Arecent study by Pompeu Fabra University (UPF) reveals that socioeconomic conditions are a determining factor in young people’s ability to identify digital advertising personalization strategies. The lower young people’s socioeconomic status is, the more their ability to detect algorithmically personalized advertising also declines. Therefore, the study concludes that digital advertising is exacerbating structural inequalities and warns that algorithmic awareness cannot be understood as an individual cognitive ability but should be linked to people’s social environment.
To date, several studies had found that youth itself is already a factor of algorithmic vulnerability, due to the younger age and lower maturity of people in this demographic group. But t his UPF study pioneers the examination of the relationship between young people’s abilities in this field (their level of digital advertising literacy) and their socioeconomic status and gender.
The study , recently published by the journal Technology in Society , was led by Mònika Jiménez, director of the Communication, Advertising and Society ( CAS ) research group of the UPF Department of Communication . The main author of the article is Carolina Sáez, a researcher from the same research group, with Isabel Rodríguez-de-Dios (University of Salamanca) as a co-author.
The researchers conducted an online survey of 1,200 young people in Catalonia aged between 14 and 30 who had different profiles in terms of sociodemographic status and gender. To determine their socioeconomic status, each subject’s place of residence -one of the data collected by the survey- was cross referenced with the Government of Catalonia’s Territorial Socioeconomic Index ( IST ).
In the survey, the young people had to rate eight statements related to various aspects of algorithmic advertising as being true or false. Among other issues, they were asked about companies’ ability to cross-reference data between different devices used by the same person; to segment people into different groups according to their habits and behaviours to send them different types of advertisements, and the role of cookies to enable this; or the possibility of displaying different adverts on the same web page according to each user’s profile or personalizing them based on what is entered in a search engine, the contents of an email or what is said out loud.
The results of t he study show that overall young people have a good level of digital advertising literacy. On average, they correctly answered 6 of the 8 statements for classifying as true or false. However, the success rate of young people with a more privileged economic background is higher than that of those in the lowest socioeconomic stratum . Although the gender factor carries less weight than social class, it should be noted that girls always score better than boys in the same socioeconomic group. Girls' scores range from 6.2 (lower class) to 6.8 (upper class), while boys' range from 5.8 (lower class) to 6.4 (upper class).
The study also reveals that young people understand the most explicit strategies of persuasion, but less so the implicit mechanisms that enable this (related to data collection and processing). “ They are also better at detecting algorithmic advertising strategies that they learn from their own experience than those that are transmitted to them in an abstract way”, Carolina Sáez (UPF) explains. As they tend to consult social networks on their mobile phones, they are not so skilled at detecting advertising personalization through websites. In fact, the statement with the highest percentage of wrong answers is the one that refers to the possibility of seeing different ads when browsing the same website (4 out of 10 considered that this is not possible).
In addition, on a scale of 1 to 5 the young people had to rate their level of self-confidence in the answers provided in the previous statements, since another of the objectives of the study was to compare their objective and subjective knowledge. The average score for this self-confidence scale was 3.94. “This shows that young people generally feel confident in their own abilities, although their subjective perception of what they know does not always coincide with their real level of knowledge”, Carolina Sáez (UPF) warns.
However, this score varies greatly according to the young people’s profile and, in this case, the gender factor is far more relevant than in the case of objective knowledge . While in the case of boys self-confidence increases as socioeconomic status rises (from 3.8 to 4.2), in the case of girls it is the opposite (it falls from 4.15 to 3.7).
In another part of the study, the youths had to identify the target audience of four Instagram ads about cryptocurrencies and gambling (in these two cases, with regard to gender); and about financial education programmes and stock market investments (in these two cases, with regard to social class). The results reveal that, with little critical distance, young people reproduce the same gender and class stereotypes as perpetuated by the algorithms themselves.
Most of the youths responded that the first two ads were targeting men , based on the stereotype that links financial risk-taking and the male gender. This bias manifested itself more clearly among upper-class respondents. In other words, gender bias is lower among young people from lower socioeconomic backgrounds.
Concerning financial education adverts, the youths mostly responded that they were targeting lower-class young people, based on the cliché that this group requires more training to learn how to manage its money. In contrast, they mostly associated advertising on stock investments with upper-class people.
In view of these results, the study deems that there is a need to critically monitor algorithmically personalized advertising systems that can reproduce structural social and gender inequalities and calls for greater transparency on this issue from the companies responsible for them. It also insists on strengthening young people’s digital literacy to address not only knowledge gaps, but also overconfidence.
Reference article:
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, https://doi.org/10.1016/j.techsoc.2026.103287 .
Technology in Society
Survey
People
Algorithmic personalization and social inequality: young people's knowledge and perceptions of bias in digital advertising
2-Mar-2026