Insights on Canadian Society
What is the connection between working from home and the languages used at work?

Start of text box

Overview of the study

Using 2021 Census data, this article examines the link between working from home and the languages used at work. It focuses on three Census metropolitan areas (CMAs) of Moncton, Montréal and Ottawa–Gatineau—three regions where both English and French are used widely at work.

• In 2021, in the CMAs of Moncton, Montréal and Ottawa–Gatineau, people working from home were generally more likely than other workers to use English as their main language at work. They were also more likely to use English and French on a regular basis.
• People who work from home differ from other workers in many respects. For example, they are more likely to work in certain industries and occupations, where the use of English is the main language or the combined use of English and French are more common (e.g., professional, scientific and technical services; finance and insurance; or the information industry).
• Differences in job characteristics (industry, occupation, etc.) and the language characteristics of workers largely account for the differences in language use by type of workplace (i.e., home or elsewhere).
• However, accounting for these factors does not completely explain why people who work from home in Moncton, Montréal and Ottawa–Gatineau use English as their main work language more often than other workers.

End of text box

Introduction

During the COVID-19 pandemic and the associated public health measures, the proportion of people who usually worked from home increased significantly. Almost 1 in 4 Canadian workers (24.3%) worked from home at the time of the 2021 Census, compared with less than 1 in 10 workers (7.4%) at the time of the 2016 Census (Chart A.1 in the appendix).

The work-from-home rate has declined slightly since then, but it remains well above pre-pandemic levels. Data from the Labour Force Survey indicate that 20.1% of workers usually worked from home in the spring of 2023.^Note 

Working from home can take many different forms. For example, it includes telework (i.e., remote work based on the use of information and communication technologies) and various types of self-employment, as well as working for a business or farm with the same address as a person’s home. That said, the growth in working from home during the pandemic is primarily attributable to an increase in telework.

People who work from home have a different reality from other workers. The transition to telework, caused by the pandemic, changed the daily work experience of millions of workers in Canada. However, the effects of working from home on different aspects of work, including languages used at work, remain unknown. Until this point, the relationship between working from home and languages of work has not been studied in Canada.

Compared with onsite work, telework affects the frequency and manner of communication with colleagues or clients. This raises the question of whether these changes in communication have an impact on the languages used at work. In addition, studies have shown that the use of languages at work in a multilingual setting depends on a set of elements relating to the nature of the tasks and the work environment.^Note  It can, therefore, be assumed that telework is a work environment that is sufficiently different from onsite work to have an impact on language use.

Telework also enables long-distance employment situations where a person works for an employer that is located in another country, another province or just farther away within the same province.^Note  According to the Labour Force Survey, as of June 2022, approximately one in five employees who worked from home were working for an employer located outside their region of residence, that is, a region they could not travel to daily.^Note  This could, in this case, have an impact on the languages these workers use at work. For example, it can be assumed that a person residing in Montréal but working for an employer based in Ontario or the United States would be more likely to work in English than their counterparts working onsite locally.

This may result in different language usage by workplace type, and this appears to be supported by data that show that in the Montréal census metropolitan area (CMA), home-based workers (32.8%) were twice as likely to mainly use English at work, compared to those who were not working from home (16.8%).

However, this does not mean that working from home in itself has an effect on language use. In fact, people working from home are different from other workers in terms of language and job characteristics. There is a greater proportion occupied in some industries and occupations where the use of English and the combined use of English and French are more common.

Using 2021 Census data, this article examines the link between working from home and languages used at work. The article focuses on three CMAs: Moncton, Montréal and Ottawa–Gatineau. These regions were chosen because both English and French are used extensively at work (see the text box “Data sources, methods and definitions” for more details).

Languages used by workers vary by workplace type

Overall, in 2021, people working from home (81.7%) used English as their main work language slightly more often than other workers (75.6%) (Table 1). They also more often used a combination of English and French on a regular basis at work (regardless of the language used most often). This was true for both Canada as a whole and the CMAs of Moncton, Montréal and Ottawa–Gatineau, although to varying degrees in each of these regions.

