Political Microtargeting Research Articles

More Than Meets the Eye: Understanding Political Microtargeting Processing With Gaze-Cued Retrospective Think-Aloud Methodology

Political microtargeting is a popular campaign tool in elections worldwide. However, it is associated with democratic risks. Foremost, scholars and policymakers are concerned that citizens cannot cope with political microtargeting and, thus, stand vulnerable to persuasion. To assess this risk, an in-depth understanding of how citizens make sense of and cope with political microtargeting is required. However, empirical studies are scarce, partially inconclusive, and provide global rather than nuanced insights. This study contributes to this research by employing an innovative, qualitative gaze-cued retrospective think-aloud design to distinguish coping patterns and, based on that, assess citizens’ vulnerability to persuasion via political microtargeting (<em>N </em>= 25). The results reveal similarities regarding conceptual persuasion knowledge activation (i.e., advertising and targeting awareness) but differences in attitudinal and behavioral coping, illustrated along five coping patterns (avoidance, coherence assessment, ad quality assessment, partisanship bias, and neutral observation). Only individuals who cope by neutrally observing the political message seem vulnerable to political microtargeting. For individuals who cope with political messages based on their partisanship, political microtargeting might strengthen existing ties but backlash when employed by a non-preferred party. This study informs educators and policymakers about citizens’ coping mechanisms with political microtargeting and their potential vulnerability, which may guide intervention and regulation decisions.

Are Certain Types of Microtargeting More Acceptable? Comparing US, German, and Dutch Citizens’ Attitudes

Much of the research on political microtargeting has focused on growing public concerns about its use in elections, fuelling calls for greater regulation or even a ban on the practice. We contend that a more nuanced understanding of public attitudes toward microtargeting is required before further regulation is considered. Drawing on advertising psychology research and the results of academic analyses into microtargeting, we argue that individual concern, and by corollary, acceptance of microtargeting will vary based on socio-demographic characteristics and political orientations, and the type of personal data used. We hypothesise that microtargeting that relies on observable or publicly accessible personal information will be more accepted by voters than that which uses unobserved and inferred traits. We test these expectations and the expected variance of public acceptance by individual characteristics using comparative survey data from the US, Germany, and the Netherlands. We find that across countries and socio-demographic groups, not all microtargeting is considered equally problematic. For example, whereas the use of age and gender is generally deemed acceptable, the use of sexual orientation is not, and right-leaning individuals are more accepting than those who lean left. Additionally, overall, the US is more accepting of microtargeting than Germany or the Netherlands. Thus, we find that not all microtargeting is considered equally problematic across countries and socio-demographic groups. We conclude by calling for a more contextualised debate about the benefits and costs of political microtargeting and its use of “sensitive” data before the expansion of current regulation.

Editorial: Data-Driven Campaigning in a Comparative Context—Toward a 4th Era of Political Communication?

Contemporary political campaigning takes place both online and offline, and can be data-driven. In this piece, we review existing knowledge around data-driven campaigning (DDC) and introduce the new contributions made by the pieces within this thematic issue. We reveal how the studies included in this thematic issue of <em>Media and Communication </em>contribute to this existing knowledge by providing an up-to-date account of how DDC in general, and political microtargeting in specific, have been employed in election campaigns between 2021 and 2023, in a range of countries: France, Germany, the Netherlands, Sweden, and the US. As a collection, these studies highlight the variance that exists in the degree to which DDC is practiced, the range of DDC tools used, and attitudes toward DDC. In recent election campaigns, DDC takes many forms, and disapproval of DDC varies depending on how it is implemented.

Political Microtargeting – on the Possibility of a Polish Version of the American Cambridge Analytica Scandal from 2016

The use of political microtargeting, a technique of political market segmentation based on an algorithm that allows personalising campaign materials catered to individual voters, is on the rise with political campaigns across the world. Trump’s 2016 campaign provides an example of it being used in an ethically-questionable way that entails personal data infringement. The article looks at the possibility of using it in a similar way in Poland. To this end, the author presents a comparative analysis of a number of elements that make up the political landscapes of the United States and Poland.

Evaluating the persuasive influence of political microtargeting with large language models

Recent advancements in large language models (LLMs) have raised the prospect of scalable, automated, and fine-grained political microtargeting on a scale previously unseen; however, the persuasive influence of microtargeting with LLMs remains unclear. Here, we build a custom web application capable of integrating self-reported demographic and political data into GPT-4 prompts in real-time, facilitating the live creation of unique messages tailored to persuade individual users on four political issues. We then deploy this application in a preregistered randomized control experiment (n = 8,587) to investigate the extent to which access to individual-level data increases the persuasive influence of GPT-4. Our approach yields two key findings. First, messages generated by GPT-4 were broadly persuasive, in some cases increasing support for an issue stance by up to 12 percentage points. Second, in aggregate, the persuasive impact of microtargeted messages was not statistically different from that of non-microtargeted messages (4.83 vs. 6.20 percentage points, respectively, P = 0.226). These trends hold even when manipulating the type and number of attributes used to tailor the message. These findings suggest-contrary to widespread speculation-that the influence of current LLMs may reside not in their ability to tailor messages to individuals but rather in the persuasiveness of their generic, nontargeted messages. We release our experimental dataset, GPTarget2024, as an empirical baseline for future research.

Who Does(n't) Target You?

Political campaigns are increasingly investing in targeted advertising on social media platforms to reach voters. Despite critical implications for citizens and elections, little is known about the targeting strategies deployed by political parties — especially in countries beyond the global north. This paper provides a comprehensive descriptive analysis of political microtargeting practices on Facebook and Instagram across 95 countries during 113 national elections. By analyzing the Meta Ad Targeting dataset, we explore targeting strategies of 54k political advertisers who ran 2.5 million ads between August 2020 and December 2022. The findings indicate that election campaigns worldwide utilize targeted advertising. Most commonly, spending is allocated towards a single targeting criterion, however, in wealthier countries and electoral systems with proportional representation, a greater amount of money is spent on microtargeting by combining multiple criteria. Furthermore, targeting strategies vary along ideological lines of political parties who seek out voters more typically aligned with each side of the political spectrum. Nonetheless, parties use microtargeting irrespective of political ideology. Our findings offer the first comparative analysis of political microtargeting on Meta platforms for countries across all continents. Methodologically, we introduce a semi-automatic method to identify worldwide political advertisers using multiple data sources. Our study deepens the understanding of how country and party contexts explain differences in targeting strategies, highlighting the need for more research beyond the global north. Finally, our results have important implications for policy makers, and other stakeholders who seek to develop regulations to address the challenges posed by political microtargeting techniques.

A Comparative Analysis of Advertising in the 2020 Presidential Elections & Phoenix Mayoral Elections using Natural Language Processing

This language model examines advertisements used for political campaigns, specifically the 2023 Phoenix mayoral election. Political microtargeting looks at the tone of content throughout Instagram and Facebook ads during the 8 months leading up to the election. Through various processing techniques, both written and spoken language within ads was collected and used Facebook ad API, associated with a target demographic or group within the platform. This allows campaigns to send targeted messages to voters based on factors like location, media preference, associated political beliefs, age, and other critical variables. This data was then compared to a peer-reviewed study by the University of Amsterdam, that studied Facebook and Instagram ads 8 months before the 2020 US Presidential election. Media companies often offer paid advertisements to a targeted audience, and during highly polarized world events such as an election, uncivil negative ads can be pushed to groups to which the shown ideology appeals. (Votta, F.) The question posed by Votta in the comparative study was: “TO WHAT EXTENT IS POLITICAL MICROTARGETING USED BY POLITICAL ADVERTISERS TO DELIVER TOXIC CAMPAIGN MESSAGES?” To study whether this optimization was occurring, the advertisements were run through a multilevel ordinal regression(Votta [https://www.aup-online.com/content/journals/10.5117/CCR2023.1.001.VOTT] et al., 2023). They took ads from official political campaigns and compared them to third-party groups. They then analyzed the size of the desired audience and compared the size of the audience to the verbal tone of the listed advertisement. Our results indicate that while traditional media remains an important part of electoral voting, show out, and postage signs and campaign events serve a purpose in reaching a wide, unrestricted audience, social media advertising presents a valuable resource in targeting specific groups of individuals who will be most susceptible to guided advertisement. Last, we analyze the implications of our findings in the 2020 Phoenix mayoral election, how city elections differ from a presidential election, and how micro-targeted advertisements might affect voter behavior.

Microtargeting, voters’ unawareness, and democracy

Abstract Recent technological developments have raised concerns about threats to democracy because of their potential to distort election outcomes: (a) data-driven voter research enabling political microtargeting and (b) growing news consumption via social media and news aggregators that obfuscate the origin of news items, leading to voters’ unawareness about a news sender’s identity. We provide a theoretical framework in which we can analyze the effects that microtargeting by political interest groups and unawareness have on election outcomes in comparison to “conventional” news reporting. We show which voter groups suffer from which technological development (a) or (b). While both microtargeting and unawareness have negative effects on voter welfare, we show that only unawareness can flip an election. Our model framework allows the theory-based discussion of policy proposals, such as to ban microtargeting or to require news platforms to signal the political orientation of a news item’s originator (JEL C72, D72, D82, D83).

The persuasive effects of political microtargeting in the age of generative artificial intelligence.

