Eia-112 Ssb_9.16.2024

­ ­­

Part B: Collections of Information Employing Statistical Methods
OMB No. 1905-0XXX

Form EIA-112, Residential Utility Disconnections Survey

Supporting Statement for Residential Utility Disconnections Survey

www.eia.gov

U.S. Department of Energy

Washington, DC 20585

October 2024

The U.S. Energy Information Administration (EIA), the statistical and analytical agency within the
U.S. Department of Energy (DOE), prepared this report. By law, our data, analyses, and forecasts are independent of approval by any other officer or employee of the U.S. Government. The views in this report do not represent those of DOE or any other federal agencies.

1. B.1. Respondent Universe

The target population universe for the form EIA-112 is all utilities and non-utility companies that provide natural gas or electricity service to residential customers and are responsible for service disconnection and reconnection. The universe excludes electric entities that provide less than 200 gigawatt hours annually. Information on service disconnections, reconnections, and final notices due to bill nonpayment will be collected. The first year, a census of all companies in the target population will be conducted. The subsequent two years (hereafter referred to as ‘non-census years’), a cutoff sample will be used.

The frame for the form EIA-112 comes from two established EIA census surveys: the form EIA-176 Annual Report of Natural Gas and Supplemental Gas Supply and Disposition for natural gas and the form EIA-861 Annual Electric Power Industry Report for electricity. Both annual surveys have corresponding monthly surveys, the form EIA-861M for electricity and the form EIA-857 for natural gas. Small electric entities that report on the Form EIA-861S, Annual Electric Power Industry Report (Short Form), will be excluded from the form EIA-112 to reduce respondent burden.

The form EIA-176 is a natural gas survey of all known utilities and nonutility companies that sell and/or deliver natural gas to residential customers, as well as other kinds of natural gas companies. Only respondents that have residential customers will be included on the EIA-112. The EIA-861 is a survey for retail power sales entities and like the EIA-176, only entities with residential customers will be included in the EIA-112 universe. Energy only service providers that do not provide delivery service and do not issue disconnections and so will be excluded. The EIA-112 would have 2,407 respondents from the EIA-176 and EIA-861 combined.

2. B.2. Statistical Methods 

The first year, a census of all respondents will be conducted. The subsequent two years (hereafter referred to as ‘non-census years’), a cutoff sample will be used in order to reduce respondent burden. In the census year, the EIA-112 will have 2,407 respondents from the EIA-176 and EIA-861 combined. In the non-census years EIA will use cutoff samples from Form EIA-857 and Form EIA-861M, which are based on utility size and state coverage of sales. This reduces the number of respondents to 491 in each of the final 2 years, making an average of 1,130 respondents each year. This will allow EIA to reduce respondent burden on smaller entities, while also producing population estimates at the state-level by modeling data for the entities on the frames that were subjected to sampling but not selected. The table below shows the numbers of companies in the sample for both census and non-census years for electricity and natural gas.

	112 Census			112 Non-Census
	Electricity	Natural Gas	Total	Electricity	Natural Gas	Total
Population	1208	1199	2407	1208	1199	2407
Total Number of Respondents	1208	1199	2407	340	151	491
Coverage*	100%	100%		87%	90%

*Coverage calculations exclude Form EIA-861S respondents

Estimation and imputation will both be performed using a weighted least squares regression model. In the first year of collection, Form EIA-112 will collect from a census of eligible respondents. Nonrespondents and responses which fail data quality edit checks will be imputed using a weighted least squares regression model using residential customer counts as the independent variable. The weight will be calculated using residential customer counts. In non-census years, a weighted least squares regression model will also be used, but EIA will investigate using a company’s prior census value as the independent variable and consequently may use a model that is not the same as in the census year. This estimation methodology would be similar to the estimation methodology used in the Form EIA-861M for other types of data items. The model developed will be used to do both imputation for nonrespondents and edit failures, as well as estimation for Non sampled companies.

The estimates generated will be both national and state level totals of final notices, disconnections, and reconnections. Relative standard errors will also be calculated for the state level estimates and published with the corresponding estimates.

3. B.3. Maximizing Response Rates 

To maximize response rates, the form has been designed and the instructions written to be clear and concise. 

Nonrespondents will be contacted by e-mail, telephone, and letter to request their data submission. If no response occurs, additional correspondence will be sent to higher-level management officials to request the submission of the data.

4. B.4. Test Procedures and Form Consultations 

EIA conducted a pilot study to support the collection of this new data series. Data collection for the pilot study began on November 1, 2023. The reference period for this survey is October 2023. All reported statistics herein are only indicative of this single-month reference period. Data collection was handled through two modes: a webform and a spreadsheet form. Webforms were submitted through the EIA Respondent Portal, and spreadsheet forms were submitted through a Secure File Transfer Protocol (SFTP).

Respondents had until November 30, 2023, to complete their questionnaires. For any respondents who had not submitted, we conducted a targeted non-response phone follow-up. We stopped data collection efforts on December 15, 2023. Any late submissions were reviewed on a case-by-case basis.

For this pilot study, data were collected using Computer-Assisted Self-Interviews (CASI), also known as a web survey, from November 1, 2023, to December 15, 2023. The sample, as described below, received a survey invite email on October 25, 2023, and one reminder recruitment email on the opening day of data collection, November 1, 2023. Nonrespondents were sent two additional reminders on November 20, 2023, and November 28, 2023. Data collection stopped on December 15, 2023.

The sample for the pilot study comprises electric and dual-service utilities from the electricity-based EIA-861M, Monthly Electric Power Industry Report, cutoff sample and natural gas and dual-service utilities from the natural gas-based EIA-857, Monthly Report of Natural Gas Purchases and Deliveries to Consumers, cutoff sample. Dual-service utilities provide both electricity and natural gas service.

Because of state law, some utilities may not have had authority to conduct disconnections, and as such, reported zeros for the appropriate questions. Power marketers and third-party owners of rooftop solar were excluded from the sample because these respondents to the Monthly Electric Power Industry Report do not provide electricity service directly.

In total, the EIA-861M sample included 419 unique electric utilities, the EIA-857 included 157 unique natural gas utilities, and entities in both the EIA-861M and EIA-857 included 43 unique dual-service utilities. The selected sample units are located across different regions of the United States, are large utilities in terms of total sales volume, are of various sizes based on the number of residential customers (small, medium, large), and are of different types based on the ownership model (investor, municipal, public, cooperative, state, and federal).

In total, 201 entities completed the pilot study, 16 entities partially completed the pilot study, 30 entities refused, and 415 entities did not respond. The response rate was 32.53%, excluding the partial completes from the numerator.

The pilot study found that most utilities appear to be able to report this data without difficulty. In terms of difficulty, over half of respondents (52%) found it “very easy” or “somewhat easy,” and 73% of respondents said that completing the survey was either easy or neutral. Even respondents that reported difficulty were able to provide some but not all data as requested.

For additional information on forms review, please see Part A, section A.8., “Summary of Consultations Outside of the Agency.” For a summary of the comments received and EIA’s responses, see the Public Comment File submitted with this OMB package.

5. B.5. Statistical Consultations 

For additional information concerning this proposed information collection, please contact [email protected].