- As explored in LMI Insight Report 31, the O*NET system developed by the US Bureau of Labor Statistics (BLS) contains one of the most utilized skills and competencies taxonomies today. In particular, the skills ratings of almost 1,000 occupations are useful for a variety of applications, including informing Employment and Social Development Canada’s new Skills and Competencies Taxonomy.
- Leveraging the insights in O*NET requires a comprehensive, rigorous, validated concordance (or crosswalk, correspondence) between Canada’s National Occupational Classification (NOC) and the US Standard Occupational Classification (SOC) linked to O*NET.
- Statistics Canada and Employment and Social Development Canada have worked together to develop two concordances: one from Statistics Canada’s mapping of the Canadian NOC (4-digit level) to the US SOC (6-digit level) to develop statistical indicators; the other from Employment and Social Development Canada's mapping of the Career Handbook (5-digit NOC) to O*NET-SOC (8-digit level) to create occupational profiles using a skills lens.
- In some cases, the skills profile of O*NET for one US occupation applies to several Canadian occupations; in others, several O*NET profiles apply to one Canadian occupation; and for 62 Canadian occupations at the 5-digit level (representing 10% of total employment), a gap remains in our ability to match O*NET skills-related data.
Employment Development Canada (ESDC), Statistics Canada (STC) and the Labour Market Information Council (LMIC) have been investigating the feasibility of using the US Occupational Information Network (O*NET) to link job–worker characteristics, such as skill importance and level ratings, to Canadian jobs. With its strong theoretical and empirical foundations and the similarity between the Canadian and US economic systems, O*NET is a credible source of occupational data for the Canadian labour market.
Since the O*NET database classifies occupations differently than in Canada, drawing on O*NET requires mapping from one classification to another. This process is achieved via a “concordance” (sometimes referred to as a “crosswalk” or “correspondence table”). In this LMI insight report, we present the first publicly accessible and methodologically transparent concordances that match O*NET data to Canadian occupations for the purpose of creating detailed occupational profiles with skills content. In addition to presenting these concordances in an open, transparent way, we articulate some considerations, especially in drawing insights related to skills.
In 2019, LMIC, STC and ESDC began assessing the various approaches to linking skills to occupations. The first step in that process was the release of the Skills and Competencies Taxonomy, a dictionary of over 280 skills, knowledge and ability descriptors, which expands the level of detail, provides consistent terminology, and improves communication of occupational information in Canada. In developing this taxonomy, ESDC produced a set of methodological principles to underpin descriptor development:
- Definitions must be mutually exclusive
- Only one concept per descriptor
- Definitions must apply in a variety of contexts yet have specific meaning
- Descriptors must be measurable in order to be assessed
As we assess the various approaches to leveraging this Skills and Competencies Taxonomy, one obvious source is the US O*NET system. Indeed, various governmental and non-governmental organizations in Canada — RBC, the Brookfield Institute for Innovation and Entrepreneurship, and British Columbia’s Ministry of Advanced Education — have built concordances between the US System of Occupational Classification (SOC) and the Canadian National Occupational Classification (NOC) to assess the skills and competencies of different segments of the Canadian labour market. These concordances, however, lack the granularity and specificity needed to leverage the full breadth of O*NET data and create detailed occupational descriptions.
The challenge is that, for most occupations, there is no neat one-to-one mapping from the Canadian NOC to the US SOC used in O*NET. For instance, in the US system there are three different skills profiles for Mathematicians (SOC 15-2021), Statisticians (SOC 15-2041) and Actuaries (SOC 15-2011). By contrast, the Canadian NOC treats these as one occupation (NOC 2161). As a result, previous concordances have focused only on occupations with neat one-to-one mappings. Moreover, these concordances are often proprietary — with little information or explanation of how they were constructed — which prevents others from using them or limits any analysis of the approach taken.
Structural Differences in the NOC and SOC Systems
To appreciate the need for a more granular concordance, we must understand the differences in how occupations are structured in the US and Canada. While the purposes of both the SOC and the NOC are similar, important differences in structure and information captured within occupational profiles are evident. For statistical reporting (e.g., number of people employed or average wage), the US SOC contains 867 6-digit classifications. On the other hand, the Canadian NOC has 500 4-digit classifications. In both systems, occupations are grouped according to a 4-level hierarchy (see Table 1). Despite the smaller number of categories, the NOC contains greater detail about the occupations than the SOC. For example, each occupation specified in the NOC is accompanied by example job titles, main duties, employment requirements, and licensing and regulatory considerations by province/territory.
