# Indistinct Information Habitats: Information and Attitudes in the 2023 Alberta Election August 2023 ![](_page_0_Picture_3.jpeg) Aengus Bridgman, Blake Lee-Whiting, Thomas Bergeron, Thomas Galipeau, Alexei Abrahams, Saewon Park, Sara Parker, Taylor, Owen, and Peter John Loewen ![](_page_0_Picture_7.jpeg) # **Table of Contents** | Table of Contents | 1 | |-------------------------------------------------|----| | Executive Summary | 2 | | Methodology | 3 | | Survey Research | 3 | | Digital Trace | 3 | | The Alberta Election Information Environment | | | News Media Consumption | 4 | | Engaging with News and Politics on Social Media | | | Alberta Election: Issues and Attitudes | 9 | | Online Campaign Discourse and Voter Priorities | 9 | | Political Trust | 11 | | Misinformation | 13 | | Conclusion and Reflections | 16 | | References | 17 | ## Executive Summary Amidst concerns of rising polarization and increasing information environment segregation, the Media Ecosystem Observatory (MEO) conducted a two-pronged study during the 2023 Alberta provincial election to evaluate the possibility of increasing ideological and affective polarization. 12 MEO's study set out to examine the presence of distinct information environments or 'habitats' among Alberta New Democratic Party (NDP) and the United Conservative Party (UCP) by surveying 948 Albertans to understand their media consumption habits and collecting digital trace data on Instagram and Facebook to understand social media information networks. Our analysis was guided by the following questions: (1) what was the news media landscape in Alberta during the election? (2) Do NDP voters and UCP supporters consume different media? (3) What attitudes are those patterns of consumption associated with? We find that the information ecosystem is not heavily divided in terms of media consumption within the province along party lines: NDP and UCP voters do not exist in vastly different information habitats. Voters from both parties have similar news media consumption preferences and generally obtain their news from similar outlets, with the notable exceptions of CBC being favoured by NDP voters and Rebel News being favoured by UCP voters. We also find that voters from both parties use similar social media platforms. However, the digital trace data hints at the presence of distinct information habitats on social media platforms along party lines. Using a social media engagement network from comments on candidate and media outlet posts, we find suggestive evidence that a small number of Albertans – those most heavily interested in politics – do help produce a divided social online discourse. This divergence in information environments online is an important consideration, given our study also revealed that, despite commonalities in media consumption, there are large differences in attitudes. There are large differences in attitudes related to trust in politicians and political institutions, the importance given to particular political issues, and opinions about how to best respond to the democratic threat of misinformation. 1 [Affective' polarization is the extent to which people like \(or feel](https://www.zotero.org/google-docs/?5xhML9) warmth towards) their political allies and [dislike \(or feel lack of warmth towards\) their political opponents, see Iyengar, Sood, and Lelkes, "Affect,](https://www.zotero.org/google-docs/?5xhML9) [Not Ideology."](https://www.zotero.org/google-docs/?5xhML9) 2 Johnston, "Affective Polarization in the Canadian Party System, 1988–2021."; Boyd, "'The tale of two Albertas': Danielle Smith's province is already divided. Are things about to get worse?" # Methodology The [Media Ecosystem Observatory \(MEO\)](https://www.mediaecosystemobservatory.com/) is interested in the dynamics of information production, dissemination, and consumption across digital media, our goal is to empower Canadians to navigate the complexities of the modern digital age. To that end, MEO regularly collects representative survey data from Canadians across the country, complemented by robust data collection in Canadian online spaces. #### Survey Research MEO surveyed 948 Albertans between May 25th and May 29th, the last five days of the campaign period terminating on election day. The survey sample is generally representative of Albertans in terms of age, gender, and general location. Moreover, all the analyses use weights on age, gender, and education.3 Survey respondents were provided by [Dynata,](https://www.dynata.com/) a leading survey sample provider; the survey itself was administered on the Qualtrics survey platform.4 In addition to standard demographic questions, we collected information on media consumption habits, trust in institutions, opinions on political parties and leaders, how Albertans interpreted the implications of misinformation during the campaign period, and more general questions about preferences related to news media. #### Digital Trace Using official