SOCIAL MEDIA ALGORITHMS AND DATA MANAGEMENT

Ivana STAMENKOVIĆ

Assistant professor, University of Niš, Faculty of Philosophy, Department of

Communicology and Journalism, Niš, Serbia
E-mail:[email protected]

Dušan ALEKSIĆ

Assistant professor, University of Niš, Faculty of Philosophy, Department of

Communicology and Journalism, Niš, Serbia
E-mail:[email protected]

Tatjana ĐUKIĆ ŽIVADINOVIĆ

Ph.D., University of Niš, Faculty of Philosophy, Department of

Communicology and Journalism, Niš, Serbia
E-mail:[email protected]

Abstract
Although the audience in the digital media space has more power
than in the traditional media environment, as indicated by their
ability to create, reshape and share content, media users’ behavior
is shaped by the use of algorithms and big data management.
Taking into consideration the fact that students use the internet
and social media platforms daily, this paper aims to examine their
perceptions and viewpoints on the operation of algorithms and
data management on the Internet. According to a survey
conducted by the authors, which consists of 200 respondents, two-
thirds of students notice the results of the algorithmic
personalization, filtered selection of content and news, and the
customized display of content on social media. Even though 70%
of them realize that user activities are continually monitored and
that control over personal data online is taken over by large
companies and/or a third party, most respondents express only
moderate concern for their data online (82%), which further
confirms the fact that only a small percentage of students (18%)
almost always read the terms of use on a website, application, or
internet service.

Keywords: social media, big data, algorithms, students

 Ivana STAMENKOVIĆ, Dušan ALEKSIĆ, Tatjana ĐUKIĆ ŽIVADINOVIĆ

Introduction

In the modern communication environment which relies on advanced

technology, people become the captives of the interests of “surveillance
capitalists” (Zuboff, 2020), due to their actions in the digital domain. Despite
the optimistic predictions that digital technology and the introduction of new
communication platforms will encourage the formation of a new democratic
zone and civic activism, there are increasing attempts to control, monitor and
direct individuals and groups toward particular commercial and political goals.
Such pretensions are demonstrated by companies and institutions, which
collect, store and analyze “every trace of human activity” while having some
benefit from big data (Vaidhyanathan, 2018, p. 74). The use of big data
observed as the next level of datafication and described as the "transformation
of social action into online quantified data, thus allowing for real-time tracking
and predictive analysis" (Van Dijck, 2014, p. 198), is linked to algorithms
involved in the organization, classification, and categorization of digital
content. “Algorithms contain a series of digital instructions programmed into
computerized systems so that they recognize patterns and enable data mining
and collection” (Chayko, 2018, p. 103). Simultaneously with finding and
displaying relevant, customized information to network users, a database
containing a personal archive and profile about the user is created for the future
exploitation of their personality and behavior shaping. These processes are
mostly carried out without the knowledge and awareness of the users of digital
services, and the effects reach deep and touch the most hidden level of their
personality - "The greatest change happens when using new media becomes our
routine, when we accept new media to the extent that we neither think about
their functioning mechanisms, nor their consequences on our behavior (Aleksić
& Stamenković, 2018, p. 104). Needs, feelings, attitudes, affinities, interests,
and other segments of the lives of networked people become available and
visible to the trading elite that exchanges the data to successfully sell political
ideas, and ideological beliefs, as well as necessary and redundant commercial
products to users. Any use of digital media and smart devices leads deeper into
a surveillance experiment in which privacy, autonomy, and freedom of choice
are lost so that users become the captives of other people’s interests, decisions,
and recommendations.
In addition, every online activity becomes a trace that is monitored,
analyzed, and used for different purposes, such as creating algorithmic profiles
based on a user’s online activity which is valuable for various companies from
a commercial point of view. These traces provide data about the identity and
lifestyles of users, who, almost entirely unaware of such manipulative games,
supported by artificial intelligence in recent times, continue to participate in the
modern forms of totalitarian surveillance. Based on the people’s previous
online experiences, algorithms select, filter and display content that is
personalized and customized to the individual, by the existing attitudes,
opinions, choices, and emotions. Also, it can produce a distorted image of
reality, and hide a different world, rich with alternatives, from the digital

200 Balkan Social Science Review, Vol. 21, June 2023, 199-217
 Social media algorithms and data management

universe users. On the other hand, the level of criticism towards the algorithmic
programming of social reality can vary among media users, which depends on
numerous factors, such as user behavior, the degree of participation online, the
level of education, as well as the knowledge of the methods of functioning of
modern, digital communications based on the principles of algorithmic
organization of information space. Moreover, the unpredictability of content
that differs from the media users’ preferences and interests, may cause
skepticism of young people towards a social media platform as a provider of
information, specifically the awareness of the reciprocal relationship between
user behavior and algorithm-dependent decisions (Swart, 2021). If users can
detect latent forms of manipulation carried out by organizations, institutions,
groups, and individuals on the internet analytically and critically, the belief that
the benefits of the digital world are easily accessible, with plenty of knowledge
available and social barriers removed, becomes a delusion.
In a world free from hierarchical forms of influence of traditional and
official news sources, social media platforms and search engines, together with
the companies that own them, have become the dominant distributors of
information and programmed consciousness. These groups are taking over the
role of information gatekeepers and social agenda editing, thus suppressing the
former information masters. In addition, the difference is that instead of the
human factor, in the domain of social media and search engines, the editorial
function is taken over by an algorithm, under the influence of the criteria of
those who created and implemented it (Presuel & Martinez Sierra, 2019).
Taking into account that students use the Internet and social media platforms
daily, this paper aims to examine their perception of and attitudes toward the
operation of algorithms and data management on the Internet.

