Algorithmically personalized advertising plays a role in reproducing structural inequalities, and digital advertising literacy is widely recognized as necessary to address its consequences. However, little is known about how these inequalities manifest within digital advertising literacy itself. This study examines how structural inequalities shape young people's knowledge of algorithmically personalized advertising and their perceptions of algorithmic bias in advertisement delivery systems. Based on an online survey of 1,200 individuals aged 14 to 30, the research assesses both objective and subjective knowledge of personalization mechanisms, alongside tasks exploring perceived gender- and class-based targeting in social media advertising. Gender identity and a multidimensional socioeconomic index are used as independent variables. The results show that women and participants from higher socioeconomic status perform better in the knowledge test, while men from privileged backgrounds report higher confidence levels. The participants generally perceive gender and class bias in advertisement delivery, although those from less privileged socioeconomic groups are less likely to recognize these patterns. These findings indicate that digital advertising literacy is unevenly distributed and linked to structural conditions. The study highlights the risks of normalizing biased targeting through repeated exposure and underscores the need for digital literacy efforts that address not only knowledge gaps but also overconfidence. It calls for greater algorithmic transparency and critical oversight of algorithmically personalized advertising systems that may reproduce structural inequalities. • Youth understand ad personalisation better through experience than formal knowledge. • Socioeconomic background strongly shapes algorithmic advertising literacy. • Confidence in knowledge does not always reflect actual understanding. • Young people perceive gender and class bias in targeted advertising. • Lack of algorithmic literacy increases vulnerability to ad manipulation.
Sáez-Linero et al. (Sun,) studied this question.