party lists, MEO collected posts published on Instagram by all UCP and NDP candidates, alongside eight Canadian news outlets operating in Alberta. All posts published by these handles during the electoral writ period (May 1 - 29) were manually retrieved, along with up to 100 of their top comments.5 MEO also collected Facebook posts using [Meta's](https://crowdtangle.com/) [Crowdtangle API](https://crowdtangle.com/) from the public pages of UCP and NDP candidates during the electoral writ period. In total, 3,150 posts and 53,877 comments posted by 22,134 commenters on Instagram as well as 2,598 posts on Facebook were analyzed. In this report, we analyze both the text content of each post caption as well as the pattern of networked engagement implied by commenter behaviour. The text content of the Facebook and Instagram posts was analyzed using zero-shot classification with a large language model, which allowed us to categorize 3 Weighting survey data is a method used to maximize the representativeness of the sample. by correcting the difference between the collected sample and the population of Alberta. This technique is used to increase the generalizability of our findings. We used an iterative proportional fitting (or raking) method to adjust the weights of the sample to align with the populations' distributions collected from the census. In this report, the weighting variables were raked according to Alberta's distribution in age, gender, and education. 4 Anonymized survey data from this project will be made freely available online. 5 The data collection occurred from July 17 to 19, 2023. Several UCP and NDP candidate Instagram and Facebook handles were no longer available. posts based on the main issue they addressed (Healthcare, Inflation, Climate Change, Economy, Housing, Education, General Campaign Support, Other). ## The Alberta Election Information Environment Canadians are said to be becoming more ideologically and affectively polarized – driven, in part, by the information they consume.6 During an election period, are the information consumption patterns of Albertans divided along partisan lines? Do voters reside in distinct information habitats? For all survey results, we examine only those respondents who indicate a desire to vote for the two leading parties: the UCP and NDP. #### News Media Consumption We asked Albertans where they go to get news online and how often they use each news website. We presented them with eight options: four are national Canadian media outlets [\(CBC,](https://www.cbc.ca/) [The Globe and Mail,](https://www.theglobeandmail.com/) [the Toronto Star,](https://www.thestar.com/) and [The National Post\)](https://nationalpost.com/), two major Albertan newspapers [\(Edmonton Journal](https://edmontonjournal.com/) and [Calgary Herald\)](https://calgaryherald.com/), and two independent (now national) Canadian media outlets with Albertan roots [\(Rebel News](https://www.rebelnews.com/) and [True North News\)](https://tnc.news/). ![](_page_4_Figure_5.jpeg) Figure 1. Albertans Online News Consumption 6 [Merkley, "Mass Polarization in Canada: What's Causing It? Why Should We Care?"; Bridgman, "The Role](https://www.zotero.org/google-docs/?broken=noz6bZ) [of Social Media in Polarizing Canadians."](https://www.zotero.org/google-docs/?broken=noz6bZ) *Figure 1* represents Albertan news consumption preferences. There are two main differences between NDP and UCP voters: NDP voters are more likely to visit CBC than UCP voters, and UCP voters read Rebel News more often than NDP voters. Aside from these differences, the media consumption habits of partisan Albertans are relatively similar across the six other news sites. The National Post, Calgary Herald, and Edmonton Journal have crossparty appeal, whereas few readers engage daily with True North News or the Toronto Star. #### Engaging with News and Politics on Social Media Accessing news online is often done through social media platforms by clicking on shared links or reading posts made directly on social media.7 Examining social media use during the Alberta election may further, and even better, capture divides in the information environment. ![](_page_5_Figure_3.jpeg) Figure 2. Albertans News Gathering Behaviour on Social Media *Figure 2* shows that Albertans rely heavily on social media platforms to follow news. The three most popular Canadian social media platforms are Facebook,8 which is also the most 7 Wojcieszak, "Avenues to News and Diverse News Exposure Online: Comparing Direct Navigation, Social Media, News Aggregators, Search Queries, and Article Hyperlinks." 