Data quality, concepts and methodology: Data quality, concepts and methodology

View the most recent version.

Introduction

The statistics in this publication were collected under the authority of the Statistics Act (Revised Statutes of Canada, 1985, chapter S19). The statistics are derived from two surveys: the Coastwise Shipping Survey (Survey number 2751) and the Marine International Freight Origin Destination Survey (Survey number 2791). ¹  The purpose, populations, data sources, and data quality for these surveys are discussed in the following sections.

Purpose of the surveys

These surveys collect data on:

1. Coastwise or domestic marine transport - the carriage of cargo by ships between Canadian ports in Canadian waters; and,
2. International marine transport – the carriage of cargo between Canadian ports and foreign ports in the United States or overseas.

The information collected in these surveys includes the description and quantities of commodities loaded and unloaded at Canadian ports and marine terminals by vessels calling at those facilities, and the origin and destination of those commodities. Data are aggregated for publication of commodity traffic by Canadian port and foreign country of origin or destination.

Statistics Canada uses these data as one input to Canada’s System of National Accounts. The data are also used by Transport Canada, other federal departments, provincial ministries, transportation companies, consulting firms, universities and foreign governments. This information is used to analyse transportation activities, for marketing and economic studies, and to assess the industry’s performance.

Survey populations

For each survey it is possible to define a target population as all the units (e.g., ships) for which data were intended to be collected and a survey population as all the units for which the survey actually provides data. The survey population may differ from the target population when it is not possible to identify or to survey all of the units within the target population.

The target populations of the Coastwise Shipping Survey and the Marine International Freight Origin Destination Survey are all ships that carry cargo between Canadian ports or between Canadian ports and foreign ports.

The Coastwise Shipping Survey population is somewhat different from the target population. It consists of ships engaged in domestic shipping at Canadian ports, with the exception of:

1. cargo vessels less than 15 net register tons (NRT);
2. tugs or other vessels under 15 gross register tons (GRT);
3. Canadian naval ships;
4. fishing vessels;
5. research vessels;
6. ballast movements for towboat and ferry operators on the West Coast ports;
7. cargo carried in trucks and rail cars on domestic vehicle ferries.

This limited survey field may have a significant impact on the statistics reported for some commodities, regions and ports such as the Pacific Coast, where a number of tugs that are under the GRT threshold are engaged in the towing log booms. The exclusion of ferry traffic will also affect the statistics for cargo transported to and from islands such as Newfoundland, Prince Edward Island and Vancouver Island. However, this cargo should be reflected in statistics collected for freight transported by the truck and rail modes.

The survey population for the Marine International Freight Origin Destination Survey includes vessels engaged in international shipping at Canadian ports with the exception of (i) fishing boats registered in Canada or abroad, (ii) maintenance and service ships such as icebreakers, (iii) research vessels and (iv) other non-commercial vessels such as hospital ships.

Data sources

Data for the Coastwise Shipping Survey are collected directly from the owners, operators, agents or officers of ships engaged in domestic shipping. The data are reported on one of two questionnaires, either the S.1, Domestic Shipping Report or the S.4, Towboat and Ferry Operators Shipping Report. The operators of cargo-carrying ships arriving and departing Canadian ports or marine terminals complete these reports for each such arrival and departure. The S.1 Report provides the following information: vessel name, country of registration, the vessel’s gross and net registered tonnage, cargo description, tonnage and ports of loading and unloading. The S.4 Report is used only on the West Coast by operators of tugs, cargo barges and ferries for monthly reporting of coastal shipping. These operators do not report ballast movements. The S1 and S4 Reports are sent to Statistics Canada each month.

Information on international shipping is collected from the Advance Commercial Information (ACI) program & the A6 General Declaration and manifests required by the Canada Border Services Agency (CBSA) for each ship loading or unloading cargo in Canada that either originated in or is destined to a foreign country. Local CBSA employees send copies of these documents to Statistics Canada. In some cases, marine agents or port authorities send equivalent reports to Statistics Canada. The data provided by the ACI program & the Form A6 and its annexes provide information on: the vessel’s name and country of registration, registered gross and net tonnage, and information such as the description of the cargo, tonnage containerizes information and ports where commodities will be loaded and unloaded. These variables constitute the database for statistics on international shipping.

Data quality

All statistical surveys are subject to errors. The total error of a survey is defined as the difference between the survey estimate and the actual value for the target population. Survey errors can be classified as “sampling” or “non-sampling” errors. The marine transport surveys are censuses and are therefore not subject to sampling errors.

There are four main types of non-sampling errors:

1. Coverage errors (when the survey frame does not thoroughly cover the target population);
2. Response errors (when a respondent provides inaccurate information);
3. Non-response errors (when a respondent does not respond to all or part of the questionnaire) and,
4. Processing errors (errors made during data capture, coding, and editing).

The Coastwise Shipping Survey is thought to be more affected by coverage errors than the Marine International Freight Origin Destination Survey. The survey population is smaller than the target population due to the size threshold and the various ship types that are exempted (e.g., vehicle and rail ferries) and the practical difficulty of identifying the fleet of all ships that are active in the domestic transport of cargo. With respect to the Marine International Freight Origin Destination Survey, customs rules apply to all ships arriving from or departing for foreign ports. However, Statistics Canada has only limited tools for identifying missing shipping documents.

Both surveys have significant exposure to response error, particularly in terms of the reported commodities. The marine carrier of domestic and international cargo is not always able to provide descriptions that are sufficiently detailed to accurately code the commodity. Terms such as general cargo are often used in cases where the carrier is unsure of the exact contents of a container or when there are a number of commodities being transported in one shipment. As result, residual categories such as “Other manufactured and miscellaneous goods” and “Other chemical products and preparations” may be overestimated while the true commodity is understated. Respondents may also use incorrect conversion factors in estimating the weight of a shipment.

Non-response errors are expected to have more of an effect on the Coastwise Shipping Survey. Shipping reports are not submitted for all vessel traffic Canadian ports. It is known that there is a significant amount of under-reporting for the Pacific Coast and the Mackenzie River. While it is assumed that the reporting to the Canada Border Services Agency (CBSA) of international traffic is fairly comprehensive, there are occasions when reports are not forwarded to Statistics Canada, particularly for ports in the Canadian Arctic and other remote areas that do not have local customs offices.

Both surveys are subject to processing errors during data entry, including errors in transcribing and coding information and in converting cargo quantities from non-weight units (e.g., barrels or cubic metres to kilograms). Errors may also be introduced during editing and imputation procedures that are used to correct data that appear to be erroneously reported or partially reported. A number of techniques are used to detect and resolve these errors including micro and macro analysis of the data for congruency and consistency with information acquired from other sources (e.g., cargo statistics from port authorities and economic statistics from other Statistics Canada divisions).