Survey of Commercial and Institutional Energy Use (SCIEU)

Detailed information for 2014





Record number:


The purpose of this survey is to provide detailed information on the energy demand and consumption patterns of Canadian businesses, organizations and institutions.

Data release - September 16, 2016


The purpose of this survey, sponsored by Natural Resources Canada, is to provide detailed information on the energy demand and consumption patterns of Canadian businesses, organizations and institutions. The survey collects data on the types and quantities of energy (such as electricity, natural gas etc.) consumed by businesses and institutions in Canada and by their buildings.

The data will be used (by utilities, as well as provincial and federal governments) to develop programs and policies that will improve the energy efficiency of commercial and institutional buildings in Canada. In addition, they will be used to support target programs, such as ENERGY STAR Portfolio Manager, providing the basis for updating the ENERGY STAR 1-100 performance scores currently available and for creating new scores for building types not currently eligible for a score. Industry associations, building managers and business owners will have up-to-date data on similar buildings with which to compare their own energy consumption. Energy specialists and consumers can use the data to learn more about their building consumption patterns.

Reference period: Calendar year


  • Energy
  • Energy consumption and disposition

Data sources and methodology

Target population

The survey targets two populations: one to collect data on the floor space and energy used by establishments (establishment component) and the other to collect similar information for the buildings which the establishments occupy (building component).

Establishment Component:
This target population includes all active establishments in Canada in industries which belong to the commercial and institutional sector. Defined based on the North American Industry Classification System (NAICS), these industries include: 41, 44, 45, 48, 49, 51, 52, 53, 54, 55, 56, 61, 62, 71, 72, 81 and 91. These establishments must rent or own commercial or institutional space and have at least one employee. Establishments located in the Canadian territories (Yukon, Northwest Territories and Nunavut) are excluded.

Building Component:
This target population includes buildings which have at least 50% of their space used for commercial or institutional purposes and have at least one full-time employee working during the shift when the largest number of employees is in the building of interest. Military bases, embassies and portable structures, as well as buildings located in the territories are excluded.

The targeted buildings can be classified by activity to the following 10 types:

- Office Building (non-medical)
- Office Building (medical)
- Primary or secondary school
- Assisted Daily Care Facility and/or Residential Care Facility
- Warehouse
- Hotel, Motel and/or Lodge
- Hospital
- Food and Beverage Stores (excluding restaurants and bars)
- Retail Store (non-food)
- Other activity or function

Instrument design

In partnership with Natural Resources Canada, five collection instruments were developed to collect information on energy use by establishments and their buildings.

The first collection instrument was a script used during telephone interviews with sampled establishments to confirm that they were in-scope for the survey; to identify contact persons for the collection of establishment and building information; and to confirm the number, type and size of the buildings occupied by the establishment. The latter information was used to build a frame for the selection of a sample of buildings.

The four remaining collection instruments are used to collect information on the energy used by establishments (establishment questionnaire), by establishments' buildings (building questionnaire) or for both (hybrid questionnaire). The latter instrument is used in cases where the establishment is the sole occupant of only one building.

Questionnaire design specialists were consulted in the design and testing of the survey questionnaires. The questionnaires were tested with commercial and institutional establishments sampled from the Business Register. Their comments on the design and content were incorporated into the final versions.


This is a sample survey with a cross-sectional design.

Two distinct but related samples are selected for the Survey of Commercial and Institutional Energy Use. A sample of establishments is selected first (establishment component). Information collected from in-scope establishments about their buildings is used to create a frame from which a sample of buildings is selected.

Establishment Component:
The survey frame for establishments is created from two lists. The first is a list of primary and secondary schools (private and public) compiled based on administrative data and research by the Center for Special Business Projects. The second list includes establishments in the industries targeted by the survey (excluding schools) drawn from the Statistics Canada's Business Register (BR) for reference month June 2014. In total, the frame consisted of 925,356 establishments. The frame of establishments is stratified based on industry group, region and size.

Establishments with more than 10,000 employees are selected with certainty. The largest establishments associated with buildings believed to be influential, are also incorporated with certainty. The establishments associated with each of the 11 largest government departments are also selected with certainty.

The allocation of the establishments is done in two steps. The first step ensures that the establishment-level estimates will be of sufficient quality. The second step adds sample to the initial allocation when required to support the production of building-based estimates. The post-allocation sample of 11,530 establishments was selected using the stratified simple random sampling approach.