8 Gruzd and Mai, "The State of Social Media in Canada 2020." used for news access9 , Youtube, and Instagram, whereas Twitter, TikTok, Linkedin, and Reddit are less popular. Very few Albertans receive news from Twitch, the live streaming service, or Discord, the voice, video, and text chat social media app. We observe no political divide on where Albertans get their news on social media: UCP and NDP voters use the same social media platforms to find news. Shared usage of social media platforms, however, may not result in shared exposure to news and political information. Even though a similar share of UCP and NDP voters may obtain their news and political commentary from a given social media platform, they may very well be receiving and consuming substantially different streams of political content. We decided to investigate this possibility further on Instagram, a platform from which 30% of Albertan voters (and 67% of young Albertan voters) report obtaining their political news. We conducted a scrape of Instagram to map the relative positionality of UCP and NDP candidates to one another, in order to detect and characterize divisions in the information habitat. Using comments on posts published on Instagram by the eight media outlets and all NDP and UCP candidates, we constructed a directed network where each node is an Instagram user (an author A or commenter C) and each edge (C, A) represents one or more comments written by C on a post authored by A. The network is visualized in *Figure 3*. 9 Bridgman et al., "Mis- and Disinformation during the 2021 Canadian federal election." ![](_page_7_Figure_0.jpeg) ![](_page_7_Figure_1.jpeg) *Figure 3* shows Network positioning UCP (blue) and NDP (orange) candidates, and news outlets (green), according to who comments on their Instagram posts. Network contains 22,205 nodes (commenters plus authors) and 24,950 edges representing 53,877 comments on 3,150 posts published during the electoral writ period of May 1-29, 2023. Larger nodes garnered more comments, with a maximum size factor of six. Network plotted using Gephi (Force Atlas 2 layout). In total, the network contains 22,205 nodes (commenters plus authors) and 24,950 edges representing 53,877 comments. But since our analytic interest is the relative positionality of candidates and news outlets, and to preserve the privacy of ordinary users who commented on the posts, we filtered the visualization to hide ordinary commenters, showing only the candidates and news outlets. For ease of viewing, we colored the news outlets green, the UCP candidates blue, and the NDP candidates orange. To highlight authors that enjoyed more engagement (i.e. more comments) on their posts, we rendered their circles (nodes) and names larger. Setting node size exactly proportional to engagement, however, would have rendered the network illegible. Instead, we substantially softened the differences, letting the largest node (Rebel News) be no more than six times the size of the smallest node.10 10 Note that as an artifact of this intervention many of the smaller nodes appear similar in size. We rendered the network in such a way that authors appear closer together whenever they share a common pool of commenters.11 For example, if an Instagram user who comments on a post by Premier Danielle Smith [\(@abdaniellesmith\)](https://www.instagram.com/abdaniellesmith/reels/) also comments on a post by the National Post [\(@NationalPost\)](https://www.instagram.com/nationalpost/?hl=en), then Smith and the National Post will appear closer together visually. This visualization strategy leads to a very specific prediction: if the information habitat is unified, then users would read and comment on the posts of *both* the UCP and NDP candidates, and all the news outlets, leading to a visualization in which blue, orange, and green nodes would be completely intermingled. Instead of being intermingled, *Figure 3* demonstrates that blue and orange nodes appear separately clustered. This spatial divide between NDP and UCP candidates is a strong sign that their readers are engaging with the rhetoric of mostly one or the other party, not both. Blue UCP candidate nodes appear clustered together, as do orange NDP candidate nodes. This kind of sorting along party lines implies that NDP candidates enjoy a common readership – a set of Instagram users who comment on the posts of multiple NDP candidates. Separately, UCP candidates also enjoy a common readership. These two readerships are distinct: the users who comment on NDP posts do not comment on UCP posts, and vice-versa. While *Figure 2* shows that Albertan NDP and UCP voters access Instagram at similar rates to each other, it is evident from *Figure 3* that they are not engaging with the same content on Instagram. Notably, however, the UCP and NDP clusters are not entirely separate from each other. In fact, both camps are 'infiltrated', with some orange nodes deep in blue territory, and some blue nodes deep in orange territory. For example, NDP candidate Janet Eremenko [\(@janeteremenko\)](https://www.instagram.com/janeteremenko/?hl=en) is surrounded by blue UCP nodes, and quite far off from her fellow orange UCP candidates. The same people who are commenting on the posts of UCP candidates are also commenting on Janet Eremenko's posts. Likewise, UCP candidate [Prasad