Data Trading on the Internet - Circumstances and Consequences

Contrary to the expectations that the internet will enable greater

freedom and autonomy of users, that they will be connected with others,
eliminate spatial distance and facilitate learning, obtaining information, and
providing entertainment, the reality framed by digital experience reveals its
negative side. Although the internet provides various possibilities for civic
engagement, whether of social, political, or entertainment nature (Obradović &
Mitrović, 2019)1, every time an individual accesses the online space, is the
beginning of their activity being tracked based on the content of their search,
the websites they visit and the action they take, such as clicking on news,
selecting a product of purchase, or liking on social media. The digital world has
evolved into a stage where network users are permanent participants that attract
all the attention, particularly interested institutions, advertisers, organizations,
and social media owners. However, digital technology itself cannot be the cause

1
All translations of quotations made by non-English authors were made by the authors
of the article.

Balkan Social Science Review, Vol. 21, June 2023, 199-217 201
 Ivana STAMENKOVIĆ, Dušan ALEKSIĆ, Tatjana ĐUKIĆ ŽIVADINOVIĆ

of the current situation; its use is always influenced by social, political, and
economic factors.
Commercial imperatives dictate the further direction and cannot be
equated and justified by technological inevitability. For this reason, according
to Shoshana Zuboff, we should look for puppeteers “who secretly, behind the
scenes, control the machines and tell them what to do” (Zuboff, 2020, p. 27).
Zuboff also pointed out the modern form of capitalism, the so-called
surveillance capitalism, which claims unilateral ownership of the human
experience and behavioral data. While one part of the data is used to improve
services and products, the other part - a private behavioral surplus, is used to
predict future behavior thanks to machine intelligence. Additionally, these data
are transformed into prognostic products and become key items on the market
of ideas (Zuboff, 2020, p.18). An insight into these data enables the prediction
of future behavior and directs millions of people around the world toward
acceptable choices dictated by surveillance capitalists. Behavior modification
is becoming a goal, and how it is achieved is becoming more hidden, perfidious,
and complex. Through simple decisions such as purchasing a product, watching
a certain movie, or visiting a restaurant, to forming a lifestyle or a political or
ideological orientation, a new market project is imposed and is managed from
the shadows by the most intimate part of our personality, and cannot be avoided.
Data mining has become a lucrative strategy that ensures a long-term impact
on the user and their behavior. It is easy to influence people’s future activities
with filtered content if it is known what they want, their personal and
professional interests, plans, and habits. Therefore, based on the extracted data,
one can assume the structure of an individual’s personality, character, affinities,
what they want to have, buy, do, or what they want to be. It has become
generally accepted that personal data that users leave online is used, classified,
and grouped and this process is carried out with the help of search tools, the
most popular of which is Google. Also, these search tools serve users by
providing information they consider relevant based on individual user’s
previous online activities, while at the same time collecting, sorting, storing,
and selling data to advertisers and other third parties. Only a small number of
users are aware that their privacy has been violated, and the issue emerges as
the result of the unread terms of use, which are complex, incomprehensible, and
changeable. One of the problems is the insufficient critical potential of users to
understand how they can protect themselves and share as little personal data as
possible.
Taking into consideration that surveillance is a constant phenomenon
in the world (Zuboff, 2020) and that its tentacles are expanding and multiplying
in the circumstances of digital communication, Mary Chayko distinguishes
between vertical (asymmetric) and horizontal (social) surveillance (Chayko,
2019, p. 100-109). Online surveillance relies on the use of the internet to track
or observe someone’s behavior. In the case of a “solid hierarchical structure of
power”, the surveillance that is carried out for commercial, political, or legal
purposes is called vertical. It involves the government with its institutions and
various corporations, which possess the power to influence, direct, shape, and

202 Balkan Social Science Review, Vol. 21, June 2023, 199-217
 Social media algorithms and data management

protect network users. Horizontal or social surveillance, on the other hand,

describes surveillance roles assigned to other people online, with whom data is
shared. Their power is equal, and the potential difference is related to
communication skills. As Chayko indicates, race, gender, age, class, sexual
orientation, and social status can produce a difference in the power of
communication. “[P] perhaps the most striking difference between horizontal
and vertical surveillance is reflected in the degree, type, and expectation of
reciprocity in relationships” (Chayko, 2019, p.107). Whether or not there is an
awareness of constant surveillance, either vertical or horizontal, the fact
remains that in technologically developed societies, people find it difficult to
conceal information about their behavior and personality. The methods of
collecting and combining data have progressed, and user profiles are now
created using algorithms that recognize the patterns of behavior. Users’
behavior is predicted based on these algorithms, and users’ lives can be
impacted and further controlled by knowing various personal data (Chayko,
2019; Vaidhyanathan, 2018; Zuboff, 2020). According to Zuboff, people are no
longer “the subjects of value realization”, but “the objects from which raw
materials are extracted and expropriated for Google’s prediction factories” and
“predictions about our behavior are Google’s products, and they are sold to its
actual customers, but not to us; we are the means to other’s ends” (Zuboff, 2020,
p. 109).
The state of widespread and continual digital surveillance inevitably
raises concerns about the right to privacy and its protection. Who protects the
right to privacy in modern market conditions? And is it possible to control the
data we leave online? The right to privacy encounters an obstacle in the
contemporary communication and information environment. There are many
ways in which private data can be misused, and one of the examples is the
Cambridge Analytica data scandal, which was discovered in March 2018.
Namely, the company Cambridge Analytica, which advised then-President
candidate Donald Trump in the 2016 election campaign, used illegally private
data of more than 50 million users of the social network Facebook, and in this
way influenced political campaigns in several countries. The data was obtained
through a Facebook quiz to create an algorithm that determined personality
traits that are, among other things, related to voting preferences. Facebook
learned about data collection at the end of 2015 but did not inform users about
it, it was only ordered to remove the application and destroy the collected data
(Đukić & Stamenković, 2019, p. 159). Due to the collection and storage of large
amounts of data (contact lists, photos, messages, etc) that we leave by accessing
websites and applications, our right to privacy is jeopardized. Traces containing
data can be visible and clear, such as first names, last names, and email
addresses, or invisible and hidden, such as location and time of access.
(Mitrović, 2019). Although the right to privacy is protected by numerous
international documents, many social media platforms, among which Facebook
is the leader, define the term privacy differently and interpret it differently.
There is a combination of different social contexts on Facebook, while the
control over shared content is left to users by this social media itself.

