Digital Media Discourse and the Secularization of Germany: A Textual Analysis of News Reporting in 2020–2024

Abstract

This article examines how digital media discourse mediated Germany’s secularization between 2020 and 2024. Drawing on a corpus from Der Spiegel, Die Zeit, and other outlets, this study combines text mining, sentiment profiling, and frame analysis. Mentions of “faith” and “religion” rose sharply, while reporting stabilized around a neutral tone with a slight positive drift. Coverage centered on public order and institutional roles, relocating religion from moral authority to a civic function. Time-series evidence shows that peaks in discourse preceded increases in non-affiliation by about two months, suggesting a spiral relation in which media both mirror and accelerate secularization. The analysis identifies a three-stage mediating mechanism: communicative reconfiguration, which decentralizes religious narration; cognitive shaping, where neutral framing encourages functional evaluation; and feedback reinforcement, where expanding secular discourse normalizes non-religion. Generational and regional contrasts sharpen this process. Nearly 60% of Germans aged 18–30 identify as unaffiliated, and Eastern regions display both higher secularization and more critical frames than the West. This study contributes empirical evidence for how digital discourse structures belief change in a high-penetration society and extends secularization and mediatization theories by showing how discursive circulation links structural transformation with individual cognition in the digital age.

1. Introduction

1.1. Digital Penetration and Belief Transformation: The Historical Context of Secularization in Germany

Germany is fully embedded in the digital age. Internet penetration reaches 93.3%, with roughly 77.7 million users. The figures do more than index diffusion; they mark digital media as infrastructure—platforms for information flow, public deliberation, and everyday interaction on a national scale. Daily routines confirm the shift; 80% of residents go online each day, 95% at least once per week. What once seemed dispensable has now become a necessity, shaping the intake of news, social relationships, consumers’ choices, and even the scaffolding of an individual’s and a collective worldview. Social platforms remain comparatively less central for German users than for some of their Western peers, yet Facebook sustains a weekly active rate near 31% and Instagram sustains a rate measuring about 15%; among younger cohorts, these arenas amplify topics, accelerate diffusion, and intensify networked exchange. It is worth noting that religious discourse on digital media in Western countries shows significant cross-national differences. G. Perreault and Montalbano (2023) found that, in the United States, some journalists focused on how marginal religious groups challenged the dominance of Protestantism, while conservative media exacerbated religious polarization through narratives of “Christianity versus Islam” (Montalbano, 2019). However, existing research has mostly concentrated on content analysis while paying little attention to those who actually construct the text—whether ordinary journalists or religion-specialized professionals. G. Perreault and Montalbano (2023) also point out that U.S. religious reporting is dominated by generalist journalists, whose lack of religious literacy tends to amplify discursive extremity. By contrast, there is still no study that clearly identifies the profile of reporters in German religious coverage. If German reporters are more professionally specialized, this may help explain the absence of extreme religious discourse in German media. Current cross-national research also lacks analytical models that explicitly link journalistic professionalism to discursive patterns. Montalbano (2019) and G. P. Perreault (2023) emphasized that the key variable is the identity of a journalist, which can create discourse differences. This indicates that when interpreting German media, the professional composition of the author group should be taken into consideration. This may constitute a latent factor behind the moderation of Germany’s religious coverage and provide a conceptual lens for subsequent analysis. Unlike the situation in the United States, digital media in Germany does not exhibit extremist religious discourse. This feature is closely related to the cautious attitude towards religious conflicts in history and “objective and neutral” journalistic ethics. This creates a mild discourse environment, which is consistent with the ongoing secularization process in Germany and echoes Davie et al.’s (2022) observation of the digitalization of the religion–secularization interaction in Europe. As religious debates, expressions of faith, and secularization disputes become increasingly digitalized, the mechanism of the transformation of traditional beliefs is facing reconstruction, and this trend simultaneously provides a new analytical framework for understanding the dynamic logic of secularization.

The second movement unfolds in parallel. Germany’s secularization trajectory has reached a milestone stage. In 2024, 47% of people believed they had no religious beliefs, surpassing the combined total of Catholics (about 28%) and Protestants (about 23%) for the first time. Since unification in 1990, the proportion of people without religious beliefs has more than doubled (from 22% to 47%), with a particularly notable intergenerational difference—among those aged 18 to 30, this proportion is close to 60%. Martin’s characterization of secularization as non-linear—long periods of accumulation punctuated by episodic leaps—fits the curve (Martin, 2017); Pickel et al. (2024) document the persistent East–West asymmetry that extends through socialist legacies, with near 55% of people in Saxony reporting themselves as unaffiliated versus Bavaria, with around 38% (Pickel et al., 2024).

