Justifying an Invasion: When Is Disinformation Successful?

Abstract

Conventional wisdom suggests that social media, especially when used by authoritarian powers with nefarious aims, leaves citizens of democratic countries vulnerable to psychological influence campaigns. But such concerns overlook predispositions among recipients of false claims to reject (or to endorse) conspiratorial narratives. Analyzing responses from a survey fielded in 19 countries, we find that it is a preexisting conspiracy outlook at the individual level, more so than media diets, which consistently predicts rating Russia's pretenses for the invasion as more accurate. In all countries, individuals who view the world in general with a conspiratorial mindset are more likely to believe war-related disinformation. Receiving news via YouTube, Facebook, or TikTok is also associated with greater belief in Russian narratives justifying the invasion in several countries, but these relationships are weaker compared to those observed for conspiracy thinking. Without downplaying a potential positive role for media interventions, the findings highlight the importance of a deeper understanding of the underlying causes of conspiratorial thinking.

Keywords: Conspiracy thinking; Russia; Ukraine; conflict; public opinion; social media.

Figure 1.

Share of respondents endorsing narratives justifying the invasion.

Figure 2.

Conspiracy thinking across the sampled countries (the densities show distributions of the additive index, rescaled to range from 0 to 1; the vertical lines denote the median value for each country). The wording of four survey instruments used for these calculations, and distributions of item-by-item responses by country are shown separately in the Supporting Information (Figure SI.1).

Figure 3.

Coefficients from country-by-country regressions (using the richest specification, the pooled version of which is shown as column (6) in Table 1). Each row contains estimates from a separate model, and specific predictors of interest are pulled into six panels to facilitate cross-country comparisons. The outcome variable (average belief in disinformation narratives) ranges from 1 to 5, and all predictors are scaled to range from 0 to 1. To preserve space, some predictors (other digital platforms) and control variables (education, age, gender, political interest, and political orientation) are not displayed.

Figure 4.

Average predicted beliefs in disinformation by news consumption via social media (interacted with the conspiracy thinking predisposition). The predictions are based on a model where conspiracy thinking is interacted with both social media consumption and country dummy variables. Controls include: political orientation (incl. its squared value), political interest, news consumption via traditional channels (TV and newspapers), education, gender, and age categories (coded in the same way as shown in Table 1).