Balkan Social Science Review, Vol. 21, June 2023, 199-217 203
 Ivana STAMENKOVIĆ, Dušan ALEKSIĆ, Tatjana ĐUKIĆ ŽIVADINOVIĆ

Facebook’s terms and conditions discourse is dominated by the attitude that the
user bears the responsibility and that privacy is a personal matter. According to
Siva Vaidhyanathan, “privacy protection is a problem of the environment that
Facebook approaches as a matter of personal responsibility; for Facebook,
privacy is a structural problem that can be only solved by the work and insight
of users” (Vaidhyanathan, 2018, p.83). In the desire to keep the attention of
users and save it, the issues of privacy and dignity become inconvenient and
redundant (Vaidhyanathan, 2018).
We depend on how the algorithm determines is based on the data we
have consciously or unconsciously left in the digital space. “Data sets are in
themselves meaningless. They must be made useful. Therefore, we are “made
subject not to our data, but to the interpretations of that data” (Cheney-Lippold,
2017, p. 254). As the author adds, in the process of shaping knowledge about
the world and ourselves, the central place belongs to algorithms, databases, and
their logic (2017). Another explanation offered by Cheney-Lippold is that the
citizen is, for example, the one that currently produces data that the algorithmic
logic defines as a citizen. Algorithmic identification begins as soon as an
individual enters the online world, and its consequences extend into a future
that is becoming increasingly uncontrollable.
Distribution of content on social media is closely related to what
Riemer and Peter describe as the algorithmic audience. They use this term "to
refer to the automatic and ad-hoc configuration of audiences for speech through
algorithmic content distribution, as a by-product of profit maximization
(Riemer & Peter, 2021,p. 9). The consequences of algorithmic programming
include fragmentation and polarization of the audience, the decline in news
quality, and the radicalization of public discourse (due to the presence of
misinformation and fake news), as well as uncivilized and hate speech (Stark
& Stegmann, 2020, p. 6; Spohr 2017). As it was argued multiple times: "This
happens because more extreme, more outrageous and thus more polarising
content is often found to be most engaging and thus amplified by the algorithm"
(2021, p. 11; Marantz 2020; Edelson et al. 2021; Vaidhyanathan 2021). Many
authors consider that these negative phenomena originated from the operation
of the algorithmic logic of social media, the personalization of content
according to the user, as well as the so-called filter bubbles and echo chambers.
Back in 2011, Eli Pariser, a writer, internet activist, and entrepreneur, drew
attention to the phenomenon that appeared on social media, driven by market
imperatives and the desire to achieve a commercial benefit and called it the
“filter bubble” (Pariser, 2011; Vaidhyanathan, 2018). It is about delivering
content that is similar that what users searched for and responded to by clicking,
liking, sharing, and commenting. In this way, Facebook has become a machine
that has great power in organizing our information space, our connections and
relationships with others, and our perspectives on social reality. As noted by
Cetina Presuel & Sierra: "Distributing any type of content that may interest
their users means that those users will remain engaged and spend time on their
services and that more data will be collected" (Cetina Presuel & Sierra, 2019,
p 265). Closely related to this phenomenon is the phenomenon known as the

204 Balkan Social Science Review, Vol. 21, June 2023, 199-217
 Social media algorithms and data management

“echo chambers”, which refers to the formation of a group of like-minded

people, where there is an echo of similar opinions and beliefs. The “filter
bubble” refers to an individual around whom the information world is built
based on algorithmic recommendations, independently separate from social
connections and discussions, whereas the “echo chamber” refers to an
individual being paired with like-minded people and participating in social
contacts and discussions (Stark & Stegmann, 2020). If users are constantly
guided cognitively and emotionally, then the power of this social media action
takes on immeasurable proportions and can have severe consequences on an
individual and society. The narrowing of the field of vision due to the effect of
the “filter bubble” and “echo chambers” decreases the possibility to see the
wider information spectrum and to develop grounded attitudes and long-term
beliefs. Communication in the so-called “niches” (Livingstone, 2012) of like-
minded people is opposite, in terms of effects, to open communication where
different opinions and feelings meet. The starting point of such discussions,
“unpolluted” by different opinions, creates the preconditions for the
fragmentation of the audience into subgroups that rarely come into contact with
one another. This also makes it difficult for people to reach an agreement on
important social issues, which can affect the stability of democratic societies.
Polarization is often interpreted as the division of society into different political
camps, although it can refer to divisions in viewpoints, such as ideological or
thematic divisions, as well as divisions in effects or polarization of experiences
(Stark & Stegmann, 2020, p.15). Existing beliefs are strengthened and the
affective public produced on Facebook further secures its position, whereas
people are divided into polarized groups, without achieving mutual
communication. Increasing polarization, political segregation, the lack of civic
dialogue, and distrust towards public institutions represent the greatest problem
of the modern, digitally networked society (Vaidhyanathan, 2018; Chitra &
Musco, 2019). Dialogue and tolerance are powerful barriers to the rule of one-
sidedness (Šušnjić, 2007, p.104). Experiencing a difference brings a new
quality and value to the development of society and the personality of
individuals, while the echo of identical, cognitively harmonized, emotional
tones leads individuals into spiritual poverty, intolerance, and animosity
towards others. Although the discussion in like-minded enclaves can lead to a
liberated and positive communication between marginalized communities in
real life (Sunstein, 2018), such “deliberative enclaves” (Barbera, 2020) lead to
a homogenized view of social reality and radical attitudes and opinions
(Sunstein, 2001), enabling extreme attitudes and extremist groups easy to form
in social media spaces. In addition to filter bubbles and algorithm operations,
which contribute to the affirmation of existing opinions, and attitudes and the
emergence of polarization, homophily, a tendency to associate with people who
are similar to them, has been noticed among people. This can be interpreted as
personalization led by an individual. The combination of these factors,
technologies used for modern forms of surveillance, and people with natural
tendencies to turn to ideologically like-minded people, creates a fertile ground
for the development of various forms of manipulation (Vaidyanathan, 2018).

