Today in Energy

February 7, 2017

EIA’s residential and commercial studies require significant data collection and analysis

timeline of residential energy consumption survey, as explained in the article text
Source: U.S. Energy Information Administration

EIA’s two studies of energy-related characteristics and energy consumption, the Residential Energy Consumption Survey (RECS) and the Commercial Buildings Energy Consumption Survey (CBECS), require years of preparation, data collection, analysis, modeling, and dissemination for each survey cycle. Ongoing improvements to both surveys aim to increase data quality while reducing the time between data collection and release.

These efforts are already resulting in shorter timelines for data releases. Notably, RECS household characteristics data for a 2015 sample that were collected through the first half of 2016 are scheduled for release this month, to be followed with release of corresponding energy consumption and expenditure data early next year.

Survey preparation: Before the survey is conducted, EIA works to update the questionnaire and establish the survey frame, or the large pool of buildings from which samples are picked. For the residential sector, the survey frame is primarily based on an existing list of U.S. Postal Service addresses. However, for commercial buildings, EIA must construct its own survey frame because no comprehensive list exists for U.S. commercial buildings.

Data collection: Survey data are collected at the end of or in the year following the reference year. The most recent reference years are 2015 for residential housing units and 2012 for commercial buildings. The data collection process has traditionally involved dozens of interviewers conducting in-person surveys at thousands of buildings, but EIA is evaluating some alternative methods that could accelerate and streamline the process without compromising quality. Prior to the 2015 RECS, EIA tested the effectiveness of online and mail surveys and subsequently blended these alternative modes into the official 2015 RECS household data collection.

In the commercial sector, even a seemingly simple task like identifying the appropriate building to survey can be challenging. The most straightforward cases—freestanding, single-occupant buildings that are not part of a multibuilding campus—made up only 40% of the commercial buildings sample in 2012. Data collection is more complicated when buildings are connected, are part of a multibuilding campus, contain multiple tenants, or do not have a standard street address.

timeline of commercial buildings energy consumption survey, as explained in the article text
Source: U.S. Energy Information Administration

Energy bills data collection: After the building characteristic responses are received, EIA initiates a second wave of data collection that involves collecting energy consumption and expenditure data from respondents’ utilities for fuels such as electricity and natural gas and from bulk fuel suppliers for fuels such as propane and fuel oil. About half of commercial building respondents are able to provide their own energy usage and cost data. For the rest of the commercial survey respondents and all of the residential survey respondents, EIA works directly with energy providers to obtain customers’ monthly energy consumption and expenditures during the reference year.

Data processing: Both waves of responses go through stages of quality assurance, editing, and (for missing values) imputation, to arrive at a full set of data for characteristics and consumption data. EIA combines processed data with other data, such as weather information from the National Oceanic and Atmospheric Administration (NOAA), which is used to determine weather-related effects on energy consumption. For privacy reasons, statistical noise (small amounts of random variation) is added to certain data items that could be used to identify individual respondents.

End-use consumption modeling: EIA’s end-use models aim to disaggregate total consumption into end uses such as heating, cooling, lighting, and more. This task is one of the most valuable parts of RECS and CBECS, as utility meters are not able to distinguish how much energy was used for each particular use within a housing unit or commercial building.

Data releases: After all data collection and associated data quality assurance processes are complete, EIA publishes the characteristics and energy usage data. These data are first available in summary tables and later in detailed microdata files. Data for the 2012 CBECS were published between 2014 and 2016. EIA expects to begin publishing RECS characteristics data in February 2017 and energy usage estimates in early 2018.

More information about the survey methodologies and improvements in data collection is available on EIA’s websites for the Residential Energy Consumption Survey and the Commercial Buildings Energy Consumption Survey.

Principal contributors: Chip Berry, Joelle Michaels