Building Component:
The frame of buildings is the list of the buildings collected from in-scope, sampled establishments. No additional stratification of buildings is required beyond what has already been done for the establishments to which the buildings are linked.

An upper limit is placed on the number of buildings selected from those listed by each establishment. This is done in order to manage the response burden imposed on multi-building establishments. All buildings are selected for establishments that had 3 buildings or less. Three buildings are selected for establishments that had at least 4 but less than 50 buildings and 4 buildings are selected for the establishments with 50 or more buildings.

The sample of buildings for each establishment is selected using a systematic sampling approach whereby the buildings are first sorted in descending order by building size. The size of each building is requested when the list of buildings is collected from each establishment.

Data sources

Data collection for this reference period: 2014-08-04 to 2015-12-18

Responding to this survey is mandatory.

Data are collected directly from survey respondents.

The survey is conducted through multiple modes. The establishment or building worksheets are administered using computer-assisted personal interviews. The paper worksheet is provided to respondents in advance to help them prepare for the interview. The establishment, hybrid, and building questionnaires are administered by mail using a paper questionnaire which is mailed back. A supplemental paper questionnaire is being sent to warehouses to collect additional detail for this building activity type. The warehouse questionnaire is also returned by mail. Follow-up for these questionnaires is done through a computer-assisted telephone interview.

The decision to collect the information through paper questionnaires or through personnel interviews was made based on the number of questionnaires and the types of questionnaires a respondent would be required to complete.

All collection for Hybrids is completed through paper questionnaire with CATI follow-up. The collection mode for the establishment questionnaire will depend on whether the contact person for the establishment and for the buildings which the establishment occupies, are the same. In the case where they are the same, both the establishment and the building data are collected using the CAPI application. For establishments with two unique contact persons for the establishment and building information, the establishment information is collected using a paper questionnaire with CATI follow-up and the building information collected using CAPI.

View the Questionnaire(s) and reporting guide(s).

Error detection

Throughout the collection period, attention was given to the interviewer case remarks and notes to identify missing or incorrectly recorded values. Validity and consistency edits were performed in the collection application to limit potential errors. There was ongoing review of the write-in fields of the questionnaire for re-coding the responses.

In the post-collection stage, validity, distribution and consistency edits were performed to identify incorrect, missing and invalid responses for usable records. Edits based on ratios were applied to detect errors and inconsistencies in the reported survey data following collection. Outlier detection was also used to identify extreme values requiring imputation.


Imputation was performed to treat partial non-responses (also called item non-responses). Note that total non-responses were accounted for during the weighting process by adjusting the weight of all responding records.

A returned questionnaire was considered as a partial non-response when some variables were fully completed by a respondent, but one or more other variables were left blank and would need to be completed via imputation. Excluding instances where certain questions do not apply to some respondents, we observed that about 10-20% of the data for the various variables required imputation (on average). The imputation of non-responses was performed using the nearest neighbour donor imputation procedure in the generalized system BANFF. This procedure uses a nearest neighbour approach to find, for each record requiring imputation, the valid record that is most similar to it and that will allow the imputed recipient record to pass the specified imputation edits and post edits.

Nearest neighbour donor imputation was applied when variables in a record requiring imputation (the recipient) were identified and imputed using a donor record mostly similar to the recipient record. These similar records were found by taking into account other variables that were correlated with the missing/incorrect values via the customized imputation classes and matching variables for each variable to be imputed. If nearest neighbour donors were not found for all recipients, then it was necessary to be less restrictive by changing the imputation classes and reprocessing the data. For example, similar units would be located by region rather than by province and thus more donors in each class would be available for the recipients requiring imputation. This imputation processing continued by a predetermined sequence until nearest neighbour donors were assigned to all records requiring imputation or until no nearest neighbour donors were available. Once nearest neighbour donors were found, the new imputed values replaced the previous missing or incorrect values for the variable. During imputation, edits and post edits were applied to ensure that the resulting record did not violate any of the specified edits.


For establishments, the estimation involves use of weights to be used alongside the variables. These weights are simply defined as the inverse of the selection probabilities within the strata. These initial weights are later adjusted to take into account total non-response. As for variance estimation, the Generalized Estimation System was used in the context of a stratified simple random sampling approach.