In Montréal, of those working from home, 56.3% mainly used French at work, compared with 74.8% of other workers. At the same time, 32.8% of home-based workers mainly used English, compared with 16.8% of other workers. Furthermore, a greater proportion of home-based workers used both English and French on a regular basis at work (47.7% versus 35.0% of other workers). In short, the use of English as the main work language was significantly more common among those working from home than among the rest of the workers, as was the use of a combination of English and French.

The findings were similar for Ottawa–Gatineau, where 10.0% of people working from home mainly used French at work (compared with 22.1% of other workers), 84.5% mainly used English (compared with 73.3%), and 31.1% used both English and French on a regular basis (compared with 23.4%).

While these findings were somewhat less pronounced in Moncton, the same pattern emerged: 14.1% of those working from home mainly used French at work (compared with 16.5% of other workers), 77.9% mainly used English (the same proportion as other workers), and 31.8% used both English and French on a regular basis (compared with 27.0% of other workers).^Note 



Work from home varies according to job and language characteristics of workers

People working from home differed in various ways from other workers, which may explain their different language practices at work. First, people working from home are concentrated in certain occupations and industries where this type of work arrangement is possible. For example, in 2021, more than half of people worked from home in professional, scientific and technical services (57.8%); finance and insurance (57.7%); and information and cultural industries (54.9%) (Chart 1). Conversely, this was the case for about 1 in 10 people in the transportation and warehousing (10.6%), construction (9.7%), and retail trade (9.2%) sectors, and only 1 in 20 people in the accommodation and food services sector (5.3%).

Data table for Chart 1



Data table for chart 1
Table summary
This table displays the results of Data table for chart 1. The information is grouped by Major industry sector (appearing as row headers), 2021, calculated using percentage units of measure (appearing as column headers).

Major industry sector	2021
	percentage
Total	24.3
Accommodation and food services	5.3
Retail trade	9.2
Construction	9.7
Transportation and warehousing	10.6
Health care and social assistance	11.9
Manufacturing	12.0
Mining, quarrying, and oil and gas extraction	17.8
Other services	23.0
Wholesale trade	26.1
Educational services	30.4
Utilities	32.4
Arts, entertainment and recreation	33.6
Public administration	38.3
Real estate and rental and leasing	39.2
Agriculture, forestry, fishing and hunting	41.3
Information and cultural industries	54.9
Finance and insurance	57.7
Professional, scientific and technical services	57.8
Source: Statistics Canada, Census of Population, 2021.

The use of languages at work varies greatly depending on occupation and industry (Appendix Table A.1). For example, in the CMA of Montréal in 2021, 34% of workers in the professional, scientific and technical services sector and 32% of workers in the information industry worked mainly in English, compared with 10% of those in construction and 13% of those in health care and social assistance. Therefore, could it be that the differences between the language practices of people working from home and other workers depend on the characteristics of their jobs?

A similar question can be raised about the other language characteristics of workers (language knowledge, languages spoken at home, etc.), which are also linked to their language use at work. For example, in the Montréal CMA in 2021 (Table 2), the work-from-home rate was higher among workers who were English–French bilingual (30.5%) or who knew English but not French (28.7%), compared to workers who knew French but not English (11.9%). Beyond knowledge of languages, there were also differences depending on languages spoken at home. Among unilingual workers, those who knew English and spoke English most often at home^Note  (34.6%) were significantly more likely to work from home than those who knew French and spoke French more often at home (12.3%). Among bilingual workers, those who spoke English most often at home worked more often at home (35.5%) than those who spoke French more often (29.4%) or another language (26.4%).



These differences in language characteristics may themselves be related to other factors. For example, bilingual workers tend to be more concentrated in certain occupations and industries where telework opportunities are more common (e.g., professional, scientific and technical services; finance and insurance; and public administration). There are also differences in the spatial distribution of language groups in metropolitan areas.^Note  These differences can result in different relationships to commuting, and by extension, to telework: telework would in principle be more advantageous when commuting times are longer.