Balkan Social Science Review, Vol. 21, June 2023, 199-217 205
 Ivana STAMENKOVIĆ, Dušan ALEKSIĆ, Tatjana ĐUKIĆ ŽIVADINOVIĆ

Due to structural changes in the media market, where information

intermediaries such as Facebook, Google, and other social media platforms are
becoming the leaders in providing information to people, traditional media are
gradually conforming to economic imperatives and new business conditions.
One of the consequences of the market pressures and the actions of the
algorithmic logic of social media, which have become important actors in the
online space, is reflected in the softening of news, adapting to tabloid principles,
becoming sensational, spectacular, and shocking. Provoking emotions is the
most common instrument of ensuring that the news will be read, shared with
others, or commented on by a user. While the effects of echo chambers and
filter bubbles are overestimated (due to “weak connections” with people
online), the risks of political polarization and social fragmentation remain
(Stark & Stegmann, 2020; Barbera, 2020). Echo chambers and polarization can
occur if there are more homogenous groups, when topics are effective or
controversial, and there are clear political predispositions. At that point people
are easily divided into opposing groups with such affective polarization being
based on stereotypes and negative assessments of others, non-belonging (Stark
& Stegmann, 2020). Polarization on social media is frequently a result of very
active individuals spreading political ideas, and their influence encouraging
others to engage in online party agitation. Barbera and Rivero obtained
evidence for this viewpoint after analyzing the hyperproduction of pro-party
content on Twitter ahead of the presidential elections in 2012. An active,
visible, and pervasive minority encouraged the majority to participate in the
production and distribution of ideological ideas (Barbera & Rivero, 2015).
Furthermore, although people tend to connect with others who have similar
beliefs and attitudes when it comes to the social media space, they may be
exposed to opposing political tones, different news, and disagreements, to a
much greater extent than in the physical environment (Fletcher & Nielsen,
2017). The disparities in the findings of the research on the scope and effects
of polarization on social media can be explained by pointing out the difference
between ideological and affective polarization (Barbera, 2020).
While the first implies diverse political viewpoints, the second refers
to the affective level, to the perception of difference and distance towards the
other group, in short, to the sense of belonging and identity. The blending of
social and political contents, feedback, political identity profiling, and the use
of “incendiary” speech on social media contribute to the creation of affective,
psychological polarization. As a result of this process, political and social
identities are strengthened, and the distance from the members of other political
parties becomes greater (Settle, 2018). This may indicate the absence of a real
ideological difference, well-founded and solid political ideas, or the fact that
the real political dispute is driven by effects and investing in a manipulative
game of playing with identity recognition (Barbera, 2020). Although the
algorithmic logic of the functioning of social media is susceptible to the
distribution of misinformation and fake news, the scope of such impact is
overestimated. This is likely the case with echo chambers and filter bubbles as
well. Their effects are limited, whereas the gap between groups with opposing

206 Balkan Social Science Review, Vol. 21, June 2023, 199-217
 Social media algorithms and data management

viewpoints is widening. Groups with right-wing attitudes, in which already

distorted opinions are intensified, are particularly vulnerable to the effects of
misinformation (Stark & Stegmann, 2020). Political ideology has a substantial
impact on polarization on social media with numerous studies confirming the
finding that conservatives are less open to communicating with political
dissidents than liberals (Barbera et al., 2015).

Method

The modern media environment has undergone significant changes,

and as a consequence, the audience’s relationship with the media has changed.
Even though the audience has gained more power in the digital media space, as
evidenced by their ability to create, reshape and share content, media users have
become the subjects of surveillance experiments by global capitalists hoping to
sell ideas and products. A “software culture” (Manovich, 2013) has emerged,
in which algorithms play a central role in tailoring the world of information and
news to the individual’s needs. Also, all of this occurs with media users’
selective awareness or the lack of awareness concerning the algorithmic
programming of their cognitive, emotional, and behavioral spheres.
Taking into account that young people and students are important
social actors who should be able to responsibly create and share content in the
new communication environment, the subject of this paper is students’ attitudes
toward the operation of algorithms. This paper aims to examine how students
view and understand the function of algorithms and data management on social
media platforms.
The main research question is: Is there a difference in the perception of
the mechanisms of the algorithm in an online environment between students of
journalism and communication and public relations on the one side and students
of social policy and social work on the other? These groups of students were
taken as a sample to determine the difference in perception between those who
acquire certain competencies in the field of digital and media literacy during
formal education and those who do not develop these competencies in their
study programs. Accordingly, students of journalism and communication, and
public relations are expected to demonstrate a greater degree of knowledge
regarding algorithmic mechanisms.
In this research, a survey was used as the method of data collection.
The questionnaire includes questions related to the knowledge of basic terms
about the algorithms mechanisms, privacy problems, and protection of personal
data. In addition, through the questionnaire we collected data on how students
interpret the content obtained through the action of algorithms, that is, whether
they understand the principles of the functioning of algorithmic programs on
social networks. The research sample consists of 200 students at the Faculty of
Philosophy, University of Niš, Serbia both from the Department of
Communication and Journalism and the Department of Social Policy and Social
Work - 150 students from the first department and 50 students from the second
department (Table 1).