In order to calculate estimation weights for selected buildings, the Weight Share Method (WSM) was used. Typically in sample surveys, the estimation weights are the inverse of the selection probability, which leads to unbiased estimates. With the complex relationships between buildings and establishments, calculating the selection probability for each building was not feasible, and the Weight Share Method provided a simplified approach which required less information on links to non-sampled units and simpler calculations, while still producing unbiased estimates. The WSM was used to calculate the initial weights, which were then adjusted for unit non-response encountered during the personal interview phase. For variance estimation, the exact theoretical formulas where programmed. The various terms in the formulas were obtained using several runs of the Generalized Estimation System (GES).

Quality evaluation

Prior to the data release, survey results were analyzed for quality. In general, this analysis included a detailed review of individual responses, coherence with results from related indicators and information from other external sources.

Disclosure control

Statistics Canada is prohibited by law from releasing any information it collects that could identify any person, business, or organization, unless consent has been given by the respondent or as permitted by the Statistics Act. Various confidentiality rules are applied to all data that are released or published to prevent the publication or disclosure of any information deemed confidential. If necessary, data are suppressed to prevent direct or residual disclosure of identifiable data.

In order to prevent any data disclosure, confidentiality analysis is done using the Statistics Canada Generalized Disclosure Control System (G-Confid). G-Confid is used for primary suppression (direct disclosure) as well as for secondary suppression (residual disclosure). Direct disclosure occurs when the value in a tabulation cell is composed of or dominated by few enterprises while residual disclosure occurs when confidential information can be derived indirectly by piecing together information from different sources or data series.

Revisions and seasonal adjustment

This methodology does not apply to this survey.

Data accuracy

Because the data are based on a sample, they are subject to sampling error. Estimates based on a sample vary from sample to sample and are typically different from the results that would have been obtained from a complete census with a 100% response rate.
Quality indicators were calculated using Statistics Canada's general estimation system, G-EST. The method used to calculate these indicators differed depending on the component.

Building component
Sampling variability of the building component estimates can be estimated using the bootstrap method. Bootstrap weights were created for the building component and were used to estimate the sampling variability of all produced estimates. This method was chosen for the building component because of the complexity of the sampling plan.

Institution component
Taylor's linearization was used to calculate the sampling variability of the estimates.

Response rate
The collection response rate for the building component is 25.6%.
The collection response rate for the institution component is 70.7%

Non-sampling errors
Common sources of non-sampling errors include imperfect coverage, classification errors and non-response. Coverage errors, or imperfect coverage, arise when there are differences between the target population and the surveyed population. These differences can be caused by exclusion of units on the frame or inclusion of units outside of the target population in the sample. If the excluded population differs from the survey population, the results may be biased. In general, since these exclusions are small, one would expect the biases introduced to be small. Classification errors are related to stratification information such as the province and building type (or campus type). If these are not correct on the frame, units are selected in one stratum when they belong to another, and this can have an impact on the estimates. Non-response could occur either at pre-contact or during main collection. Survey estimates were adjusted (i.e., weighted) to account for non-response cases. Other types of non-sampling errors can include response and processing errors.

Non-response bias

The main method used to reduce non-response bias involved a series of adjustments to the survey weights to account for non-response as much as possible.

Coverage error

Building component
The frame for the building component was the SBgR. This new Statistics Canada product is a list of all buildings in Canada with a physical address compiled from different administrative data sources. Unfortunately, not all buildings in Canada are represented by a physical address on an administrative data file. It is not possible to evaluate the coverage error related to these exclusions, but it is known that the impact is bigger in rural areas than in urban areas.

Institution component
The frame for the institution component was the BR. Coverage error for the BR is low.

Classification error

Building component
The 2019 SCIEU frame was built using the SBgR and was then stratified by building type and province or territory. The information available to classify buildings by activity type was not present for all buildings; therefore, some buildings were added in a supplementary stratum, and a sample was selected in that unknown stratum. If certain activity types of interest were more represented than others in that stratum, it is possible that not enough sample was selected to produce good estimates for that building type.

Other non-sampling errors

A significant effort was made to minimize bias by using a well-tested questionnaire, a proven methodology, specialized interviewers and strict quality control.


Date modified: