Bakamo Global Social Intelligence
A Psychographic Divide in Hungarian Political Discourse: Tisza Supporters Show a Significantly Higher Inner-Directed Share — and Why It Matters for April 12
Budapest — Bakamo, a global social intelligence agency, publishes findings from a large-scale semantic analysis of Hungarian social media conversations ahead of the April 12, 2026 parliamentary elections. The research reveals a measurable psychographic divide between supporters of Tisza (Péter Magyar) and supporters of Fidesz (Viktor Orbán) — one with significant implications for understanding the resilience and trajectory of each support base.
Bakamo conducted a qualitative semantic analysis of over 10,000 social media expressions (generated between March 11–18). The finding — subsequently validated at scale by multi-stage AI classification — is consistent: Tisza supporters produce a substantially higher share of inner-directed expressions than Fidesz supporters, with a gap of 12.4 percentage points. Both supporter groups are predominantly outer-directed, as is typical for political discourse on social media. But the difference in inner-directed share is significant, persistent across sources, and — we argue — carries predictive weight.
Inner-Directed“03.15 ITT VOLTAM — 04.12 OTT LESZEK”
I was here on March 15 — I will be there on April 12. Personal commitment anchored in individual agency.
Outer-Directed“Együtt erő vagyunk! 🇭🇺”
Together we are strong. Group identity, collective belonging, social validation.
The Framework
What Inner-Directed and Outer-Directed Mean
These terms originate in social psychology and communication theory. David Riesman's foundational work The Lonely Crowd (1950) distinguished between individuals whose orientation is guided primarily by internalised values and those guided primarily by signals from their social environment. Julian Rotter's concept of locus of control (1966) draws a related distinction: individuals with an internal locus of control attribute outcomes to their own actions; those with an external locus attribute them to forces beyond themselves.
We apply this framework not to classify people, but to classify the language of their organic social media expressions. When a person posts on social media without being asked, the structure of their language reveals the motivational frame — the thing that caused them to speak.
An inner-directed expression references the self as the source: personal feelings, lived experience, individual values, direct personal consequence. An outer-directed expression references the social world as the trigger: group norms, collective identity, abstract systems, national narratives, or external validation.
The analytical question is not who the person is — but what frames the expression, and how that frame is encoded in the language of the post.
“How many cases like this are there within families... Like in our case, it's the older generation. Luckily there's only one... Everyone else is a Tisza Party voter.”
Hány ilyen van családon belül.. Mondjuk a mi esetünkben az idősebb generáció. Szerencsére csak egy.. Mindenki más Tisza párt szavazó.
“The Tisza crowd already soiled themselves with all things Ukrainian last year alongside the liposuctioned luxury general, their ‘Ukrainian brother’ Roland Tseber, and their MEPs parading around in Ukrainian t-shirts; and this year it's with Zelensky's threatening remarks and the shutting off of the oil pipeline. Everyone has seen it, heard it, and knows it. Even in Brussels...”
A tiszások már tavaly összeukránozták magukat a zsírleszívott luxi-tábornokkal, ukrán testvérükkel Tseber Rolanddal, az ukrán pólóban díszelgő EP-képviselőikkel, idén meg Zelenszkij fenyegető kijelentéseivel és az olajvezeték elzárásával. Mindenki látta-hallotta-tudja. Még Brüsszelben is...
The contrast is illustrative. One post anchors the stakes in a concrete family situation — a personal reality. The other anchors them in a chain of national-level associations and an appeal to shared knowledge (“everyone saw-heard-knows”). Our research shows this pattern — self vs. system, personal consequence vs. collective mythology — runs consistently through the data.
Key Findings
A consistent, cross-platform psychographic split.
Fidesz Supporters Social Media Conversations
7.2% inner-directed, 92.8% outer-directed
Tisza Supporters Social Media Conversations
19.6% inner-directed, 80.4% outer-directed
Inner-directed share — Brandwatch data, Mar 11–19, 2026
From an initial N=10,000 social media mentions, 69.6% were identified as relevant, of which 54.7% supported Fidesz and 45.3% were in support of Tisza.
Across the dataset, Tisza supporters produce a meaningfully higher share of social media expressions rooted in personal conviction, direct life experience, and individual agency — approximately 12.4 percentage points more than Fidesz supporters. Fidesz supporters more frequently produce expressions rooted in collective identity, national threat narratives, and appeals to shared group knowledge.
Both groups are majority outer-directed. This is expected: political social media is inherently a space of public positioning, and outer-directed framing is the default register. What is analytically significant is not the absolute level but the relative difference — and the consistency of that difference across sources and platforms.
Inner-directed expressions communicate a higher degree of what psychologists term self-concordance — an alignment between a person's inner needs and the position they publicly take. When someone speaks from personal experience and direct conviction, the expressed commitment is anchored in something the person has processed individually. This is not more virtuous — it is more structurally stable.
Analysis
Why the Inner-Directed Gap Matters
In qualitative consumer and communication research, inner-directed expression is understood to be contagious. When a person articulates a position through personal experience — “this is what happened to me, this is why I changed my mind” — the language carries an authenticity signal that activates personal recognition in others. The listener does not need to share the speaker's group identity. They need only recognise their own situation in the speaker's words.
“If Fidesz wins, I'll be doing everything I can so we can get the hell out of here as soon as possible. Because if this kind of politics continues, we'll crash out of the EU, and then there will be starvation and ruin, and I don't want to be here with the family.”