Timelines overlap in telling ways. Between 2010 and 2020, Internet penetration rose 75–90%, while non-affiliation moved 30–42%. This coincidence poses a central problem. Is digital news discourse a catalyst, a mirror, or an intermediary in this secularization process? Clarifying that role speaks directly to the drivers of belief change under digital modernity.

1.2. Research Significance: Bridging the Gap in Studies on the Interaction Between Digital Media and Secularization

Most scholarship follows how religious institutions use the Internet—image work by the Catholic Church, communication strategies on platforms—while few studies trace, with data at scale, how digital media discourse interacts with the rise in the unaffiliated. Germany offers a distinctive test bed precisely because a mature digital ecology coincides with a rapid secular shift. By analyzing 2020–2024 news texts from leading German outlets, this study targets the mechanisms that connect digital discourse to growth in non-religious identification. The contribution is twofold: theoretically, it extends secularization and mediatization frameworks into a socio-digital setting with contemporary evidence; empirically, it supplies a patterned account of how discourse shapes cognition and identification in a society where Internet penetration (93.3%) and non-affiliation (47%) advance in tandem.

1.3. Core Research Questions: Interactive Charateristics Between Digital Media Discourse and the Secularization Process

Based on the research background and significance, this section clarifies the core research questions, corresponding hypotheses, and their logical relationships, forming a “Question–Hypothesis–Verification” loop that focuses on the interaction between digital media discourse and the secularization process in Germany.

1.3.1. Evolution of Discursive Features

Research Question: Between 2020 and 2024, how did the discursive characteristics of “religion” and “non-religion” topics—such as keyword frequency and sentiment orientation—evolve in German mainstream digital media?

Hypothesis 1 (H1).

As the proportion of non-religious population increases, the frequency of keywords such as “religion” and “belief” will increase significantly. The emotional orientation will shift from “implicit criticism” to “neutral tolerance”, with neutral emotions gradually taking the lead and the distribution of positive and negative emotions becoming more balanced.

Rational: Drawing on Davie et al. (2022) and the concept of “digitalization of the religion-secularization interaction field” in Europe, the transformation of social belief structure is expected to drive high-frequency media discussions and attitude adjustments, while Germany’s objective and neutral news ethics curb extreme emotions.

1.3.2. Temporal Correlation

Research Question: Does the peak of media discourse on religious belief precede or follow the increase in the non-religious population?

Hypothesis 2 (H2).

Media discourse intensity (e.g., keyword frequency and topic salience) will lead the growth in the non-religious population by approximately two to three months. Concentrated media attention will occur before a significant rise in non-religious identification.

Rationale: According to Evans et al. (2022) and his “information exposure—Attitude Change—Behavioral Practice” model, individuals first come into contact with multiple perspectives through the media, then adjust their attitudes, and only then transform their identities—resulting in a time lag. Gauthier (2020) also confirms this lead–lag relationship in European contexts.

1.3.3. Framing Effects and Mechanisms

Research Question: How do media framing strategies around “religion” and “non-religion”—such as “public order” and “institutional roles”—shape public cognition and influence the secularization process?

Hypothesis 3 (H3).

A higher proportion of public order and institutional role frames will encourage the public to evaluate religion through instrumental rationality, reducing reliance on traditional religion. This framing proportion is positively correlated with the growth of the non-religious population, with particularly pronounced effects among the 18–30 age group.

Rationale: Stolz and Könemann (2016) argue in their “religion–secular competition theory” that media framing reflects discursive power struggles. Integrating religion into a secular functional evaluation framework weakens its “unquestionable” status. Younger audiences, with more frequent digital exposure, are especially susceptible to this effect.

These three research questions are logically progressive. Question 1 identifies what media discourse looks like, laying a foundation for subsequent analysis; Question 2 examines whether media or demographic shifts come first, providing temporal evidence; and Question 3 explains how media shapes secularization, moving from descriptive and temporal to mechanistic and causal levels. Together, they present a comprehensive picture of the media–secularization interaction and set the stage for the methodological design.

1.4. Key Concepts

Religion lacks a unified definition across contexts. Smith (2001) criticized the definition of essentialism, proposing that religion is a dynamic structure involving the relationships among humanity, transcendence, institutions, culture, and public affairs. This study defines religion in the German context across three dimensions: (1) subjective belief (individual sacred cognition); (2) institutional structure (the role of religious organizations); and (3) public discourse (media representation of religious topics). Particular attention is given to its intersections with social structures (Höpflinger, 2012).