Balkan Social Science Review, Vol. 21, June 2023, 199-217 207
 Ivana STAMENKOVIĆ, Dušan ALEKSIĆ, Tatjana ĐUKIĆ ŽIVADINOVIĆ

The responses of the students from both departments have been compared to
determine whether there were any differences between them in their
understanding of the algorithmic reality on the platforms, as well as to
determine whether the students were able to deal with the mechanisms of
algorithm operations in the shaping of information flows.

Table 1: The sample according to the department

Number of students
Department
N %
Communication and Public
150 75
Relations/Journalism
Social Policy and Social Work 50 25
Total 200 100

All respondents who participated in the research are social media users.
Most of them (93%) use them daily, while 6.5% tend to access social media
several times a week. Only one respondent has the habit of accessing them only
several times every month. According to the frequency of accessing social
media, the students of these departments do not differ.
It should be pointed out that the sample is not representative, but it
gives us insight into students’ perception of algorithms’ understanding. The
survey was conducted by the authors. All the questionary sessions were
conducted face-to-face.

Results

Of the total number of respondents, 21% of them answered that during

their studies, they dealt with the problem of privacy and the mechanisms of
algorithmic operations online, 18.5% were not certain, while 60.5% answered
that they had not dealt with that issue. By comparing the students of the
departments of Communication and Public Relations, Journalism, and Social
Policy and Social Work, and after applying the chi-squared test, it was
confirmed there was a significant statistical difference between these two
groups (X2 (2, N = 200) = 8.54, p = .013914.). It was determined that the
students of the Department of Communication and Public Relations, as well as
the Department of Journalism, have dealt with this issue to a significantly
greater extent.

208 Balkan Social Science Review, Vol. 21, June 2023, 199-217
 Social media algorithms and data management

Table 2: During your studies so far, have you dealt with the issue of privacy
and algorithmic operation mechanisms online?
Number of students
Yes No Not certain
Department
Devi Total
Empir Theore Devia Empir Theore Devia Empir Theore
atio
ical tical tion ical tical tion ical tical
n
Communication
and Public
36 31.5 0.64 82 90.75 0.84 32 27.75 0.65 150
Relations /
Journalism
Social Policy
6 10.5 1.93 39 30.25 2.53 5 9.25 1.95 50
and Social Work
Total 42 121 37 200
X2 (2, N = 200) = 8.54, p = .013914.

Almost a quarter of students (23.5%) were familiar with the definition

of the concepts “filter bubbles” and “echo chambers”, 13.5% were not certain,
while 63% of students did not know the meaning of these concepts. When the
two groups of students were compared on this issue, there was no statistically
significant difference between them (X2 (2, N = 200) = 0.03, p = .985102).
When asked whether they read terms of use before accessing a website,
an application, or social media, 18% responded affirmatively, and 55.5% stated
they read it sometimes, which was noticed between the two groups (X2 (2, N =
200) = 2.891, p = .235632).
A total of 12% of respondents believe that all users receive the same
content through social media, 65.5% think the opposite, while 22.5% responded
that they were uncertain. In this regard, the difference between the two
examined groups of students was not statistically significant (X2 (2, N = 200) =
0.2488, p = .883027).
The belief that social media are a reflection of reality was held by 6.5%
of the respondents, while there is a slightly higher percentage of those who are
not certain (10%). However, most of them (65.5%) consider that what they see
on social media does not represent reality. There is no statistical difference
between the two groups of students (X2 (2, N = 200) = 2.045, p = .359691).
Only 8.5% of respondents expressed a lack of concern for personal data they
leave on social media, while 10.5% were deeply concerned. However, most of
them (82%) expressed moderate concern. The students of the two surveyed
departments did not differ statistically regarding this issue (X2 (2, N = 200) =
2.045, p = .359691).
Seventy-eight percent of respondents believe that it is difficult to have
control over personal data once they are shared on social media and 6.5%
believe the opposite, while 15.5% remain uncertain. In this sense, the difference
between the two groups showed no statistical importance (X2 (2, N = 200) =
1.7833, p = .40998).

Balkan Social Science Review, Vol. 21, June 2023, 199-217 209
 Ivana STAMENKOVIĆ, Dušan ALEKSIĆ, Tatjana ĐUKIĆ ŽIVADINOVIĆ

Most respondents (69.5%) believe it is unacceptable that social media

or some third party use their personal data, 23.5% are uncertain, while 7% of
them consider this practice acceptable. Regarding this issue, the students of the
departments do not differ in the statistically significant measure X2 (2, N = 200)
= 2.6182, p = .270062).
According to 66% of the respondents, content published by social
media users is used to adjust and personalize information and to gain profit.
Somewhat less than a third of respondents (31%) are not certain, while 3% hold
the opposing view. The departments had no impact on this question (X2 (2, N
= 200) = 0.2437, p = .885269).
By being presented with similar content, conditioned by previous
experiences and interests, 67% of respondents believe they can be denied
different opinions and viewpoints, though 22.5% are uncertain, while 10.5% of
respondents believe that is not the case. After comparing the two groups of
respondents and applying the chi-square test, a statistically significant
difference between the students of the Department of Public Relations and
Journalism and the Department of Social Policy and Social Work was
confirmed (X2 (2, N = 200) = 6.3271, p = .042276) (Table 3). In this regard, the
students of the Department of Social Policy and Social Work express greater
concern on the matter of being denied different viewpoints, while the students
of the Department of Communication and Public Relations express uncertainty
on this issue.

Table 3: Do you think that by being presented with similar content, conditioned
by previous experiences and interests, you can be denied different opinions and
viewpoints?

