1
Looking for Patterns Within and Across Media Messages
Every scroll, pause, and click on a social media platform teaches an invisible system
more about its user than the user often realizes about themselves. Within days, a single search for
running shoes can produce weeks of shoe advertisements, while a video watched twice can
reshape an entire feed. This quiet reshaping is not coincidence; it is the media literacy skill of
inducing at work, translated into computer code and applied at a scale no single observer could
match. Social media companies have transformed everyday behavioral data into algorithmic
pattern recognition, and in doing so, they have redefined how audiences encounter, interpret, and
are influenced by media messages. This paper argues that social media platforms apply the skill
of induction through data-driven algorithms to identify behavioral patterns in their audiences,
and that this inductive process produces both measurable benefits and serious risks for audiences,
a dynamic best explained through cultivation theory.
Induction is the process of drawing general conclusions about a set of elements after
observing only a small sample of that set, while inference occurs when a person combines those
observed clues with prior knowledge to reach a logical conclusion (Potter, 2019). Potter (2019)
explains that this pair of skills allows audiences, and increasingly machines, to notice recurring
patterns across media messages and predict what will happen next. Social media algorithms
perform an automated version of both skills continuously. Every like, share, comment, click,
save, and viewing duration becomes a data point that platforms feed into machine learning
systems designed specifically to detect behavioral patterns among users (Singh, 2024). From an
industry perspective, these systems allow media companies to treat audiences less as a mass
public and more as a collection of individually profiled consumers whose future preferences can
be forecast from past behavior (Potter, 2021). The result is a feed curated not by human editors
2
but by statistical prediction, personalized to each user's inferred preferences and updated with
every new interaction.
This inductive process carries genuine benefits. Personalized algorithms reduce the time
users spend searching for relevant content and can connect people to communities, information,
and entertainment aligned with their interests (Singh, 2024). In a large canvassing of technology
experts conducted by the Pew Research Center, 38% of respondents predicted that algorithms'
positive effects on society would outweigh their negative effects, citing efficiency, convenience,
and improved decision-making as primary advantages (Rainie & Anderson, 2017).
However, those same experts were nearly as likely to warn of harm; 37% predicted that
negatives would outweigh positives, citing bias, manipulation, and an erosion of human
judgment (Rainie & Anderson, 2017). Because engagement, not accuracy, drives algorithmic
ranking, sensational and emotionally charged content is systematically favored over measured or
accurate content. A large-scale study of diffusion patterns on Twitter found that false news
stories spread significantly farther, faster, and more broadly than true stories, largely because
falsehoods were more novel and provoked stronger emotional reactions such as fear and surprise
(Vosoughi et al., 2018). Eisenstat (2020), a former Facebook employee, similarly argued that
because platform algorithms are optimized purely to maximize engagement, they inevitably
promote inflammatory and polarizing content over calm, accurate information, regardless of a
platform's stated intentions.
This pattern connects directly to cultivation theory, a media effects framework asserting
that repeated, long-term exposure to consistent media messages gradually shapes a person's
perception of social reality (Potter, 2021). From the audience's individual perspective, viewers
rarely notice this shift happening because it occurs gradually across many small exposures rather
3
than through any single dramatic message (Potter, 2021). Cultivation theory, originally applied
to heavy television viewing, extends naturally to algorithmic feeds, because inductive systems
continuously reinforce the same categories of content a user has previously engaged with. A user
repeatedly shown emotionally charged political content may come to perceive society as more
divided or dangerous than it objectively is, not because the world itself changed, but because the
algorithm's inductive pattern recognition kept reinforcing that particular narrative until it felt
normal.
In conclusion, social media platforms rely on the media literacy skill of inducing,
translated into algorithmic code, to observe audience behavior, identify patterns, and predict
future preferences. This inductive personalization offers real convenience and connection, yet it
simultaneously rewards emotionally provocative content and, through processes explained by
cultivation theory, gradually reshapes how audiences perceive reality itself. Media literate
consumers who understand this inductive machinery are better equipped to recognize when a
feed reflects something true about the world, and when it simply reflects a pattern the algorithm
has learned to exploit for engagement.
Word count: 750
4
References
Eisenstat, Y. (2020, August). Dear Facebook, this is how you're breaking democracy [Video].
TED Conferences.
[Link]
democracy
Potter, W. J. (2019). Seven skills of media literacy. SAGE Publications.
Potter, W. J. (2021). Media literacy (10th ed.). SAGE Publications.
Rainie, L., & Anderson, J. (2017, February 8). Code-dependent: Pros and cons of the algorithm
age. Pew Research Center. [Link]
dependent-pros-and-cons-of-the-algorithm-age/
Singh, C. (2024, January 3). Understanding social media algorithm: 2024 complete guide.
SocialPilot. [Link]
Vosoughi, S., Roy, D., & Aral, S. (2018). The spread of true and false news online. Science,
359(6380), 1146–1151. [Link]