The increasing availability of microtargeted advertising and the accessibility of generative artificial intelligence (AI) tools, such as ChatGPT, have raised concerns about the potential misuse of large language models in scaling microtargeting efforts for political purposes. Recent technological advancements, involving generative AI and personality inference from consumed text, can potentially create a highly scalable "manipulation machine" that targets individuals based on their unique vulnerabilities without requiring human input. This paper presents four studies examining the effectiveness of this putative "manipulation machine." The results demonstrate that personalized political ads tailored to individuals' personalities are more effective than nonpersonalized ads (studies 1a and 1b). Additionally, we showcase the feasibility of automatically generating and validating these personalized ads on a large scale (studies 2a and 2b). These findings highlight the potential risks of utilizing AI and microtargeting to craft political messages that resonate with individuals based on their personality traits. This should be an area of concern to ethicists and policy makers.

"I always feel like somebody's watching me"

The practice of political micro-targeting (PMT) – tailoring messages for voters based on their personal data – has increased over the past two decades, particularly in the U.S. Studies of PMT have to date concentrated largely on its effects on voters, or its implications for democracy more broadly. Less attention has been given to answering basic descriptive questions about how people perceive, feel and care about this new mode of political communication. This paper fills that gap by reporting findings from an online survey (weighted to be nationally representative on age, gender, ethnicity, region and past vote) that measured public attitudes toward PMT during the 2020 U.S. Presidential campaign. Specifically, we measure voter orientations toward PMT in four key dimensions – awareness, aversion, knowledge, and acceptability at the aggregate level – and explore how these vary according to a range of individual characteristics. Key findings are that public understanding and acceptance of PMT may be higher than current studies indicate, particularly among certain sectors of the population. Such insights are important for academic research to cognize and also policy-makers, as they move toward greater regulation of voter targeting.

What is Political Microtargeting

What is Political Microtargeting

Regulating political microtargeting

Regulating political microtargeting

Recommendations for addressing the risks of political microtargeting

Recommendations for addressing the risks of political microtargeting

Balancing perceptions of targeting: An investigation of political microtargeting transparency through a calculus approach

Over the last few years, political advertisers have moved with their audiences: to social media platforms. Advertisers on these platforms aim to persuade voters by sending messages tailored to them based on their own data: political microtargeting (PMT). A considerable problem with PMT is that users are often unaware that they are being targeted, while current transparency advances do not seem to suffice in informing users. However, increasing transparency may have consequences on users’ privacy perceptions. Thus, the current work investigates whether disclosures, as a measure to increase transparency, increase users’ recognition of a microtargeted ad, and subsequently what this means for their perceived benefits, privacy concerns, and their likelihood of engaging in privacy protection behavior, based on the privacy calculus. In a preregistered online one-factorial between-subjects experiment (N = 450) we exposed participants to either an Instagram post containing a currently used disclosure or a more salient disclosure. Our results show that exposure to this disclosure increases recognition of the ad being microtargeted, and that this relates to perceived benefits but not privacy concerns. However, the results show that users’ privacy concerns are related to their increased privacy protection behavior. Finally, we found that over four-fifths of our participants who were exposed to the more salient disclosure recalled it correctly.

Beyond the Ballot: A Comparative Analysis of Political Microtargeting Practices and Regulations in Kenya and Nigeria

Technological advancements have significantly impacted the political world. Unlike the traditional means of conducting campaigns, technology makes it possible to conduct data-driven campaigns on a larger scale and with high levels of specificity. For political parties, better clarity leads to hyper-individualized communication in a process known as microtargeting. Critics argue that political microtargeting can directly manipulate and suppress voters, exacerbate polarization, perpetuate misinformation, and indirectly lead to long-term effects by encouraging political parties to ignore individuals whom they deem unlikely to vote or those who are digitally excluded. This paper studies political microtargeting in Kenya and Nigeria because of data-driven campaigns that have been observed in these jurisdictions in the past and due to the increased reliance on social media platforms that political actors are utilizing to influence voters. It argues that political microtargeting is an issue of concern and therefore, using Kenya and Nigeria, it pushes the agenda that countries in the Global South should implement policies and regulations to curtail the negative impact of the practice. To examine the extent of political microtargeting in both Kenya and Nigeria, this paper employs a multi-phase approach that involves an analysis of paid Facebook advertisements in both countries during the past election periods.

Mitigating microtargeting: Political microtargeting law in Australia and New Zealand

AbstractTo the detriment of liberal democracy, governments have struggled to prevent the exploitation of personal data for voter manipulation in the digital era. Laws pertaining to political microtargeting are often piecemeal and tend to derive from a combination of laws on electoral advertising and privacy. Evidence indicates that this approach is insufficient to curtail microtargeting. However, little is known about the regulation of microtargeting outside of the European and US contexts within which the bulk of anti‐microtargeting research has been undertaken. Accordingly, this paper aims to shed light on the preparedness of the law in Australia and New Zealand to mitigate the potential harms of political microtargeting. A comparative analysis of legislation pertaining to microtargeting is therefore undertaken using a blended approach of comparative law and content analysis. This paper: (1) identifies current legislation relevant to microtargeting in Australia and New Zealand; (2) assesses patterns of similarity and difference between each country's laws in relation to microtargeting; and (3) evaluates the preparedness of current legislation to curtail microtargeting in an evolving social media landscape. It finds that in both countries, legislation is sufficiently robust to mitigate microtargeting in some limited circumstances, but a cohesive regulatory approach is needed to constrain the most insidious microtargeting operations.

