Who is the most typical Albertan, really?
By: Jared Wesley and Evan Walker
Over the last two decades, many people have been asking: What does it mean to be Albertan?
In this post, we explore the difference between the mythical “Average Joe” Albertan, on one hand, and the most commonly-found Albertans, on the other. This means going beyond our earlier work on stereotypes to examine Census data.
Our findings suggest the modal (most common) Albertan is not “Joe,” after all.
TL;DR:
According to Census data, the most common Albertans are blue-collar workers in Calgary: a millennial single mom (“Jane”) and a single, Gen Z man (“Connor”).
Jane and Connor are not all that common. There are fewer than 4,000 people like them living in Alberta.
Expanding our view of what it means to be “typical” means including a wider range of Albertans. This opens up new ways of thinking about provincial politics.
Who is typical?
The “average voter” figures prominently in democracies like Alberta. Consciously or not, citizens and politicians have a persona in mind when they think about what it means to be a typical member of their community. This image helps us make sense of the world, distinguishing the sorts of people who are in “the mainstream” from those who are not.
Politically, this persona defines what it is acceptable to say, do, or think in the public sphere. If the average person doesn’t think a policy is palatable or feasible, we’re unlikely to see value in pursuing it – regardless of how popular or necessary the measure might be.
For politicians, personas frame the “Overton Window,” the realm of popular and possible policy solutions they are willing to accept. Some politicians take it one step further, creating personas to shape the political landscape. Ralph Klein famously conjured up “Martha and Henry” as his pair of “severely normal Albertans,” for instance.
So, our vision of the “average person” matters politically. But just how accurate are our perceptions?
Who is “Average Joe” Albertan?
Much of our work here at Common Ground revolves around identifying the dominant image that comes to mind when we ask our research participants to “draw us an Albertan.”
The portrait foremost in our participants’ minds is surprisingly common. Most conceive the “typical Albertan” as a middle-aged man working a blue-collar job in agriculture or the oil patch, living in a rural area or small town, often with kids. When asked to name the person they picture, the most common response is “Joe.”
Putting themselves in his boots, our participants have helped us understand Joe’s worldview: he dislikes politics, believes hard work is the key to success, sees family as the foundation of society, and senses that people like him are “falling behind.”
But for all these drawings and surveys, our team was curious about one outstanding question: who exactly is the most commonly-found Albertan, and how numerous are they? Moving beyond the stereotype, what are the most common personal traits in Alberta, and how many people share them? For these answers, we need to go beyond our Viewpoint Alberta surveys to look at Census data. We built our approach on a similar project conducted in the United States.
Finding the Modal Albertan
To find the “average” Albertan, we need to identify the most commonly-found sets of characteristics in the population. In technical terms, we are looking for modal responses - the most common answers to a series of questions on the Census.
For instance, according to the 2021 Census, the most commonly-found Albertan is a Millennial. There are more people in that cohort than any other. In simpler terms, if you were forced to guess the age of the next Albertan you met on the street, you’d be wise to bet on it being someone in their thirties.
We could end our inquiry there and conclude that “Joe” is a Millennial, but that would leave a lot of unknowns. Where does the typical Albertan live, what does he do for work, and is he even a man? Unfortunately, there are limits to how much we can learn about the average Albertan. The Census only contains certain questions, for instance, and most only capture demographic information. More fundamentally, the more variables we add, the more difficult it becomes actually to identify “Joe” in the wild. With each additional variable, fewer and fewer Albertans satisfy the modal criteria.
Take our millennial Albertan, for instance. Let’s say we also want to know how many children they have. We would introduce the modal response for the number of children, which happens to be one or more. Suddenly, we have far fewer “Joes” who are both a millennial and have at least one child. Adding another category, like whether they live in an urban or rural area, further reduces the number of people in our “average” group. Taken far enough, we might end up with the most commonly-found Albertan, but the irony is this person is not that common at all.
Finding “Joe”
To keep our analysis manageable, we established three rules. First, we limited our analysis to the variables that we prompt our focus group participants to consider in their caricatures of “Joe” Alberta. These include income, education, marital status, occupation, age group, religion, and whether or not they have kids. Second, our variables must be dichotomous or discrete—in other words, coded so that the person either has that trait or does not. A good example is our education variable; this has been coded so that a respondent either has completed a bachelor's degree or not. Others, like income, were given three categories: $0-50,000, $50,001-120,000, and $120,000 and over. Third, we limit our analysis to individuals over the age of 18. See Table 1 for a full breakdown of our variable coding.
Table 1. Variable Coding
To find the average Albertan, we examined the 2021 Canadian long-form Census microdata; it contains basic demographic information for 112,878 Albertans. In 2021, this accounted for about 2% of Alberta’s total population but contains sufficient observations to make generalizations. The Census microdata is also the largest dataset available that contains all the information we need to provide a comprehensive sketch of the average Albertan. The basic frequencies for our variables of interest were then taken, reporting the modal responses as follows (Table 2)
Table 2. Modal Response Categories
Using these modal responses, we filtered the data according to whether an individual fit the most common response for each variable. To be considered one of our penultimate “modal” Albertans, an individual would have to share the most common characteristic across nine different traits. So, what did we find?