Number of students

Department Yes No Not certain

Empi Theor Devi Empi Theor Devi Empi Theoret Devi Total
rical etical ation rical etical ation rical ical ation
Communication
and Public
94 100.5 0.42 16 15.75 0 40 33.75 1.16 150
Relations /
Journalism

Social Policy and

40 33.5 1.26 5 5.25 0.01 5 11.25 3.47 50
Social Work

Total 134 21 45 200

X2 (2, N = 200) = 6.3271, p = .042276

210 Balkan Social Science Review, Vol. 21, June 2023, 199-217
 Social media algorithms and data management

Most students (76.5%) agree that the way the algorithm functions can affect the
amount of time they spend on social media, 16.5% are uncertain, while 7% state
that is not the case. There is no statistically significant difference between the
two groups of students (X2 (2, N = 200) = 3.1237, p = .20975).
Only one respondent (0.5%) believes that algorithms do not allow
social media platforms and large companies to manipulate users’ personal data.
On the contrary, 76% believe that such a practice exists, while 23,5% express
uncertainty. In this case, the groups studied did not show a significant
difference of opinion (X2 (2, N = 200) = 0.0054, p = .941425).
Furthermore, 68.5% believe that users’ activities on any social media are
constantly monitored, 23% express uncertainty and 8.5% think there is no
monitoring. In relation to their departments, the students’ opinions do not differ
in this case (X2 (2, N = 200) = 5.7224, p = .057201).
According to 85.5% of respondents, appropriate education and
information can help users to better understand and control their data on social
media. 12.5% express uncertainty, while only 2% believe that is not the case.
Regarding this issue, there is no statistically significant difference among the
students (X2 (2, N = 200) = 1.3639, p = .24287).
As far as the analyzed categories are concerned, there is a statistically
significant difference between the students of Communication, Public
Relations, and Journalism, and the students of Social Policy and Social Work
in only two categories. Also, there is an impression that the students are equally
familiar with the functioning of social media, algorithms, and potential privacy
threats. The reason for this may be found in a critical approach to social media,
both by the students who are media-focused and those who are not.

Discussion

Research aimed at understanding the operation of algorithms among

the student population is rarely conducted. In addition to the lack of data on the
so-called algorithmic literacy among youth and students, another problem is
that the knowledge about algorithmic personalization of content remains
outside the perview of media education (Swart, 2021; Head, Fister &
MacMillan, 2020; Mihailidis, 2018). Given the constant and significant
influence of algorithms on the content that young people encounter in the online
space, the focus on a critical review of information should be enhanced by
training young people to reexamine the mechanisms of content creation and
understand the accuracy and balance of information. Here, the main question
would be: Do all users receive the same information from different sources, or
are they denied a certain part of content, by previous activities and
recommendation systems of algorithms? Also, as part of their study programs,
young people should be prepared for an independent, critical evaluation of
technological and social forces that shape the circulation of information and
news at the present moment (Head et al., 2020). Their informal knowledge
about algorithms might be achieved through the use of media, but this would
depend on the frequency and extent of its usage (Cotter & Reisdorf, 2020). One

Balkan Social Science Review, Vol. 21, June 2023, 199-217 211
 Ivana STAMENKOVIĆ, Dušan ALEKSIĆ, Tatjana ĐUKIĆ ŽIVADINOVIĆ

of the qualitative studies of students from eight American universities and

colleges revealed that among this population there is an awareness of the
omnipresence of information and a hidden mechanism of personalization of
information and news distributed online (Head et al., 2020, p.27).
The research conducted in this paper confirms these results, showing
that 65.5% of students believe that not all users receive the same information
and content through social media. However, only a quarter of the surveyed
students (23.5%) are familiar with the terms “filter bubbles” and “echo
chambers”. Also, 67% believe that if they are presented with similar content as
a result of an algorithmic operation, they will be denied different opinions and
viewpoints. These results are consistent with the research conducted in the
United States, which shows that, according to most students, algorithmically
organized data can jeopardize “representative democracy and the cultivation of
an informed and active community” (Head et al., 2020, p. 27). Many have
expressed discontent and indignation over the large impact of algorithms on
social life, information, and news, as well as the inability to combat gigantic
systems and their control and surveillance mechanisms. Most students
expressed concern about the constant control of personal data, which can have
immeasurable consequences on existing social inequality. For that reason,
many of them apply certain strategies to protect their data and limit the scope
of algorithmic control (Head et al., 2020). Our research, however, shows that
78% of students believe it is difficult to have control over personal data online,
while most students (82%) express moderate concern on this issue. Only 18%
read the terms of use before accessing a website, application, or social media
platform.
The research Experiencing Algorithms: How Young People
Understand, Feel About and Engage with Algorithmic News Selection on Social
Media, which included 22 respondents aged 16-26, focused on the three
dimensions of users’ algorithmic experience: cognitive, affective, and
behavioral. The analysis of the first dimension sought to discover how young
people understand the operation of algorithms. The data showed that context
has an impact on how algorithmic operations are perceived, particularly when
the user’s expectations are not fulfilled, or when the content is confusing, and
there are clear indications of content personalization. In addition, the research
showed that, in some cases, the respondents easily noticed the impact of
algorithms, while on certain platforms it was more complicated to do so. In
summary, context affects the understanding of algorithms, which is determined
by a platform, its characteristics, and the type of content (Swart, 2021, p.5).
Research into the affective dimension of algorithmic experience showed what
types of moods, effects, and sensations are caused by algorithms. They range
from a neutral interpretation of algorithms, used for achieving a certain goal, to
a positive view of algorithms as useful filters of information chaos, to a negative
perception of algorithms used for controlling and censoring information. The
scope of user action about algorithms is thought to be quite limited and the
respondents believe that algorithms are beyond their control. Also, young
people’s knowledge of algorithms does not persuade them to participate in

212 Balkan Social Science Review, Vol. 21, June 2023, 199-217
 Social media algorithms and data management

algorithmic decisions (Swart, 2021). These findings lead to the observation that
young people are, to a certain extent, aware of the impact of algorithms on
information daily, but they have no control over the process of selection and
editing, or personalization of content. Young people do not withdraw from
social media or limit the amount of time they spend online and on various
platforms, according to the research conducted by Head, Fister, and MacMillan,
despite knowing that their activities are monitored and their data traded,
because they find application services useful (Head et al., 2020).