Panda](https://www.unitedconservative.ca/nominated-candidates/prasad-panda/) appears closer to orange nodes than blue. We manually reviewed the posts and comments of outliers such as these, and found that each of these candidates authored fewer than ten posts during the electoral period, and received only a few comments per post. The low volume of posts and comments renders the candidates' positionality more volatile. The positionality of the green news outlet nodes relative to the UCP and NDP clusters likewise tells us something about their political leanings. Rebel News and TNC are s far from the NDP as possible, further even than the UCP candidates themselves. Among UCP candidates, their closest connection is Danielle Smith, suggesting that Smith's readership exists in a distinct information habitat that is driven in large part by Rebel News and TNC. The National Post, by contrast, is embedded squarely among the UCP candidates. The Edmonton Journal, Calgary Herald, Globe & Mail, and Toronto Star, lie far closer to the border 11 We plotted the network in Gephi, a popular network visualization software. We relied upon Force Atlas 2, a standard choice for network visualization. region between the UCP and NDP clusters, suggesting these outlets are read across party lines – a finding consistent with the survey data. Their positionality suggests these four outlets function as bridges between distinct information habitats. By comparison, CBC is more proximate to the NDP environment, consistent with the survey findings. The authors who enjoyed the greatest amount of engagement12 were the two independent outlets with Albertan roots, Rebel News and True North News, which garnered 39,927 of the 53,877 comments (74.1%) and 1.63 million likes (78%). Alberta's Premier, Danielle Smith [\(@abdaniellesmith\)](https://www.instagram.com/abdaniellesmith/?hl=en) garnered merely 8.2% of comments and 6.4% of likes. In total, UCP and NDP candidates garnered 10.1% and 9.3% of comments, respectively (7.1% and 10.5% of likes). That independent news outlets capture the lion's share of engagement stands in stark contrast to the survey findings depicted in *Figure 1*, where few respondents from either party indicated receiving their news on a regular basis from either of those outlets. Given that Instagram and other social media platforms prioritize content that elicit high rates of engagement, it is likely that many Albertan voters would have seen Rebel News and True North News posts in their feed. 12 *Figure 3* substantially understates the asymmetry in engagement between the alt right outlets and everyone else, because – for the sake of legibility – we artificially restricted node sizes to vary no more than a factor of 6. # Alberta Election: Issues and Attitudes Distinct information habitats are of interest because they have the potential to lead to disagreement about fundamental political facts, increased political and affective polarization, and an inability to articulate and collaborate on a shared political future. Here, we examine differences in attitudes between UCP and NDP voters. We focus on issue priority, trust in political figures, media, and elections, and exposure to and beliefs about misinformation. #### Online Campaign Discourse and Voter Priorities The Instagram data has provided an indicator that the information habitats of NDP and UCP voters may be different, despite them generally consuming information from a common set of news outlets and social media platforms. One possibility is that voters are interested in different issues and political candidates are echoing and shaping that difference. We asked voters what their most important election-related issue is, reported by party, in *Figure 4*. UCP voters chose inflation and economy as priority issues more often than NDP voters, who emphasized healthcare and climate change as priority issues more often. We categorized text captions from the candidates' Instagram and Facebook posts into these major policy issues by party using a machine-learning model for natural language processing. 13 The data covers Instagram and Facebook posts uploaded by NDP and UCP candidates during the writ period spanning from May 1, 2023 to May 29, 2023. In addition to the list of issues from the survey, we also added "elections" and "other" as categories to differentiate posts that discuss the election campaign in general from posts that discuss any other topics. During the period covered, we saw that approximately half of the candidates' posts (51% on Instagram, 39% on Facebook) that could be classified were broad calls for votes or campaign support rather than posts focusing on a specific voter issue.14 *Figure 5* shows the distribution of the candidates' posts by policy issue and party, only including the posts that focus on one of the major policy issues reported in the survey. These results show that the issues that voters of all ages see as important are reflected proportionally in the candidates' posts on Instagram and Facebook, even though Instagram's user base is skewed younger and Facebook's user base is skewed older. Issue pattern preferences are matched in the candidates' Instagram and Facebook campaigns, with UCP candidates posting about inflation and the economy in a greater proportion than NDP candidates, while NDP candidates posted about healthcare in a greater proportion than UCP candidates. An exception to this pattern is shown in the issues of "Climate Change" and "Education". Climate Change was identified as the most important issue for 15% of NDP voters, 13 For the categorization, a zero-shot classification task was carried out with a large language model. 