Furthermore, it is also possible that language skills play a direct role in the probability of teleworking, particularly with respect to working for an employer located outside a person’s area of residence. For example, for a resident of the CMA of Montreal, becoming fluent in English could broaden remote work opportunities to jobs located in other provinces or other countries. As well, having limited French skills could hinder integration into the local labour market and make working remotely comparatively more attractive.

Job and language characteristics do not entirely explain why people who work from home use English more often as their main language of work

Regression models are used to control for the potential differences between teleworkers and other workers, first with respect to the characteristics of their job, then their language characteristics and finally various other sociodemographic characteristics (for more details, see the text box “Data sources, methods and definitions”). An analysis using 2016 Census data produced results that were generally consistent with the 2021 results.^Note 

Charts 2 to 4 present the “unadjusted” rates of language use by type of workplace for each CMA, as well as a series of adjusted rates produced by controlling for different individual characteristics. Chart A.2 in the appendix presents the results for Canada as a whole.

In Montréal, in 2021, a smaller proportion of people who work from home (56.3%) mainly used French at work than people who usually work at another location (74.8%) (Chart 2). After accounting for a set of job-related characteristics (industry sector, occupational category, class of worker and number of hours worked per week), the proportion of people who mainly used French at work increased to 63.2% among those working from home and decreased to 72.7% among other workers. These adjusted rates correspond to what might be expected if these two groups of workers did not differ in terms of the characteristics included in the model. In other words, in Montréal, differences in job characteristics accounted for almost half of the difference in the use of French as the main language of work between teleworkers and other workers. The gap shrank from 18.5 percentage points to 9.5 percentage points when these characteristics were considered. Industry sector and occupation accounted for most of this decrease, while the class of worker (being an employee, self-employed, etc.) and the number of hours worked had little impact on the results.

When both job differences and language characteristics (knowledge of official languages, mother tongue and language spoken most often at home) were accounted, the gap in the use of French as the main work language between tele workers and the rest of workers narrowed further (difference of 5.3 percentage points). However, there was virtually no additional effect on the gap (difference of 4.9 percentage points) when certain sociodemographic characteristics were considered, such as age, gender, immigration status and period, education level, residential mobility in the past year, and municipality of residence.

In short, taking into account certain worker characteristics explains much of the difference between people working from home and other workers in the main use of French at work in Montréal. However, this does not fully explain the observed differences.

A similar picture emerged with respect to the use of English as the main work language in Montréal: differences in employment and, to a lesser extent, language characteristics partly explain the differences by workplace, but a gap remains. With respect to English–French bilingualism (use of both languages on a regular basis at work), there was no longer a gap between the two groups of workers when all characteristics were taken into account. Employment characteristics explain most of the difference.

Data table for Chart 2



Data table for chart 2
Table summary
This table displays the results of Data table for chart 2 Work from home and Other workplaces, calculated using percentage units of measure (appearing as column headers).

	Work from home	Other workplaces
	percentage
Mainly French
Unadjusted rate	56.3	74.8
Adjusted rate—job characteristics	63.2	72.7
Adjusted rate—job and language characteristics	66.4	71.7
Adjusted rate—job, language and sociodemographic characteristics	66.7	71.6
English and French regularly
Unadjusted rate	47.7	35.0
Adjusted rate—job characteristics	40.8	37.3
Adjusted rate—job and language characteristics	38.6	38.1
Adjusted rate—job, language and sociodemographic characteristics	38.1	38.3
Note: For more information on the rate adjustment method, see the "Data sources, methods and definitions" box. Source: Statistics Canada, Census of Population, 2021.

For Ottawa–Gatineau (Chart 3), the results were similar to those for Montréal. Differences in the use of English and French as the main work language were largely explained by differences between the two groups of workers. However, gaps remained, though they were smaller than those noted for Montréal. The greater propensity of people working from home to use both English and French on a regular basis was fully explained by the characteristics of the jobs held.

