A recent breakthrough study conducted by Pompeu Fabra University (UPF) unveils critical insights into how socioeconomic conditions fundamentally influence young people’s capacity to discern digital advertising personalization techniques. This profound research highlights a troubling decline in the ability to detect algorithmically targeted ads among youths from lower socioeconomic strata. Ultimately, the findings cast light on how digital advertising algorithms perpetuate and intensify existing structural inequalities in society. The study compellingly argues that awareness of these algorithmic systems should not be viewed as a mere individual cognitive skill but must be contextualized within the social environments that shape young people’s experiences and knowledge.
Prior research has acknowledged youth as an intrinsically algorithmically vulnerable demographic, often due to a combination of limited life experience and developing maturity. However, this pioneering study from UPF marks the first comprehensive effort to examine the intersection between young people’s algorithmic advertising literacy and their socioeconomic backgrounds alongside gender influences. By surveying a large, diverse population of young Catalans aged 14 to 30, researchers gained significant empirical data linking social disparities with variations in the understanding of how digital ads are tailored by algorithms.
The research team employed a robust methodological framework, surveying 1,200 young individuals spread across a broad spectrum of socioeconomic conditions. To categorize participants effectively, the researchers cross-referenced each respondent’s residential location with the Government of Catalonia’s Territorial Socioeconomic Index (IST). This detailed stratification ensured that the nuances of socioeconomic status were meticulously captured. Survey participants were then asked to evaluate the truthfulness of eight carefully constructed statements relating to the nature and mechanisms of algorithmic advertising personalization.
The statements covered sophisticated concepts such as advertisers’ capability to cross-reference data from multiple devices used by the same individual, the segmentation of users into specific behavioral groups, and the pivotal function of cookies as tools enabling targeted advertising. Moreover, the participants were queried about algorithmic capacities, including the display of different advertisements on the same webpage depending on user profiling and ad personalization based on outspread user activity like search queries, email content, or even spoken words. These questions were designed to gauge an intricate understanding of both overt and subtle forms of algorithmic advertising.
Crucial findings from the study reveal that, overall, young people possess a commendable level of digital advertising literacy, with an average of six correct answers out of eight. Yet, a more granular analysis exposes a glaring disparity; youths hailing from more economically privileged backgrounds consistently outperform peers from lower socioeconomic classes. This effect notably overshadows gender differences, though females across all socioeconomic strata tend to score better than their male counterparts. Specifically, girls’ scores ranged from 6.2 in the lower class category to 6.8 in the upper class, whereas boys scored from 5.8 to 6.4 respectively, pointing to a significant intersection of class and gender in shaping algorithmic literacy.
The research also pinpointed a striking lack of recognition of certain digital advertising practices. While adolescents were relatively adept at identifying explicit persuasion tactics, their grasp of the underlying implicit mechanisms—particularly those involving data collection and processing—was less secure. For instance, 40% of respondents failed to recognize that online advertisements on a webpage could vary depending on the user’s profile. This disconnect reflects a broader challenge in how digital natives, despite frequent interactions with personalized content on social media mobile platforms, struggle to identify less visible algorithmic manipulations on websites.
Adding complexity, the study assessed participants’ confidence levels in their own answers, juxtaposing subjective self-assessments with objective knowledge. Rated on a scale from 1 to 5, the average self-confidence score was 3.94, signifying a general belief in their competence. However, this confidence did not consistently align with actual understanding, exposing an overestimation of expertise among some respondents. Gender intricately affected confidence differently than knowledge: boys’ self-confidence rose with higher socioeconomic status, whereas girls’ confidence paradoxically declined as their socioeconomic status increased, a pattern warranting further investigation into socio-cultural influences on self-perception.
The survey extended its scope to explore how young people perceive algorithmic advertising stereotypes related to gender and class. Respondents evaluated the intended target audiences for Instagram advertisements related to cryptocurrencies and gambling—where gender biases are salient—and for financial education and stock market investments, which are often framed through social class lenses. Alarmingly, many young respondents internalized and reproduced the same gender and class stereotypes embedded in the advertising algorithms themselves, demonstrating the depth of algorithmic influence on social cognition.
Specifically, the majority associated cryptocurrency and gambling ads as being targeted toward men, mirroring stereotypes linking males with financial risk-taking behaviors. This gender bias was even more pronounced among youths from higher socioeconomic backgrounds, suggesting that privilege may amplify certain stereotypical worldviews. Contrastingly, when evaluating financial education advertisements, most youths identified the target audience as lower-class individuals, reflecting the common trope of needing greater financial literacy. Stock market advertisements were generally linked to the upper class, thus perpetuating entrenched ideas about economic stratification and financial opportunity.
The study’s findings underscore a pressing societal concern: algorithmic personalization systems embedded in digital advertising do not merely reflect existing inequalities—they exacerbate them. By perpetuating biased representations and reinforcing systemic disparities, these algorithms contribute to a digital environment that disadvantages already vulnerable groups, especially young people from lower socioeconomic backgrounds. Consequently, researchers advocate for enhanced regulatory scrutiny and transparency from companies operating these personalized advertising systems.
Given the dual challenges of unrecognized algorithmic bias and excessive self-confidence among young users, the study calls for targeted educational interventions aimed at improving digital literacy in a way that accounts for both knowledge gaps and perceptual mismatches. Fostering a critical perspective on how algorithms operate and how bias manifests in digital advertising is vital to empower youth as informed digital citizens, capable of navigating and challenging these opaque systems.
This groundbreaking research published by the journal Technology in Society offers a crucial roadmap for future policies and educational frameworks. By positioning algorithmic awareness within the broader context of social environment and inequality, it reframes the discourse around digital literacy as a collective societal responsibility rather than an isolated individual skill. As digital technologies become increasingly interwoven with everyday life, such insights are indispensable for addressing the urgent challenges of fairness, inclusivity, and equity in algorithmically mediated spaces.
Subject of Research: People
Article Title: Algorithmic personalization and social inequality: young people’s knowledge and perceptions of bias in digital advertising
News Publication Date: March 2, 2026
Web References: https://doi.org/10.1016/j.techsoc.2026.103287
References: Carolina Sáez-Linero, Isabel Rodríguez-de-Dios, Mònika Jiménez-Morales, Algorithmic personalization and social inequality: young people’s knowledge and perceptions of bias in digital advertising, Technology in Society, Volume 86, 2026, 103287, ISSN 0160-791X
Keywords: Advertising, Young people, Social media, Algorithmic bias, Digital literacy, Socioeconomic status, Gender stereotypes, Algorithmic personalization, Inequality, Media literacy
