

Additional Features

  • Input data translators (including procedure for treating estimated readings);
  • Statistical summaries of temperature data, and easy creation of "normalization" files;
  • Automated model selection (to match the consumption data to appropriate model);
  • Automated identification and correction of undesignated estimated readings;
  • Identification of cases with anomalous data, and automated application of Robust PRISM® to these cases;
  • Aggregate versions of HO, CO and HC PRISM, for weather adjustment of utility aggregate sales data;
  • Interactive graphics for individual-building analysis;
  • User-specified reliability criteria;
  • Standardized summary of savings results for control vs. participant groups.

The Package

The software is distributed electronically or on CD. The package includes a Users' Guide to PRISM® in two parts: 1) Getting Started, with tutorials to help those who have never used PRISM® and to teach beginning and experienced PRISM® users; and 2) Reference Manual, with a full description of the available options, including how to access them in the software and how to interpret the results. A complete bibliography listing PRISM® reports available from Princeton University is also provided.

Hardware Requirements

PRISM® runs on all 32 bit machines, 64 bit machines require the Windows option Hyper-V.

Ordering Information

 *Advanced PRISM® research and development at Princeton University have been co-funded by EPRI and four EPRI-member utilities (Houston Lighting and Power, Jersey Central Power and Light, New England Electric System, and Pacific Gas and Electric) as a Tailored Collaboration (RP3824-01), and also by the Wisconsin Center for Demand-Side Research and four additional utilities (Central Maine Power, Ontario Hydro, Portland General Electric, and Sacramento Municipal Utilities District). Studies of its usefulness to low-income weatherization program evaluations have been funded by DOE (through Oak Ridge National Laboratory) and the Bonneville Power Administration.

 *PRISM® (PRInceton Scorekeeping Method) is copyright protected. HO PRISM, Copyright 1984; HO/CO PRISM, Robust PRISM and HC PRISM, Copyright 1993; Advanced Version of PRISM, Copyright 1995, Princeton University. All rights reserved.