Data table for Chart 3



Data table for chart 3
Table summary
This table displays the results of Data table for chart 3 Work from home and Other workplaces, calculated using percentage units of measure (appearing as column headers).

	Work from home	Other workplaces
	percentage
Mainly French
Unadjusted rate	10.0	22.1
Adjusted rate—job characteristics	12.8	19.4
Adjusted rate—job and language characteristics	16.1	17.9
Adjusted rate—job, language and sociodemographic characteristics	16.5	17.7
English and French regularly
Unadjusted rate	31.2	23.4
Adjusted rate—job characteristics	26.5	26.4
Adjusted rate—job and language characteristics	26.9	26.2
Adjusted rate—job, language and sociodemographic characteristics	26.7	26.3
Note: For more information on the rate adjustment method, see the "Data sources, methods and definitions" box. Source: Statistics Canada, Census of Population, 2021.

In the Moncton area (Chart 4), where the differences in language use by workplace were, initially, less pronounced than in Montréal or Ottawa–Gatineau, the situation was somewhat different. The slightly lower propensity of people working from home to use French as their main work language was primarily explained by the characteristics of the jobs they held, with language and sociodemographic characteristics having little or no impact. Moreover, while the unadjusted rate of English–French bilingualism was higher among home-based workers (31.8%) than among other workers (27.0%), this situation was reversed when the differences between the two groups of workers were taken into account. The adjusted bilingualism rate was slightly higher among those not working at home (28.4%) than those working from home (26.6%) when job characteristics were taken into account. The gap between the two groups widened slightly when language characteristics were taken into account.

Data table for Chart 4



Data table for chart 4
Table summary
This table displays the results of Data table for chart 4 Work from home and Other workplaces, calculated using percentage units of measure (appearing as column headers).

	Work from home	Other workplaces
	percentage
Mainly French
Unadjusted rate	14.1	16.5
Adjusted rate—job characteristics	15.4	16.1
Adjusted rate—job and language characteristics	15.3	16.2
Adjusted rate—job, language and sociodemographic characteristics	15.2	16.2
English and French regularly
Unadjusted rate	31.8	27.0
Adjusted rate—job characteristics	26.6	28.4
Adjusted rate—job and language characteristics	26.3	28.5
Adjusted rate—job, language and sociodemographic characteristics	26.3	28.5
Note: For more information on the rate adjustment method, see the "Data sources, methods and definitions" box. Source: Statistics Canada, Census of Population, 2021.

Conclusion

In 2021, there were some differences in the use of languages at work between people working from home and other workers in major Canadian metropolitan areas where English and French coexist in the labour market. People working from home tended to use English more often at work as their main language, and to use both English and French on a regular basis more often. These patterns were very pronounced in Montréal and Ottawa–Gatineau, but slightly less so in Moncton.

With respect to the combined use of French and English on a regular basis at work, the differences by type of workplace can be largely explained, if not entirely explained, by differences in the characteristics of the jobs held by people working from home and those working elsewhere. Differences in the use of English or French as the main language of work is largely explained by differences in job characteristics, but also in the language characteristics of individuals. However, gaps remain, particularly in Montréal. This means that other factors, not considered in this study, should also be considered to explain why people working from home use English as their main work language more often than other workers.

To explain the remaining differences in the use of languages at work by workplace type, two main assumptions, which are possibly complementary, can be made. The first assumption would be that telework is distinguished from onsite work by the methods of communication at work. Since distance communication (with colleagues, clients, partners, etc.) is different from face-to-face communication (technological mediation,^Note  fewer informal discussions, etc.), this would have an impact on the use of languages at work in certain situations. However, it remains to be explained how these differences would, for example, promote the use of English as the main language rather than French. Changes in work arrangements associated with the pandemic context may also have resulted in changes in language use rates at work.

A second assumption would be that there are differences in the employers of people working from home. In particular, a portion of people working from home would telework for organizations based in another region or country, or even for fully remote organizations. It is likely that these workers would be more prone to working mainly in English than their counterparts working for employers based “physically” in major cities with a large Francophone population, particularly in Montréal. In future studies, data that provide additional information on employer characteristics could be used to test this assumption.