Personality and susceptibility to political microtargeting: A comparison between a machine-learning and self-report approach

Based on recent technological advances, campaigners and political actors can use psychographic-based political marketing. Yet, empirical evidence about its effectiveness is still very limited. Based on self-congruity theory, a pre-registered experiment (N = 280) investigated the persuasion effects of personality-congruent political microtargeting on the attitude toward the political party and voting intentions of citizens. More precisely, the focus was on the thinking vs feeling personality dimension (MBTI), and it was tested whether this personality “interacts” with exposure to a matching advertising appeal: rational vs. emotional political ad. To do so, two different methodological approaches were used: 1) a machine learning approach; 2) a self-report survey measure of personality. Results revealed significant “congruence effects” between personality and ad appeal, and showed that perceived ad relevance was serving as the underlying mechanism (mediator). However, these results were only found when the self-report measure of personality was used. When the algorithmic approach was used, no significant results were found. These findings feed into timely societal, methodological, and theoretical contributions.

Trusting tech firms’ big data for political microtargeting? A qualitative analysis of parties’ communication managers risk and trust perceptions

ABSTRACT Parties use political microtargeting (PMT) to address voter subsegments individually. Due to limited resources and legal restrictions, parties often rely on Meta’s platforms and advertising ecosystem for PMT. However, using these external infrastructures and big data analyses might be risky for parties, since big data are criticized for their validity, robustness, representativeness, and usefulness. A mechanism to tolerate risks is trust. With the theoretical background of trust in technology, this study investigates the extent to which parties’ communication managers perceive risks in their strategic use of PMT on Meta’s platforms and how they evaluate big data’s trustworthiness. Based on in-depth expert interviews with German parties’ communication managers on state level, the results show that parties’ communication managers perceive various risks in relying on Meta’s platforms and are ambivalent about PMT’s big data analytics in terms of quality and reliability. To minimize risk perceptions, parties adopt strategies such as inputting their own data or creating own target audiences. Parties’ risk perceptions only partially influence their trust in big data and Meta. Despite varying degrees of trust, PMT is still used in campaigns as it has become too common and necessary to compete with other parties.

Is Personality Key? Persuasive Effects of Prior Attitudes and Personality in Political Microtargeting

Messages that are designed to match a recipient’s personality, as enabled by microtargeting, have been found to influence political reasoning and even voting intentions. We extended these findings by adding prior attitudes to a microtargeting setting. Specifically, we examined what role different microtargeting approaches play in political reasoning by conducting an online experiment with a 2 (extraverted vs. introverted communication) × 2 (attitude-congruent vs. attitude-incongruent statement) between-subject design (<em>N</em> = 368). In line with the assumptions of the theory of motivated reasoning, attitude position matching emerged as an effective microtargeting strategy, and attitude strength moderated the effect of attitude congruency on recipients’ evaluations of political ads. While extraverted messages had no direct effect, that was unrelated to attitude congruency, recipients’ level of extraversion moderated the effect of extraverted communication on their evaluation of an ad. Interestingly, the intention to vote was significantly higher when an attitude-incongruent statement was phrased in an introverted rather than an extraverted manner, suggesting that information that challenges prior attitudes might be more persuasive when it is delivered in a more temperate way. In sum, the study indicates that matching message with personality alone might not be the most effective microtargeting approach within democratic societies.

Instaworthy? Examining the Effects of (Targeted) Civic Education Ads on Instagram

The last few years have witnessed a growing societal and scholarly interest in the potential of online political microtargeting to affect election outcomes in favor of parties and candidates. It has often been rightly pointed out that political microtargeting can pose risks to electoral integrity in democracies. But can political microtargeting also benefit democratic functioning? Very little is known about the potential of political microtargeting to affect citizens’ attitudes towards politics and increase their civic participation. To address this paucity, this article presents a preregistered online experiment conducted in Germany among young adults (<em>N</em> = 445), examining whether (targeted) civic education ads on Instagram increase political interest, efficacy, and civic participation. An innovative methodological approach to studying political microtargeting is deployed, exposing respondents to civic education ads in a mock Instagram feed, personalized in real-time based on individual preferences. We find no direct evidence of (targeted) civic education ads, leading us to believe that (targeted) ads do not unconditionally affect political interest, efficacy, or civic participation.