Both the NOC and SOC systems have been extended to create additional occupational groups with greater detail and granularity. In Canada, ESDC’s Career Handbook provides information about work characteristics, aptitudes, interests, physical activities and environmental conditions, among other details, for 939 NOC categories at the 5-digit level. The NOC, for example, has one occupation for Mathematicians, Statisticians, and Actuaries (NOC 2161). The Career Handbook breaks this category down further to create three distinct occupations at the 5-digit level: Mathematicians (2161.1), Statisticians (2161.2), and Actuaries (2161.3). The Career Handbook is the counselling component of the NOC system and associates certain work characteristics to the detailed occupations.
Similarly, the O*NET system extends the standard 6-digit SOC to an 8-digit O*NET-SOC classification, increasing the number of occupational categories from 867 to 974. The information available in the O*NET system, however, goes beyond the Career Handbook, providing details on such job–worker characteristics as entry requirements, work activities, work values and styles, knowledge, skills and abilities.1 In linking skills to occupations in the Canadian context, a detailed, rigorous concordance between Canada’s NOC at the 5-digit level and the skills and related information in O*NET at the 8-digit level is the starting point.
Table 1: Canadian NOC versus US SOC
|US Occupations (SOC)||Canadian Occupations (NOC)|
|Statistical Structure||• 23 major groups (2-digit)
• 98 minor groups (3-digit)
• 459 broad occupations (4-digit)
• 867 detailed occupations (6-digit)
|• 10 broad occupational categories (1-digit)
• 40 major groups (2-digit)
• 140 minor groups (3-digit)
• 500 unit groups (4-digit)
|Information Provided||Lead statement and example titles||Lead statement, example titles, main duties, employment requirements, exclusions, additional information (e.g., mobility, career progression)|
|Non-statistical Extension||O*NET||Career Handbook|
|Description||974 (8-digit) occupational profiles covering skills, tasks, abilities, interests, knowledge, work activities, work context, work values, education and credentials, tools and technology
Skills are linked to occupations by an importance rating and a level of complexity rating
|939 (5-digit) occupational profiles covering aptitudes, interests, data/people/things, physical activities and environmental conditions|
The Building of a Concordance
Statistical Foundation: Statistics Canada’s Concordance
STC developed a correspondence table between the NOC 2016 and the 2018 SOC. For this correspondence table, an analysis of the occupations was completed for NOC 2016 at the 4-digit level (detailed descriptions), and for 2018 SOC (US) at the 6-digit level. This table was developed in accordance with internationally recognized standards for a definitional concordance between two classifications. Most notably, the table shows every occupation in both the NOC 2016 and the 2018 SOC. Additionally, the table describes any overlap between the two systems when one-to-many and many-to-one matches are identified.
This correspondence table, which identifies and standardizes the relationship between the Canadian and American occupational classification systems, is essential to leveraging the O*NET system to derive Canadian skills data. Indeed, ESDC used STC’s concordance as the basis of their own for occupational profiles.
Leveraging Skills-Related Insights: ESDC’s Concordance
The first step in creating the ESDC concordance was to identify equivalent occupational categories by comparing common SOC and NOC elements — typically job titles and main duties. For example, some internationally recognized occupations require formal certifications (e.g., Chemist, Dentist, Accountant, Plumber, etc.). In these cases, the occupation title is the main criteria for linking the 8-digit SOC with the 5-digit NOC. In other cases, a NOC–SOC concordance was established even though the tasks in O*NET were not all reflected in the main duties of the NOC. For example, a one-to-one concordance was established between Financial examiners and inspectors (NOC 1114.2) and Financial Examiners (O*NET-SOC 13-2061.00), despite the limited information provided in the Canadian profile.
When job title and/or main duties were not sufficient, a common understanding of the occupations, using the tasks described in O*NET, was established. For example, Babysitters (NOC 4411.1) and Journalists (NOC 5123.0) were associated respectively with Nannies (O*NET-SOC 39-9011.01) and Reporters and Correspondents (O*NET-SOC 27-3022.00).