Future trends in working from home are difficult to predict. Regardless, it will be important to be able to distinguish between different work-from-home arrangements: telework, hybrid work arrangements (in which workers divide their time between work onsite and at a distance), self-employment, long-distance work for an employer located in another region, etc. The aim will be to monitor changes in these different situations and to better understand their implications for various facets of the reality of workers, such as the use of languages at work. Furthermore, the links between working from home and other language dimensions—such as languages spoken at home and the spatial distribution and mobility of language groups—should be studied.

Louis Cornelissen is an analyst at the Centre for Demography at Statistics Canada.

Start of text box

Data sources, methods and definitions

Data sources

The data are from the 2021 Canadian Census of Population. For more information and data from the 2021 Census, visit the Census of Population website.

Population

The analysis is based on people 15 years of age and older who were employed during the census reference week, the week of May 2 to 8, 2021.

Data for census metropolitan areas (CMAs) are for workers residing in them. Individuals may have a usual place of work that is outside their CMA of residence.

Work from home

People usually working from home are those whose usual place of work and residence were in the same building, those who lived on the farm where they worked, and teleworkers who worked from home for most of their work week.

In the case of hybrid work arrangements, where work time is shared between home and another location, census respondents should have indicated only where they worked most of the time. The inability to distinguish between different workplaces for the same person is a limitation of the data and analysis.

Languages used at work

The 2021 Census question on languages of work had two components. The first component concerns the language or languages used regularly at work. If necessary, the second component seeks to specify, from among the languages in the first component, which of these languages was used most often. For both components, multiple responses were possible.

People who use a language regularly (or on a regular basis) include all those who reported using this language regularly, even if they did not use it most often.

For English and French specifically, those who mainly used one language are those who used only one of these two languages most often at work. This excludes people who said that they used English and French equally but includes those who used English or French equally with a non-official language.

Individuals using some languages equally are those who reported using more than one language most often.

In 2021, the question on languages used at work was changed from previous censuses, reducing response burden and improving data quality. However, this had an impact on comparability with previous census data. Data on languages used most often at work can be compared with those of previous cycles, but any comparison must be done with caution and considering the effect of the change to the question. For the data on all languages used regularly at work (but not necessarily most often), they should not be compared directly with those from previous cycles.

For a detailed definition of census concepts related to languages, labour or geography, see the Languages Reference Guide, the Labour Reference Guide or the Census of Population Dictionary.

Multivariate models

To control for compositional effects and isolate the link between the workplace and language use, binary logistic regression models were used. These models seek to predict the effect of different variables on language use at work. The workplace type is introduced as an independent variable in the form of a two-category variable: work from home and other workplaces.

Three blocks of other independent variables are introduced in turn into the model, to control for various characteristics of individuals. The first block includes a set of characteristics related to the job held: industry sector (North American Industry Classification System four-digit codes), occupational category (National Occupational Classification two-digit codes), class of worker (employee, self-employed, family worker), number of hours worked per week and employment income in the previous year. The second block concerns the language characteristics, measured by a composite variable that crosses knowledge of official languages, mother tongue and language spoken most often at home, measuring familiarity with English and French. The third and final block includes a set of sociodemographic characteristics of workers: age, gender, immigration status and period, education level, interprovincial mobility in the last year, and municipality of residence.

Adjusted rates were produced to facilitate the presentation of regression model results. These should be interpreted as the probability for a person with a given characteristic to present a language use profile at the given job, with the other variables included in the model being kept at their average value measured in the sample. For example, from Chart 4 (Montréal), if one were to imagine that people working from home were distributed in the same way as all workers in terms of the different individual and job-related characteristics taken into account in the model (same distribution by age, occupation, industry, etc.), it could be expected that 66.7% would use mainly French at work (while this was the case for 56.3% of them in reality).

Three sets of models were produced for three different dependent variables: mainly using French at work, mainly using English, and using both English and French on a regular basis.

End of text box