14 Petrova et al., "Social Media and Political Contributions: The Impact of New Technology on Political Competition" but it was rarely mentioned in posts by NDP candidates; this may be due to the Alberta wildfire crisis that was ongoing during the campaign period, which may have deterred candidates from discussing climate change in order to avoid the impression of linking an active crisis to a partisan policy issue. Supporting this conjecture, we found 152 candidate posts discussing the wildfires, but none linked wildfires to climate change. Another possibility is that the NDP did not want to draw attention to Alberta's carbon pricing program. Additionally, although both NDP and UCP voters showed relatively little interest in education as a top-priority issue, NDP and UCP candidates frequently posted about education. ![](_page_11_Figure_1.jpeg) Figure 4. Albertans' Most Important Issues by Party Figure 5. Policy Issues Discussed in Candidates' Instagram and Facebook Posts by Party ![](_page_12_Figure_1.jpeg) **Note: Percent of candidates' social media posts categorized into policy topics using zero-shot classification with a large language model. Based on 729 social media posts published by UCP and NDP candidates between 1 May 2023 and 19 May 2023.** #### Political Trust Next, we turn to political trust: to what extent do Albertans trust different political figures and institutions? We asked Albertans to indicate their trust for a leader or institution from 0 to 10, where 0 means they think this person/institution is untrustworthy, and 10 indicates trustworthy. Starting with leaders in *Figure 6*, there is a clear partisan dynamic. The figure presents box plots where the line in the rectangle represents the median (value separating the sample in two), and the edges the lower (Q1) and higher quartile (Q3). This visualization is interesting because it allows observing how the majority of partisans trust each leader. NDP voters have high trust in Notley (mean = 7.1) and distrust Smith (mean = 1.8); the same can be observed for UCP voters, who find Smith trustworthy (mean = 6.8) and Notley untrustworthy (mean = 2.3). Both differences between partisans are significant (p < 0.01). Trust of political leaders is highly polarized in Alberta, with voters from each party heavily distrusting the leader of the opposite party. Figure 6. Albertans' Trust Towards Party Leaders and Media Organizations ![](_page_13_Figure_2.jpeg) Note: Distribution of scores by party, for each media type. Based on a survey conducted with 948 Albertans between May 25th and May 29th, 2023. Weight for age, gender and education. For media organizations, as depicted in *Figure 6*, NDP voters report trusting mainstream media more than UCP voters, and independent media seem to have won the trust of a good proportion of UCP voters. Forty-seven percent of UCP voters distrust mainstream media. Trust in mainstream media by UCP voters is low, but trust in independent media is high for both parties. We also measure trust in the federal and provincial electoral institutions in *Figure 7*. There is no statistical difference in the trust given to Elections Alberta between the two groups of voters. However, UCP voters are much less trusting of Elections Canada. One possibility is that significant discussion of foreign interference in the past two federal elections has prompted an overall rise in distrust of Elections Canada. Another possibility is that the question captures a general antipathy towards federal government institutions amongst Albertans. Regardless, the large gap may be an indicator that the information habitat of UCP voters is different with regards to election-related discussions. ![](_page_14_Figure_1.jpeg) Figure 7. Albertans' Trust in Electoral Institutions #### Misinformation Finally, we turn to experiences with and attitudes towards misinformation. First, we asked Albertans if they had seen, read, or seen political misinformation in the past week. 1 in 3 Albertans report being certain of seeing misinformation (32%). The majority of Albertans are unsure if they saw misinformation in the prior week (51%; see *Figure 8*), consistent with prior work in Canadian elections,15 but roughly 36% of Albertans self-report viewing misinformation. The results are similar across party