The research, conducted by Miloš R. Milenković and Uroš G. Miljković of the Strategic Research Institute at the University of Defence in Belgrade, sits at the intersection of media psychology and security studies. The authors argue that cognitive resilience—the capacity of citizens to recognize manipulation and preserve trust in reliable information—should be treated as a functional component of Serbia’s total defence concept, not merely as a matter of media literacy. Serbia’s National Security Strategy already warns that the spread of false news within hybrid and information warfare may impair the functioning of elements of the defence system.

Data were collected in April 2026 through an online survey platform, using convenience sampling supplemented by a snowball technique across social media platforms and professional networks. Only respondents who maintained active social media profiles were included, on the grounds that online platforms are the primary environments in which users encounter news from both official and unofficial sources. The final sample of 202 adults ranged in age from 18 to 69, with a mean age of 37.7; 57.9 percent identified as female and 42.1 percent as male. More than 65 percent held a graduate, master’s, or doctoral degree, a demographic skew the authors flag as a central limitation.

Susceptibility was measured with the Online Misinformation Susceptibility Scale (OMISS), a nine-item instrument that assesses how frequently individuals critically evaluate online content rather than testing factual knowledge about disinformation. Items are rated on a five-point scale, with total scores ranging from 9 to 45 and higher scores indicating greater susceptibility; a proposed cut-off of 23 separates high from low susceptibility. In this sample, internal consistency was 0.87 Cronbach’s alpha. The mean score was 25.48 with a standard deviation of 7.17, and 60.9 percent of participants (123 of 202) fell at or above the cut-off, placing them in the high-susceptibility category.

Correlational analysis complicated any simple explanation based on schooling. Age showed no statistically significant association with susceptibility scores (Spearman’s rho of 0.027, p = 0.706). Educational level showed a weak but statistically significant negative correlation (rho = −0.172, p = 0.014), meaning higher education was linked to somewhat lower susceptibility—but the authors stress that the association is weak and that a majority of this highly educated sample was nonetheless classified as highly susceptible. They suggest that even well-educated individuals can be swayed by the professional appearance of websites and the objective tone of online content, which may mask the absence of adequate sourcing.

The second instrument, the Perceived Media Information Environment Questionnaire (PMIEQ), was developed specifically for this study and covers five domains, from perceived reliability of digital news to views on institutional responses. It deliberately uses the phrase “false and misleading information” rather than “disinformation” to avoid response bias tied to the connotation of deliberate intent. The results paint a picture of an audience that is simultaneously skeptical and heavily exposed. Some 70.3 percent of respondents rated news on digital platforms as reliable only to a lesser extent, and 14.4 percent considered it not reliable at all, while only 1.0 percent viewed it as completely reliable. Meanwhile, 58.9 percent reported encountering suspected false or misleading information several times a week or every day.

Perceptions of who shapes media content were strongly polarized along ownership and geopolitical lines. A combined 87.1 percent agreed that media owners exert greater influence over reporting than professional ethics, with 48.5 percent completely agreeing. On foreign actors, 56.4 percent agreed that foreign and international actors wield more influence than professional ethics. Asked which foreign interests Serbian media promote, respondents gave Russia the highest share of “extremely large extent” ratings at 35.6 percent, followed by China at 27.7 percent, the United States at 20.8 percent, and the European Union at 24.8 percent. Turkey and regional neighbors received markedly lower ratings, with 21.8 percent saying Turkish interests are not promoted at all.

The perceived prevalence of false and misleading news also varied sharply by topic. Nearly half the sample (47.0 percent) judged such content to be present to a dominant extent in domestic politics, and roughly three-quarters perceived high prevalence in international politics and security and in national security. Economy and finance followed a similar pattern. By contrast, science and technology and especially sports were seen as far less affected: 41.1 percent of respondents said false news is not present at all in sports coverage. The authors interpret this differentiation as evidence that respondents do not hold uniformly negative views of media content but recognize that misinformation clusters where political, economic, and security stakes are highest.

On responsibility and remedies, respondents leaned heavily toward state action. The largest share (40.6 percent) attributed primary responsibility for preventing the spread of false information to the state and its institutions, followed by media editors at 26.2 percent and journalists at 22.8 percent; only 8.4 percent named consumers themselves and 2.0 percent named educational institutions. Notably, 77.2 percent supported creating an official state body to identify and expose false and misleading news—a finding the authors read as substantial public acceptance of treating disinformation as a security issue, while cautioning that any such measures would need clear legal safeguards, media pluralism, and protection of freedom of expression.

The authors are explicit that the findings are exploratory and preliminary. The convenience sample deviates substantially from national demographics, recruitment through social media limits generalizability, and the analysis was restricted to descriptive statistics. They also caution that the OMISS cut-off was derived from a general Greek sample and has not been independently validated in Serbia, so the susceptibility classifications should be interpreted with care. Self-report instruments add further potential bias. Still, the pattern they document—high self-assessed exposure to dubious content, low trust in domestic media, and weak verification habits—aligns with earlier national studies, including a 2025 telephone survey of 800 respondents in which 47.9 percent rarely or never cross-checked information across sources and 90.3 percent rarely or never used fact-checking sites. The authors recommend embedding information security and cognitive resilience in defence planning, systematic media literacy education, stronger editorial independence for public service media, and national capacities for early detection of coordinated disinformation campaigns, subject to democratic oversight.

Subject of Research: Susceptibility to online misinformation and perceptions of the media information environment among adults in Serbia, in the context of societal resilience and total defence

Article Title: Societal resilience and total defence in the digital age: the cognitive dimension of disinformation and its implications for the national security of Serbia

Article References: Milenković, M. R., & Miljković, U. G. (2026). Societal resilience and total defence in the digital age: the cognitive dimension of disinformation and its implications for the national security of Serbia. Frontiers in Political Science, 8, 1897340. https://doi.org/10.3389/fpos.2026.1897340

Image Credits: AI Generated

DOI: 10.3389/fpos.2026.1897340

Keywords: misinformation, disinformation, Serbia, cognitive resilience, total defence, media literacy, OMISS, hybrid threats, national security, survey research, foreign influence, media trust