Conclusion

Issues that arise in an algorithmically shaped society, can be roughly

divided into two levels: individual and social. The consequences are visible at
the level of an individual, their understanding of their personality, and their
being, as well as in terms of their relationship to society. Many users are
unaware that the content they receive has already been filtered, processed, and
matched to their previous wishes, needs, aspirations, and intentions. As a
consequence, the boundaries of existing beliefs, attitudes, opinions, and
feelings are strengthened, allowing people to enter a self-affirming zone where
the same or similar thoughts and emotional tones resonate. At the level of
personality, there can be a significant discrepancy between who we are and
what algorithmic systems say about us. In addition, personal data has become
the most valuable product on the market, while the knowledge of ourselves is
built on what algorithms say about us. Generally viewed, the impact of
algorithms on society and the democratic climate becomes even more
prominent, with consequences such as polarization, social fragmentation, the
decline in news quality, the explosive growth of fake news, misinformation,
and the deformation of public discourse. In a transformed social and
communication reality, shaped by the omnipresent branches of artificial
intelligence and algorithmic programs, only those who have developed
appropriate critical potential and who understand the actual situation in
algorithmic culture can function. The research showed that two-thirds of
students of Journalism /Communication and Public Relations, and students of
Social Policy and Social Work from a Faculty of Philosophy, University of Niš,
Serbia, notice the results of algorithmic personalization, filtered selection of
content and news, and the customized display of content on social media
platforms.
Despite the fact that over 70% of students from both departments
comprehend there is constant surveillance of user activities and that control
over personal data online is taken over by large companies and third parties, for
profit, most respondents (82%) express moderate concern for the data they
leave online. This is further confirmed by the fact that only a small percentage
of students (18%) almost always read the terms of use of a website, application,
or internet service. Also, there is a statistically significant difference between
the students of Journalism, Communication, and Public Relations and those of
Social Policy and Social Work regarding their knowledge of privacy issues and

Balkan Social Science Review, Vol. 21, June 2023, 199-217 213
 Ivana STAMENKOVIĆ, Dušan ALEKSIĆ, Tatjana ĐUKIĆ ŽIVADINOVIĆ

mechanisms of online algorithms during their studies. Students of Journalism,

Communication, and Public Relations become more familiar with these
concepts and social phenomena that occur as a result of algorithm operations,
so that understanding the algorithmic logic online, which is recognized by the
majority of respondents from both departments, can be linked to experience on
multiple social media platforms. Another significant difference between the
respondents from the two departments is in their attitudes toward being denied
different opinions, views, and interpretations of reality. which is more prevalent
among the students of Social Policy and Social Work. Although a certain
number of respondents lack theoretical knowledge about the concepts most
frequently associated with algorithms, most respondents (85.5%) agree with the
viewpoint that appropriate education and information can create conditions that
will help individuals and society to control their data online, instead of being
controlled by the network.
As indicated in the methodological part, this sample is not
representative, but it is relevant, and a proposal for future research may be a
larger sample that will include respondents from various fields of study. Also,
future research should include information on the number of social networks
on which respondents have accounts, as this data would speak to the breadth of
experience they gain using these platforms. This can be reflected in the
understanding of the operation of algorithmic practice. In addition, it would be
significant to examine members of different age categories, to examine and
compare approaches to the issue from a different age perspective.

References

Aleksić, D., & Stamenković, I. (2018). The language of media in the digital
age. Balkan Social Science Review, 12, 101-113.
Barbera, P., Jost, J. T., Nagler, J., Tucker, J. A. & Bonneau, R. (2015). Tweeting
from left to right: Is online political communication more than an echo
chamber? Psychological Science 26(10), 1531–1542.
https://doi.org/10.1177/095679761559462
Barbera, P. & Rivero, G. (2015). Understanding the political representativeness
of Twitter users. Social Science Computer Review 33(6), 712–729.
https://doi.org/10.1177/0894439314558836
Barbera, P. (2020). Social Media, Echo Chambers, and Political Polarization.
In N. Persily & J. A. Tucker (Eds.) Social Media and Democracy: The
State of the Field (pp. 34-55). Cambridge University Press
https://doi.org/10.1017/9781108890960
Cetina Presuel, R., & Sierra, J. M. M. (2019). Algorithms and the news: social
media platforms as news publishers and distributors. Cetina Presuel,
R., & Martínez Sierra, J.(2019). Algorithms and the News: Social
Media Platforms as News Publishers and Distributors. Revista De
Comunicación, 18(2), 261-285.
Cheney – Lippold, J. (2017). We are data. New York: New York University
Press.

