·  by Adeen Rizwan and Tyler Morris - 2017 Steelworker Summer Students, New Media & IT Department

Report: Are the Steelworkers Trying to Recruit You?

Today, the vast majority of workers struggle to make ends meet at the centre of Canada’s worst ever income gap, where average CEO earnings are 180 times the income of an average employee.[1] In this kind of work environment, it’s easy to imagine that more people might be seeking the protection of unions to push back against the exploitation. For low-wage, precarious workers who are younger, more often women, and disproportionately from Black, Indigenous, racialized and other minority groups, the need for union protection and support is even greater.

However, union membership isn’t growing. In fact, union density, the percentage of workers covered by a collective agreement, is declining among private sector workers in North America. Scholars, researchers and union organizers agree that unbalanced and outdated labour legislation that favours corporations is largely responsible. Such laws make it very hard for people to join a union.

So we asked: what kind of organizing intervention might help shift this decline in union density?

A common challenge in private sector unions is that their membership demographic is mostly made up of middle-aged, Anglo men. The purpose of this project was to discover the demographics of workers in new member organizing campaigns run by the United Steelworkers of Canada (USW). This information could be used by the union to determine if they should be doing more outreach to certain marginalized and equity-seeking groups, and help them analyze which communities of people are generally more willing to unionize and willing to help with campaigns.

Demographic targeting is used by corporations like Netflix and Spotify in company decision-making to help guide their choices on which new television shows or music to invest in. Likewise, demographic targeting is used by progressive electoral campaigns, to help them identify ridings where they can mobilize to defeat their better funded opponents.

When workers start a campaign to join a union, interested workers fill out a card with their information, including personal information like their name, address and contact details. Organizers also track information about the campaign itself, such as which workers said they would help, who actually volunteered to help with the campaign, who was part of the core inside campaign team and which workers voted. This information is stored in separate spreadsheets, organized by campaign and then filed away.

The worker data that organizers diligently collect should not be forgotten simply because those first campaigns have ended. They contain very useful information that could help determine if they should be doing more outreach to certain minorities and other communities of vulnerable workers. They can also help answer the question: does this public image of unions as old, white men align with the organizing goals of the union?


The first step in this project was to create a sample of USW union campaigns that were won, lost or no longer active. We found 362 for use in this study, ranging from 2000 to 2016 and including alloy welding companies, bookstores and security firms. Together we collected spreadsheets, cleaned up the information for consistency and transferred the data stored in these separate spreadsheets into one big Microsoft Access dataset. This way, all the information from all those campaigns could be analyzed at once, giving us a total of 20,516 individual people who worked at non-union job sites that were targeted by the USW for unionization.

The database was then sent for analysis to Strategic Communications Inc. (Stratcom), a consulting firm that provides data analysis services. Stratcom then ran the data through an algorithm created by Environics Analytics, to determine the gender and ethnic make-up of individuals in our dataset. Using a list of more than one million individuals with their known genders, the algorithm uses first names to analyze the gender of workers with 70% accuracy. To analyze ethnicity, the algorithm uses Census SM Research to create an ethnic surname candidate list. Then the last names of members were used to determine their ethnicity.


The data analysis showed that the majority of individuals targeted for unionization were men. With 3,027 individuals having male names with 100% accuracy and 7,352 individuals labeled as having male names with 70% accuracy, 69% of workers are male.

Women made up 31% percent of the total number of individuals whose names were gendered, with 994 people identified as having female names, and 3,577 female names with 70% accuracy.

In regards to ethnicity, most workers have an Anglo or unidentifiable ethnic origin with 11,719 people in that category. This means 57% of individuals are either Anglo or had an unidentifiable ethnic last name. In second place, we have 1,311 people of Chinese origin and in third we have 1,304 people that come from a Punjabi origin. Other common ethnicities were Irish, Sri Lankan, Vietnamese, French and Hindu. This shows that 43% of workers have last names that algorithm found to be ethnic.


With the information provided and analyzed, we concluded that the stereotype of most union members being male and Anglo is mostly true for the people USW campaigned to. While more than half of the individuals fit this category, there are still a large number of people who do not. A third of its potential membership base consists of women, and little less than half is made up of people of non-Anglo ethnic backgrounds.

Some of the campaigns analyzed were ones that the union had lost. Often, immigrants, women and other minority groups are scared to unionize and continue to endure harsh working conditions because they are unaware of what unions do. “Visible minorities actually tend to generally be more pro-union, once they get over any pre-conceived notions about the unions,” said one of the USW organizers. To address this issue, the union could be doing more to appeal to them directly, such as having more multi-lingual and female organizers, translating flyers into different languages and catering directly to the needs of individuals in these groups. Not only would this benefit the workers, but this would ensure that more people would support unions and be willing to help in campaigns.

Due to time restrictions, only gender and ethnic origin could be analyzed. Also, the group labelling for Statistics Canada census data results in limitations due to its confusing roster of ethnic, racial, cultural and language group categories. These categories are sometimes additionally confused with region of origin and religion. In the future, this database and method could be used to analyze other factors as well, including age, type of dwelling, income, commute from work, etc., which would provide further information to the union about who the USW is targeting in their unionization campaigns and how to best provide assistance. In the future we expect that alternate standards and categories will need to be applied to yield reliable results.

About the authors:

Tyler Morris is a 17 year old student, currently in his last year of high school. He is interested in data, analytics, digital card signing and has a lot of experience working with Microsoft Excel and Access. He plans to study software engineering.

Adeen Rizwan is a second year mechanical engineering student at the University of Toronto. She aspires to be a leader in the realms of science and technology. Her expertise includes data analysis, computer programing, statistics and all things math.


[1] Hugh Mackenzie (Jan. 2016) Staying Power, CEO Pay in Canada, Canadian Centre for Policy Alternatives.


Press Inquiries

Thank You

Thank you for contacting us. We will try to get back to you in a timely manner.

Media Contacts

Communications Director:
Bob Gallagher
416-544-5966 or

Communications Department:
Karina Midence

Communications Department - Québec
Clairandrée Cauchy

Mailing Address

United Steelworkers
234 Eglinton Ave. E., 8th Floor
Toronto, ON M4P 1K7