Ha a fidesz nyer azon leszek h minél hamarabb elhúzhassunk innen. Mert ha ez a politika folytatodik kiesünk az eu-ból akkor pedig éhezés és pusztulás lesz én pedig nem szeretnék itt lenni a családdal
This is the mechanism through which brands grow organically and through which political movements recruit beyond their base. Inner-directed expression is recruitment language. It converts.
Outer-directed expressions work differently. They reinforce belonging among those who already identify with the group — “we all know,” “any real Hungarian,” “everyone saw what happened.” This is retention language. It holds the coalition together through social proof and conformity cues. But it does not create new supporters, because it requires prior group membership to resonate.
“We are definitely going to win.... Go Fidesz, go Hungarians!!! I am a PROUD Hungarian Patriot!!!!!”
Egyértelműen fogunk nyerni.... Hajrá Fidesz, hajrá Magyarok!!! BÜSZKE Magyar Hazafi vagyok!!!!!
Implications
This Asymmetry has Direct Electoral Implications
The fact that Tisza supporters are 2.8× more likely to post inner-directed content means public discourse contains a larger pool of expressions that can resonate with voters outside the existing Tisza base — including privately dissatisfied Fidesz voters who have not yet found a personal reason to switch. When such a voter encounters a Tisza supporter describing a concrete experience — a family argument, a lost opportunity, a moment of personal clarity — it can activate recognition: that is my situation too.
Fidesz's predominantly outer-directed discourse, meanwhile, holds its base through collective identity and shared narrative — but this same mechanism makes its supporters structurally vulnerable. Outer-directed loyalty is loyalty to the perceived winning team, to the dominant narrative, to the sense that “everyone around me thinks this way.” When that perception begins to shift — through polling leads, through visible roadshow attendance, through the sheer volume of inner-directed Tisza expression appearing in their feeds — the outer-directed frame offers no personal anchor to resist the pull.
We therefore interpret the inner/outer gap not as a static description but as a predictive indicator that Tisza's discourse is structured to grow — and we expect this tendency to manifest in widening Tisza's lead in survey-based polling.
Transparency
A Note on Interpretation and Limitations
This is a qualitative analysis validated at scale through automated classification. Our findings are not representative in the traditional polling sense — we make no claim to statistical sampling of the Hungarian electorate. What we offer is a semantically grounded reading of organic social media expression, developed through deep human analysis and confirmed through an AI-assisted methodology.
The value of this approach lies not in counting heads, but in reading the motivational texture of public discourse — something traditional polling does not capture. Polls measure stated intention under prompting. We measure the structure of what people say voluntarily, to their peers, without a researcher asking the question. Both are valid; they answer different questions.
- Platform composition: Social media users are not demographically identical to the electorate. Engagement patterns vary across Facebook, YouTube, TikTok, X, and media comment sections. We sampled broadly but do not claim proportional representation.
- Source effects: Comments under pro-government YouTube channels and comments on opposition Facebook pages carry different baseline characteristics. We have analysed across sources but readers should be aware of this structure.
- Temporal scope: This analysis reflects a snapshot of discourse in the weeks prior to publication. Social media discourse is dynamic and may shift as the election approaches.
- Classification boundaries: Inner/outer directedness operates on a spectrum. Some expressions contain elements of both. Our classification captures the dominant signal; borderline cases are unavoidable.
Methodology
How the Research Was Conducted
Our analysis is grounded in semantic content analysis — a qualitative research tradition concerned with identifying the motivational and psychological signals encoded in natural language.
Stage 1
Human Analysis
Bakamo's senior analysts conducted an in-depth qualitative analysis of a substantial sample of organic social media posts. Working within the inner/outer directedness framework — rooted in the theoretical traditions of Riesman (1950) and Rotter (1966) — analysts read, interpreted, and classified posts to establish the core finding. Data sources included API-enabled and opaque social media channels including TikTok, Instagram, Facebook, YouTube and X.
Stage 2
AI-Assisted Validation at Scale
To test whether the pattern identified by human analysts held across a larger corpus, we collected over 15,000 comments and posts via the Brandwatch social listening platform. This data was processed by a custom AI platform utilising Google's Gemini LLM technology through a three-stage automated pipeline: (1) relevance filtering, (2) political alignment classification, and (3) inner/outer directedness scoring.
Bakamo complies with GDPR and the ESOMAR ethics codex. Quotes in this report have been slightly modified to retain meaning and tonality while protecting social media users' privacy. Data Sources: Brandwatch, Mar 11–19, 2026.
Earlier Work
Bakamo's Election Research Has Been Cited by International Media
Previous Work
The 2017 French Presidential Election
Bakamo's first public election study analysed the psychographic texture of social media discourse during the 2017 Macron–Le Pen runoff, mapping how organic expressions flowed through the media relay before reshaping mainstream narratives. The methodology and findings were picked up and cited internationally.
About Bakamo
Reality-first intelligence, at scale.
Bakamo is a global social intelligence agency that listens to the open web to reveal truths that conventional research cannot reach. Our methodology combines technology, human analysis, and academic rigour to surface the motivational and cultural drivers buried beneath surface-level opinion.
It is this foundation in reality — in what people actually say, not what they say when asked — that gives Bakamo's psychographic analysis its distinctive credibility.