lines. Following this question, we asked respondents to describe the misinformation that they report seeing. Several answers to the open-ended question referred to politicians and parties and how they spread misinformation during the election. Respondents believe that both provincial and federal leaders spread misinformation through their ads, debates and speeches; these comments are often partisan, targeting one politician or party (e.g. Smith, Notley, or the UCP) but also target, to a lesser extent, politicians and institutions in general. When asked to describe the misinformation encountered, respondents targeted political ads (e.g., "Exaggerated 15 Bridgman et al., "Mis- and Disinformation During the 2021 Canadian federal election." and misleading political ads" or "NDP election advertising"), the leaders of the UCP or NDP (e.g., "Any of Danielle Smith's pre-election promises and political platform…" or "Provincial leaders debate"), Justin Trudeau ("Anything that Justin Trudeau [says] in the media.") or politics in general (e.g., "Politics is very much about misinformation"). Another important topic is media, where respondents think mainstream media propagates misinformation and unfair news about events. For example, a respondent wrote that "anything that the CBC says [is] lies," while another notes that "almost everything the media tells are lies." Participants also noted misinformation about more general issues such as climate change, health, and wildfires. For example, one noted that they encountered news about the idea that "the wildfires were started on purpose by NDP party members and supporters in the hopes that it would cause the current UCP government to look inept in the eyes of voters which would result in the NDP winning the upcoming election." Overall, few respondents report misinformation related to voting or the voting process, again consistent with our survey findings on high trust in Elections Alberta. ![](_page_15_Figure_2.jpeg) Figure 8. Albertans' Self-reported Exposure to Misinformation Do Albertans think that encountering misinformation online is a problem for democracy? As *Figure 9* demonstrates, there is a consensus among Albertans on the belief that misinformation threatens Canadian democracy (72%). While the agreement is similar among NDP and UCP voters, a substantial intensity difference exists. NDP voters are more likely to strongly agree with the two statements than UCP voters. Nonetheless, Albertans agree that misinformation is a real and threatening issue for the country's democracy. What do Albertans think about policy responses to answer misinformation online? We asked Albertans about the role of social media companies and the federal government. There is a consensus among voters for both parties that misinformation increases political polarization (77%), threatens Canadian democracy (78%) and that social media platforms should remove misinformation from their websites (77%). The majorities of both UCP and NDP voters support this consensus, although support is softer amongst UCP voters. However, public opinion is much more divided about the role of the government in monitoring misinformation on social media. This division is observable among NDP and UCP: a majority of NDP voters believe that the government should monitor misinformation on social media (60%). In comparison, only 34% of UCP voters agree with this statement. We observe a clear partisan divide in the potential implications of the federal government monitoring misinformation online. ![](_page_16_Figure_1.jpeg) Figure 9. Albertans' Attitude Toward Misinformation # Conclusion and Reflections This study set out to examine the extent to which distinct information habitats existed among UCP and NDP voters during the 2023 Alberta Provincial General Election. We find mixed evidence, with survey data suggesting a higher-than-expected consistency of news and social media platform use: there is little evidence that Albertan partisans have different news diets. In the Alberta information ecosystem, CBC and Rebel News demonstrate the most partisan sorting, but even those outlets are followed by voters across the political spectrum. However, collected digital trace data of an engagement network with candidates and media outlets on Instagram tells another story: one of a divided information environment in online spaces where shared viewership is limited and highly concentrated on independent media sources. One set of Instagram users engage more with independent news media outlets while another is more likely to engage with CBC, and Instagram users who engage with candidates online (with some exceptions) are unlikely to do so across party lines. Despite similar cross-partisan news consumption diets, we find large differences in attitudes. Our data show that most Albertans have low trust in the leader of the party that they do not support. UCP voters are less trustful of Notley and NDP supporters are less trustful of Smith. Moreover, we