As a result, 342 of the 939 5-digit NOC categories were associated one-to-one with an 8-digit O*NET-SOC code. As shown in Table 2, this accounts for 36% of all occupation groups, or approximately 45% of the workforce. That left two-thirds of NOC occupations without a one-to-one correspondence. Of these, 135 NOCs (14%) are associated with multiple O*NET-SOCs, and 400 cases (43%) have multiple NOCs associated with at least one O*NET-SOC. Details about how these important cases of one-to-many and many-to-one linkages were determined are discussed below.
Finally, for 62 5-digit NOCs there is no clear equivalent in the SOC, accounting for 7% of all 5-digit NOCs or about 10% of the workforce. Work is ongoing to link these remaining NOCs to SOC codes accurately.
Table 2: Overview of ESDC’s Concordance
|Type of Concordance||Occupation Coverage
|One SOC to many NOCs||43||400||25|
|Many SOCs to one NOC||14||135||20|
|Concordance not yet established||7||62||10|
* Estimates based on Census 2016
Note: The concordance is based on the current NOC-2016 structure, as such, the number of occupational profiles may change with the next revision to the NOC (NOC-2021).
For the 400 instances where one O*NET-SOC code maps to several 5-digit NOC codes, the skills profile of the single SOC in question was applied consistently across the many NOC codes. Subsequently, for each NOC in question, an industry expert reviewed the skills profile to ensure that it was a reasonable fit.
For those instances where several O*NET occupations map to one NOC, we explored several approaches. First, we considered whether a weighted average of the various skills profiles could apply to the NOC based on US employment data by occupation. After a systematic review, however, the differences in how employment statistics are gathered were too great to make this work. Instead, we looked at these mismatches on a case-by-case basis. Only O*NET-SOC codes with strong linkages to a single NOC code were retained from the potential list.
In choosing from among the many 8-digit O*NET-SOC occupations, we chose by tasks most resembling the NOC in question. In some instances, this still left a many-to-one relationship. In these cases, the average of the SOC skills profiles were applied to the NOC.2
Finally, in the few instances where no immediate match was discernible, work is underway to investigate how best to close this gap.
The Way Forward
These concordances — along with our two most recent LMI insight reports on the benefits and limitations of using O*NET in the Canadian context (no. 31) and drawing insights on skills from online job postings (no. 32) — will enable our stakeholders to generate and access skills-based insights, inferences and data in Canada. The release of these concordances also provides an opportunity to validate our methods in devising them.
Our aim remains to develop a pan-Canadian system that links skills to occupations to help reveal the changing nature of jobs. Developing skills profiles for Canadian occupations is one of several priorities in ESDC’s Occupational and Skills Information System (OaSIS) project. Using existing comprehensive sources of occupational data — such as adapting O*NET database information to the Canadian context — is the most efficient way to provide publishable Canadian skills profiles in a reasonable timeframe.
OaSIS abides by a set of principles and guiding framework, including the following: 1) ensuring that every occupational profile developed adheres to the skills taxonomy; 2) developing a new Canadian rating guide to measure and assess skills descriptors and help transpose O*NET data to the Canadian labour market context; and 3) informing and maintaining profiles via a network of stakeholders, including those involved in other national skills-related initiatives.
This work must ultimately help Canadians make more informed decisions using the skill categories, descriptions, indicators and other occupational information. To that end, later this year we will assess whether surveying employers directly regarding their skills needs could complement the methods evaluated thus far. As well, we will review international examples and best practices and their potential influence on the development of OaSIS.
This LMI insight report was prepared jointly by the staff of the Labour Market Information Council, Statistics Canada (Centre for Labour Market Information) and Employment and Social Development Canada (Labour Market Information Directorate).
Your feedback is most welcome. We invite you to provide your input or share your views on how best to approach our work on mapping skills to occupations by emailing us at email@example.com.
- O*NET offers two clear metrics for skills: each occupation is associated with an importance rating (1 to 5) and a level rating (1 to 7) to identify the complexity of the skill required.
- ESDC is currently analyzing the impact of giving greater weight to the SOC skills profile that is a better fit.