214 Balkan Social Science Review, Vol. 21, June 2023, 199-217
 Social media algorithms and data management

Chitra, U. & Musco, C. (2019). Understanding Filter Bubbles and Polarization

in Social Networks. Available at: https://arxiv.org/abs/1906.08772 Last
accessed on February 17, 2023.
Cotter, K., & Reisdorf, B. (2020). Algorithmic knowledge gaps: A new horizon
of (digital) inequality. International Journal of Communication. 14,
745–765.
Čejko, M. (2019). Superpovezani (Super connected). Beograd: Clio.
Đukić, T. & Stamenković, I. (2019). "Političke kampanje na mreži: Kriza
demokratije?" ("Online political campaigns; The crisis of democracy")
In: G. Đigić i L. Milošević Radulović (eds.) Savremeno društvo i nauka
(Contemporary society and science).Tematski zbornik radova sa
konferencije Nauka i savremeni univerzitet, knjiga 1 (Thematic
collection of papers from a science conference Science and
Contemporary University) (pages: 164-174). Niš: Faculty of
Philosophy, UDK 324:004.738.5; 316.774:324 ISBN 978-86-7379-
516-4 COBISS.SR-ID 280737292
Edelson, L., Nguyen, M.-K., Goldstein, I., Goga, O., Lauinger, T., and McCoy,
D. 2021. "Far-Right News Sources on Facebook More Engaging."
Medium, from https://medium.com/cybersecurity-for-democracy/far-
right-news-sources-on-facebook-moreengaging-e04a01efae90 Last
accessed on June 6, 2023.
Fletcher, R. & Nielsen, Kleis R. (2017). Are News Audiences Increasingly
Fragmented? A Cross-National Comparative Analysis of Cross-
Platform News Audience Fragmentation and Duplication. Journal of
communication. Vol. 67, Issue 4, 476-498. DOI:
https://doi.org/10.1111/jcom.12315
Head, J. A, Fister, B. & MacMillan, M. (2020). Information literacy in the age
of algorithms: Student experiences with news and information, and the
need for change. Available at:
https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&c
d=&cad=rja&uact=8&ved=2ahUKEwjp8bC9hJT3AhUY_7sIHZstAQ
UQFnoECAIQAQ&url=https%3A%2F%2Fprojectinfolit.org%2Fpub
s%2Falgorithm-study%2Fpil_algorithm-study_2020-01-
15.pdf&usg=AOvVaw0sIhr5QrHeA4diwLDsF_nA Last accessed on
February 17, 2023.
Livingstone, S. (2012) Exciting moments in audience research – past, present
and future. In: H. Bilandzic, G. Patriarche & P. Traudt (Eds.) The social
use of media: cultural and social scientific perspectives on audience
research (pp. 257-274). ECREA Book Series. Intellect Ltd, Brighton,
UK, ISBN 9781841505121
Manovich, L. (2013): Software Takes Command. New York, London:
Bloomsbury Publishing.
Marantz, A. 2020. "Why Facebook Can’t Fix Itself." The New Yorker, from
https://www.newyorker.com/magazine/2020/10/19/why-facebook-
cant-fix-itself Last accessed on June 6, 2023.

Balkan Social Science Review, Vol. 21, June 2023, 199-217 215
 Ivana STAMENKOVIĆ, Dušan ALEKSIĆ, Tatjana ĐUKIĆ ŽIVADINOVIĆ

Mihailidis, P. (2018). Civic media literacies: Re-imagining engagement for

civic intentionality. Learning, Media and Technology, 43(2), 152–164.
https://doi.org/10.1080/17439884.2018.1428623
Mitrović, M. (2019). Uloga države i internet intermedijatora u zaštiti prava
internet korisnika, doktorska disertacija (The role of the state and
internet intermediaries in protecting internet user's right, doctoral
dissertation). Beograd: Fakultet političkih nauka.
Obradović, N., & Mitrović, M. (2019). Civic engagement and the
internet. Balkan Social Science Review, 14, 225-241.
Pariser, E. (2011). The filter bubble: What the Internet is hiding from you.
Penguin UK.
Presuel Cetina, R. & Martinez Sierra, J. M. (2019). Algorithms and the News:
Social Media Platforms as News Publishers and Distributors. DOI:
https://doi.org/10.26441/RC18.2-2019-A13.
Riemer, K., & Peter, S. (2021). Algorithmic audiencing: Why we need to
rethink free speech on social media. Journal of Information
Technology, 36(4), 409-426.
Settle, J. E. (2018). Frenemies: How Social Media Polarizes America.
Cambridge University Press.
Spohr, D. (2017). Fake news and ideological polarization: Filter bubbles and
selective exposure on social media. Business information
review, 34(3), 150-160.
Stark, B. & Stegmann, D. M. A. (2020). Are Algorithms a Threat to
Democracy? The Rise of Intermediaries:A Challenge for Public
Discourse. Available at: https://algorithmwatch.org/wp-
content/uploads/2020/05/Governing-Platforms-communications-
study-Stark-May-2020-AlgorithmWatch.pdf Last accessed on
February 17, 2023.
Sunstein, C. R. (2001). Echo Chambers: Bush v. Gore, Impeachment, and
Beyond. Princeton University Press.
Sunstein, C. R. (2018). #Republic: Divided democracy in the age of social
media. Princeton University Press.
Swart, J. (2021). Experiencing Algorithms: How Young People Understand,
Feel About, and Engage With Algorithmic News Selection on Social
Media. Social Media + Society, DOI:
https://doi.org/10.1177%2F20563051211008828.
Šušnjić, Đ: (2007). Dijalog i tolerancija (Dialogue and Tolerance). Beograd:
Čigoja štampa.
Vajdijanatan, S. (2018). Antidruštvene mreže (Antisocial networks). Beograd:
Clio.
Vaidhyanathan, S. 2019. "The Real Reason Facebook Won’t Fact-Check
Political Ads." The New York Times, from
https://www.nytimes.com/2019/11/02/opinion/facebook-
zuckerbergpolitical-ads.html Last accessed on June 6, 2023.
Zubof, Š. (2020). Nadzorni kapitalizam (Surveillance Capitalism). Beograd:
Clio.

216 Balkan Social Science Review, Vol. 21, June 2023, 199-217
Copyright of Balkan Social Science Review is the property of Balkan Social Science Review
and its content may not be copied or emailed to multiple sites or posted to a listserv without
the copyright holder's express written permission. However, users may print, download, or
email articles for individual use.