Annual Environmental Protection Expenditures Survey (EPES)
Summary of changes
Activity on this program started: 1994
No major changes since 2018. A few questions have been added to the content of the questionnaire to reduce follow-up calls. The sample size has been slightly reduced in order to diminish the response burden following the complex situation caused by COVID-19.
The survey has been redesigned for the reference year 2018 to harmonize the environmental protection activities with international classifications and extend the survey scope to include all expenditures made to protect the environment, whether or not they are in response to current or anticipated Canadian or international regulations, conventions or voluntary agreements. It introduces new questions on goods, technologies and services purchased by industries during the reference year, as well as questions on resource management.
The survey is now annual.
The target population now includes the pipeline transportation industry. Smaller establishments have been excluded from the target population using the Royce-Maranda method (meaning that 10% of the smallest establishments in term of the size measure (revenue) have been excluded from the population).
The sample size has increased and is now of 6,000 establishments.
In order to align with the Integrated Business Statistics Program (IBSP), an electronic questionnaire (EQ) was used for the first time for the survey, Bernoulli sampling was done instead of simple random sampling, and the error detection and imputation methodologies were modified. Quality indicators were also defined differently from the previous cycle and thus are not comparable.
Target population - The target population has been reduced. Establishments in the Pipeline Transportation Industry and those involved in contract logging are no longer included in the survey coverage.
For the 2008 reference year, sampling was performed exclusively at the establishment level. The oil and gas extraction industry, which was previously sampled at the enterprise level, was sampled at the establishment level for 2008.
A new sampling strategy was used for 2008 for establishments with fewer than 50 employees. In 2006, a stratified sampling strategy was used which included establishments with fewer than 50 employees. In 2008, these smaller establishments were sampled separately from the larger ones, and the sample size for them was reduced.
The survey underwent a thorough redesign for the 2006 reference year to improve the overall methodology and data quality for smaller businesses. The redesign also facilitated the production of data quality indicators. Because of the redesign, comparisons with survey estimates for years prior to 2006 are not recommended.
Target population - The target population has been expanded to include smaller establishments with 20 to 49 employees. In previous years only establishments with 50 or more employees were part of the target population.
For the 2006 reference year, questions dealing with technologies to improve energy efficiency replaced similar questions on greenhouse gas emissions asked on the past two cycles of the survey. During the development phase of this material, a series of one-on-one interviews with businesses were conducted to test the new material, as well as to collect feedback on existing survey questions.
Sampling - The sampling methodology underwent a complete redesign for the 2006 reference year. For the first time the Business Register was used as the survey frame, the sample size was increased and the employment threshold was reduced to include smaller units in the sample. In previous cycles only establishments with more than 49 employees were selected, but for 2006 establishments with 20 or more employees were selected.
Imputation - The imputation process was updated as part of the redesign of the survey. In previous years, imputation was used for all non-response records, including total non-response. Imputation programs were used to identify best-fit donors, moving systematically through increasingly general donor pools. Criteria for donor identification were various combinations of industry group and geographical location (province, region, or Canada). The programs used for this process were developed by Environment Accounts and Statistics Division. For the 2006 reference year redesign, Statistics Canada generalized edit and imputation system (Banff) was used for this process.
Estimation - In previous years, estimation was done for establishments that had 49 or more employees but were not surveyed. The mean of the environmental protection expenditures to employment ratio for a particular industry group and province/region was multiplied by the number of employees to determine an estimate for a non-sampled establishment. No estimation or imputation was done for questions dealing with qualitative information. For 2006, estimation and imputation was done for qualitative information and a measure of sampling error was calculated for all estimates.
Collection instrument - Unlike past years where two questionnaires were used (one long and one short), for the 2006 reference year only one survey questionnaire was used. The 2006 questionnaire is similar to the former long questionnaire.
Instrument design - For the 2002 reference year, questions dealing with greenhouse gas emission reduction technologies were added. During the development phase of this material, a series of one-on-one interviews with businesses were conducted to test the new material, as well as to collect feedback on existing survey questions.