find that UCP voters are less trustful of the media in general, especially traditional outlets, but a majority of supporters of both parties trust both Election Alberta and Election Canada. Democratic institutions are still trusted in Canada, even as we observe a decrease in trust for other institutions like the media. We also observe that, while a vast majority of Albertans think misinformation is a danger to Canada's democracy and that social media platforms should remove it, UCP and NDP voters disagree on its source and content. Specifically, many believe that misinformation comes from the party they do not support, often accusing political advertisements or campaign promises of spreading misinformation. Although UCP and NDP supporters share news diets and social media platforms, coexistence in this information habitat is marked by severe distrust of information from political outgroups. This study provides a snapshot of the Alberta information ecosystem in 2023 – a snapshot both because it captures a moment in time, and it provides an incomplete picture. We cannot say decisively how many Albertans reside in secluded information habitats where the danger is greatest to social and political cohesion and a collective capacity to collaborate and govern a shared political future. Further detailed investigation at the population level is needed to better identify for whom information habitats exist. For our part, we will be further examining the presence or absence of information habitats and centering questions of interpretation and behavioural response to information from news and social media. # References - Boyd, Alex. ["'The tale of two Albertas': Danielle Smith's province is already divided. Are things](https://www.thestar.com/news/canada/the-tale-of-two-albertas-danielle-smith-s-province-is-already-divided-are-things-about/article_5982a4ac-b757-5804-b92a-5faa35ea9567.html?) [about to get worse?](https://www.thestar.com/news/canada/the-tale-of-two-albertas-danielle-smith-s-province-is-already-divided-are-things-about/article_5982a4ac-b757-5804-b92a-5faa35ea9567.html?)" *Toronto Star*, May 31 2023. - [Bridgman, Aengus. "The Role of Social Media in Polarizing Canadians."](https://www.zotero.org/google-docs/?broken=JDwj9s) *[Canadian Commision on](https://www.zotero.org/google-docs/?broken=JDwj9s) [Democratic Expression](https://www.zotero.org/google-docs/?broken=JDwj9s)*[, 2023.](https://www.zotero.org/google-docs/?broken=JDwj9s) [https://www.mediatechdemocracy.com/all-work/the-role](https://www.mediatechdemocracy.com/all-work/the-role-of-social-media-in-polarizing-canadians)[of-social-media-in-polarizing-canadians](https://www.mediatechdemocracy.com/all-work/the-role-of-social-media-in-polarizing-canadians)[.](https://www.zotero.org/google-docs/?broken=vQMu2K) - Bridgman, Aengus., Mathieu Lavigne, Melissa Baker, Thomas Bergeron, Danielle Bohonos, Anthony Burton, Katharine McCoy, Mackenzie Hart, Mathieu Lavault, Rupinder Liddar, Pangying Peng, Christopher Ross, Jaclyn Victor, Taylor Owen, et Peter Loewen. "Mis- and Disinformation During the 2021 Canadian Federal Election." *Media Ecosystem Observatory*, 2022. - Government of Canada. "[Bill C-18: An Act respecting online communications platforms that](https://www.justice.gc.ca/eng/csj-sjc/pl/charter-charte/c18_1.html) [make news content available to persons in Canada](https://www.justice.gc.ca/eng/csj-sjc/pl/charter-charte/c18_1.html)." *Charter Statements*, 2023. - Gruzd, Anatoliy, and Philip Mai. "The State of Social Media in Canada 2020." *Ryerson University Social Media Lab*, 2020. - Johnston, Richard. "Affective Polarization in the Canadian Party System, 1988–2021." *Canadian Journal of Political Science/Revue canadienne de science politique*, 2023. - [Merkley, Eric. "Mass Polarization in Canada: What's Causing It? Why Should We Care?"](https://www.zotero.org/google-docs/?broken=61QD0e) *[Canadian](https://www.zotero.org/google-docs/?broken=61QD0e) [Commision on Democratic Expression](https://www.zotero.org/google-docs/?broken=61QD0e)*[, 2023.](https://www.zotero.org/google-docs/?broken=61QD0e) [https://www.mediatechdemocracy.com/all](https://www.mediatechdemocracy.com/all-work/mass-polarization-in-canada-whats-causing-it-why-should-we-care)[work/mass-polarization-in-canada-whats-causing-it-why-should-we-care](https://www.mediatechdemocracy.com/all-work/mass-polarization-in-canada-whats-causing-it-why-should-we-care)[.](https://www.zotero.org/google-docs/?broken=zzbQzI) - Petrova, Maria., Sen, Ananya., & Yildirim, Pinar. "Social media and political contributions: The impact of new technology on political competition." *Management Scienc*e, 2021. 67(5), 2997-3021. - Wojcieszak, M., Menchen-Trevino, E., Goncalves, J. F., & Weeks, B. "Avenues to news and diverse news exposure online: Comparing direct navigation, social media, news aggregators, search queries, and article hyperlinks." *The International Journal of Press/Politics*, 2022, 27(4), 860-886.