Retail Trade Survey (Monthly)

Detailed information for October 2003

The Monthly Retail Trade Survey collects sales and the number of retail locations by province and territory from a sample of retailers.

Data release - December 22, 2003

• Questionnaire(s) and reporting guide(s)
• Description
• Data sources and methodology
• Data accuracy

Description

The Monthly Retail Trade Survey collects sales and the number of retail locations by province and territory from a sample of retailers. Sales estimates obtained from retailers are a key monthly indicator of consumer purchasing patterns in Canada. Furthermore, retail sales are an important component of the Gross Domestic Product, which measures Canada's production, and are part of many economic models used by public and private agencies. The Bank of Canada relies partly on monthly retail sales estimates when making decisions that influence interest rates. Businesses use retail sales estimates to track their own performance against industry averages and to prepare investment strategies.

Retail sales estimates do not include any form of direct selling that bypasses the retail store, e.g., direct door-to-door selling; sales made through automatic vending machines; sales of newspapers or magazines sold directly by printers or publishers; and sales made by book and record clubs. Internet retailing activities are included in the survey only when conducted through the same legal structure as the retail establishment.

For additional information on Methodology, see Appendices II, III and IV of the "Retail Trade" publication, Statistics Canada catalogue no. 63-005-XWE available through the link "Publications" included in the side bar menu.

Reference period: Month

Collection period: The 1st day of the month following the reference month to The last day of the month following the reference month

Subjects

• Retail and wholesale
• Retail sales by type of store

Data sources and methodology

Target population

The Monthly Retail Trade Survey universe is based on the 1980 Standard Industrial Classification.

The sample of the Monthly Retail Trade Survey is drawn from Statistics Canada's Central Frame Data Base (CFDB), part of which comprises businesses engaged in retail activity.

The target population for the Retail Trade Survey consists of all statistical companies on the CFDB that have a location identified in the retail trade sector. These units comprise the sampling frame for the Retail Trade Survey.

Instrument design

The questionnaire collects monthly data on retail sales and the number of retail locations by province and territory from a sample of retailers. The items on the questionnaire have remained unchanged for several years. However, should modifications become necessary, proposed changes would go through a review committee, and a field test with respondents and data users to ensure its relevancy.

Sampling

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

The Monthly Retail Trade Survey sample design is a stratified simple random sample without replacement. The frame is derived from Statistic Canada's Central Frame Data Base (CFDB). The stratification is based on trade groups, provinces, territories and some Census Metropolitan Areas (CMAs - i.e. Montreal, Toronto, Winnipeg, and Vancouver). Trade groups are a grouping of 3 and 4 digit 1980 Standard Industrial Classification Codes (SIC). Each trade group / geography combination is further subdivided into 3 substrata, based on size. The first of these substrata is a take-all stratum (census) and contains all complex businesses as well as all businesses in a trade group / geography combination whose revenues exceed a given threshold. The other two strata are take-some (partially sampled) and contain medium and small businesses respectively. Businesses are assigned to these strata based on estimates of their revenues found on the CFDB.

The sample size is about 14,000, representing slightly less than ten percent of the target population.

Data sources

Responding to this survey is mandatory.

Data are collected directly from survey respondents and extracted from administrative files.

Data collection, data capture, preliminary edit and follow-up of non-respondents are primarily performed by staff in the Statistics Canada regional offices. Sampled companies are contacted either by mail or telephone, whichever they prefer. Data capture and preliminary editing are performed simultaneously to ensure the validity of the data. Companies from which no response has been received or whose data may contain errors, are followed up immediately.

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

Error detection

There are edits built into the data capture application to verify the entered data for unusual values, as well as to check for logical inconsistencies. Whenever an edit fails, the interviewer is prompted to correct the information (with the help of the respondent when necessary). For most edit failures the interviewer has the ability to override the edit failure.

Once the data is received back at head office an extensive series of processing steps is undertaken to thoroughly verify each record received. Edits are performed at a more aggregate level (trade group by geographic level) to detect records which deviate from the expected, either by exhibiting large month-to-month change, or differing significantly from the remaining companies. All data failing these edits are subject to manual inspection and possible corrective action.

Imputation

Imputation is applied to missing records. The imputation system automatically selects the appropriate method depending on the availability of the data. Possible imputation methods are based on month-to-month trends, year-to-year trends, historical data, annual data, etc. Records that fail statistical edits are considered as outliers and are not used in calculating imputation variables (such as monthly trends) used by the imputation system.