Discovering “Jane”
Right away, we noticed that “Joe” Alberta is actually “Jane,” as the modal response to gender identity tilts slightly in favour of women. The rest of our categories mirror Table 2: “Jane” is a millennial living in Calgary, earning under 50,000, is single and has at least one child, works a blue- or pink- collar job, and adheres to an Abrahamic religion (Christianity, Judaism, or Islam).
Only 647 individuals out of our original 112,878-person sample, or just over one-half of one percent (0.6%) fit his description. Extrapolated to the entire provincial population, and we find that there are 3245 “Janes” in Alberta.
However, Jane is not really a “real” person in that she is a composite of several different modal attributes. She is a collection of modes rather than a pure cross-section of the population.
Discovering “Connor”
With this in mind, we tacked a different course to try and find a more accurate picture of our modal Albertan. To do this, we took the same traits from Table 1, with the exception of whether or not the respondent has kids (only because our software only allows for a frequency table so large) and created an 8-way cross-tabulation. From here, it was only a matter of looking for which cell contained the largest number of people.
As it turns out, the largest group consists of slightly over 500 single men from Generation Z in our 112,878 person sample. Much like Jane they are very few in number. This modal Albertan works a blue-collar job, does not have a bachelors degree, earns less than $50,000, practices an Abrahamic religion, and, like Jane, lives in Calgary. We name him “Connor.”
Connor’s younger age and urban environment suggest that his political beliefs, lifestyle, and priorities may deviate significantly from Joe.
Clusters of Typical Albertans
Perhaps a better way to think about the most “common” Albertan is by conceiving ourselves in clusters. In statistics, cluster analysis groups observations into categories based on their association with each other. The basic objective of cluster analysis is to find groups of individuals who share similar characteristics compared to other respondents in the sample. For example, Cluster A consists of individuals who share similar traits, creating a relatively homogeneous group. However, compared to Cluster B, this group is distinct due to differences in a few key traits (more on this below).*
Table 3. Clusters of Albertans
We identified several distinct groupings of Albertans, but few of them capture the essence of the stereotypical “Joe” Albertans tend to picture (Table 3). For example, in Cluster 1, we find a fairly typical professional healthcare worker in Alberta: She is a Canadian by birth, Catholic (or at least raised Catholic), has a household income of around $110,000, has at least one kid, and lives with a partner. In Cluster 5 we find perhaps the closest articulation of “Joe” in that we find a Canadian-born “blue-collar” worker in the technical trades; however, he is secular and living in Calgary— a contrast to the various renditions of “Joe” in the foothills or wheat fields common across our focus group sessions. We also find a cluster of Albertans that deviate from conventional imagery, featuring newcomers of South Asian heritage who work in the sales and service support industry. Others can also be found in Table 3 and accompanying photos.
In the end, when it comes to common clusters of Albertans, much like our initial approach, the conventional imagery of “Joe” is noticeably absent.
Concluding Thoughts
In sum, Albertans’ image of themselves is seldom accurate. If someone were to randomly pick from a crowd of Albertans, they would be highly unlikely to choose someone who fits the conventional “Joe” stereotype.
This perception gap has real-world implications. Their attitudes and behaviours change when Albertans see themselves as someone they’re not. If they picture “Joe” as being the typical Albertan, they’re more likely to prioritize issues that matter to him, and to defer to his judgment on what’s politically necessary or feasible. Like people in other democracies, Albertans like to “go along to get along.” This means remaining silent about issues that are important to them if they don’t think their views are in the mainstream. This is detrimental to vibrant democratic discourse.
More deeply, the “Joe Albertan” stereotype downplays the amount of diversity that exists in this province. To the extent that the vast majority of Albertans don’t look like “Joe” – and only a fraction resemble “Jane” or “Connor” – these personas create the illusion that the province is more homogeneous than it actually is. This, in turn, creates an atmosphere where people who don’t fit the stereotype feel less attached to the provincial community and a weaker sense of belonging.
Findings like these should encourage those seeking to build a more pluralistic Alberta society to find new, more inclusive ways of telling Alberta’s story. This does not mean erasing “Joe” from the narrative. Nor does it mean building Alberta around “Jane” or “Connor.” Rather, it means contextualizing them as a few of the many faces of Alberta society.
*We ran a cluster analysis of our Census data on all Albertans using a machine-learning algorithm that groups categorical data based on their modal frequencies. Modal clustering is a bit different than standard clustering techniques that operate based on distance metrics (usually averages) by instead clustering the most frequent values returned in each cluster. Since it is a machine learning algorithm with the ultimate objective of configuring clusters in such a way as to minimize the dissimilarity of responses in each group, it is entirely possible to get different clusters based on how the algorithm configures the data to minimize such distance. Indeed, it may take multiple attempts. To allow for a degree of precision, we allowed our algorithm to predict 500 ‘runs’ of the data to find the optimal clustering of the Albertan census data and asked for a return of 8 clusters of Albertans based upon demographic traits (Table 3).