There is an identifiable lag between the time a company opens and its appearance on the survey frame. To compensate for the effect this time lag has on monthly estimates, sales for sample births are imputed back to the actual date of birth or the beginning of the current year (if the actual date of birth is prior to the current year).

Estimation

Retail sales are estimated by increasing the in-sample sales results by an estimation weight. An initial weight equal to the inverse of the original probability of selection is assigned to each entity. The weights are subsequently adjusted for achieved sample size, in order to inflate the estimate to represent the entire current population. The calculated weighted sales values are summed by domain, to produce the total sales estimates by each trade group/geographic area combination. A domain is defined as the most recent classification values available from the CFDB for the statistical entity and the survey reference period. These domains may differ from the original sampling strata because records may have changed size, industry, or location. Changes in classification are reflected immediately in the estimates and do not accumulate over time.

The variance is derived directly from a stratified simple random sample without replacement.

Quality evaluation

The methodology of this survey has been designed to control errors and to reduce their potential effects on estimates. However, the survey results remain subject to errors, of which sampling error is only one component of the total survey error. All other errors arising from the various phases of a survey are referred to as non-sampling errors. For example, these types of errors can occur when a respondent provides incorrect information or does not answer certain questions; when a unit in the target population is omitted or covered more than once; when a unit that is out of scope for the survey is included by mistake or when errors occur in data processing, such as coding or capture errors. While the impact of non-sampling errors is difficult to evaluate, certain measures such as response and imputation rates can be used as indicators of the potential level of non-sampling error.

Prior to publication, combined survey results are analyzed for comparability; in general, this includes a detailed review of individual responses (especially for the largest companies), general economic conditions and historical trends.

Disclosure control

Statistics Canada is prohibited by law from releasing any information it collects which 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.

Confidentiality analysis includes the detection of possible direct disclosure, which occurs when the value in a tabulation cell is composed of a few respondents or when the cell is dominated by a few companies.

Revisions and seasonal adjustment

Revisions in the raw data are required to correct known non-sampling errors. These normally include replacing imputed data with reported data, corrections to previously reported data, and estimates for new births that were not known at the time of the original estimates.

Raw data are revised, on a monthly basis, for the month immediately prior to the current reference month being published. That is, when data for December are being published for the first time, there will also be revisions, if necessary, to the raw data for November. In addition, revisions are made once a year, with the release of the preliminary February data, for all months in the previous year. The purpose is to correct any significant problems that have been found that apply for an extended period. The actual period of revision depends on the nature of the problem identified, but never exceeds three years.

The seasonally adjusted data, obtained with the X11-ARIMA program, also need to be revised. In part, they need to reflect the revisions identified in the raw data. Also, since the trend is sensitive to the most recent values reported in the raw data, seasonally adjusted values are revised for the previous three months with the release of each month of new data.

Once a year, all the seasonally adjusted parameters are revised to incorporate the most recent data. For example, trading day weights are adjusted to reflect the changing importance of Sunday as a shopping day. Revised seasonally adjusted estimates for each month in the previous three calendar years are released at the same time as the annual revision to the raw data.

Data accuracy

Coefficients of variation (CV), response fractions for sales and inventories, and response rates are major data quality measures used to validate results from the Monthly Retail Trade Survey.

The coefficient of variation is used to measure the sampling error of the estimates. The coefficient of variation, at the national level for total retail sales, has ranged, on a monthly basis, from 0.7% to 1.3%, with the introduction of a new sample design in late 1997. Normally, (19 times out of 20) the error caused by sampling, expressed as a percentage of the sample estimate, will be within twice (plus or minus) the coefficient of variation.

The response fraction is the proportion of the estimate that is based upon reported data. For example, a cell with 20 active sample units in which 10 respond for a particular month would have a response rate of 50%. However, if the 10 reporting units represented $9 million out to a total of $12 million, the response fraction would be 75%. Thus, 25% of the estimate would come from imputed data. The final response fraction is usually in the 94%-96% range. The response fraction for the first month for which the data are published is normally a couple of percentage points below the final response fraction.

The response rate, which is about 88%, is a measure of the proportion of those sample units that have responded in time for inclusion in the estimate.