Secularization is not equivalent to “religious opposition/absence” or a necessary outcome of modernity. Smith (2001) identified it as a dynamic process of competition between religion and secularism, and Anderson (2015) emphasized its historical contingency and regional heterogeneity (for example, the East–West differences in the non-religious population in Germany). In this study, secularization is defined through three dimensions: (1) structural (the differentiation of religion from other social domains), (2) ideational (the transformation towards secular rationality), and (3) behavioral (the diversification of belief choices). The concept of secularization is the spiral interaction between religious and secular forces, mediated by digital media and shaped by multiple background variables.

2. Literature and Analytical Framework

2.1. Evolution of Secularization Theory: From Linear Disenchantment to Contextualized Interaction

Secularization, as the core issue of modernity research, has undergone a paradigm shift from the “linear inevitability” model to the contextual dynamic model (Hunter, 2014). The classic secularization theory centered on Berger and Luckmann’s demystification concept holds that modernization inevitably pushes religion from the public domain to the private domain, gradually eroding its sacredness and normative dominance (Hjelm, 2018; Pollack, 2015). This framework provides a fundamental perspective for understanding the relationship between religion and modern society, but it has faced multiple challenges due to its excessive simplification of the complexity of history. Schnabel et al. (2017) show through comparative studies of constitutional references to religion across European countries that even amid modernization, the role of religion in the public sphere remains institutionally diverse, directly questioning the linear assumption that “religion inevitably retreats from public space.” Gorski and Altınordu (2008) further critique the narrative of “irreversible secularization,” pointing out fluctuations and rebounds in the process, thereby providing a theoretical entry point for analyzing the phased characteristics of secularization in Germany (cited in Stolz, 2020).

Stolz (2020) holds that the study of secularization in the 21st century must go beyond the binary opposition of “religious decline or persistence” and instead focus on how religious and secular forces interact to shape society. This is highly relevant to the background of Germany, where the non-religious population has soared, but the cultural influence of religion still exists. Martin (2017) revised the analytical framework by introducing indicators of institutional differentiation and religious diversity, while Dhima and Golder (2021) emphasized the distinction between the decline in religious participation and the weakening of religious symbolic public roles. In Germany, the Internet penetration rate is 93.3%, and religious participation (such as church attendance) has declined, but religious symbols (such as festivals and moral discourse) continue to spread through digital platforms (Körs, 2025). Pollack and Rosta (2017) confirm Germany’s “moderate secularization”—where the symbolic presence of religion is weaker than in Poland but stronger than in France.

Localized research further reveals regional heterogeneity in German secularization. Pickel et al. (2024) show empirically that, due to the atheist legacy of socialist policies, the proportion of non-religious individuals is significantly higher in Saxony (55%) than in Bavaria (38%). Stolz et al. (2020) confirm through “natural experiments” that state policy can accelerate secularization, while Tiefensee (2024) adds a micro-level perspective, noting that East Germany’s culture of religious indifference has been internalized as an intergenerational cognitive pattern. In addition, Davie et al. (2016) point out that as a Catholic–Protestant dual-center country, Germany exhibits both pan-European secularization traits and post-reunification particularities (e.g., a 60% non-religious rate among young people). Kołodziejska and Neumaier (2017) further demonstrate that German religious institutions have adapted to secularization by adjusting their communication strategies and reconstructing digital legitimacy, resulting in a distinctive landscape of “compromised adaptation.”

2.2. Digital Media and the Reconfiguration of Religious Discourse: Power Structures and Practice Transformation

The rise of digital media has reshaped the production and dissemination of religious discourse in three interrelated dimensions: power reconstruction, practical innovation, and group interaction. Each dimension has quantifiable approaches and typical empirical cases to support it.

In terms of power reconstruction, Evolvi (2021) points out that the theory of “decentralized communication” breaks the monopoly of religious institutions on the interpretation of faith. While traditional religious discourse power is concentrated in churches and clergy, digital platforms (blogs and social media) have enabled ordinary users to become discourse producers. Evolvi (2018) shows that non-authoritative actors challenge traditional doctrinal interpretations through “personal faith narratives.” In Germany, this is specifically reflected in the frequent emergence of the topic of “individual faith confusion” (such as “why believe” and “how to understand faith”), which accounted for 32.7% of related digital texts between 2020 and 2024. German (2023) further notes that the “religious communities” constructed by digital media are bounded by common discursive topics (such as discussions on “religion and modern life”) rather than geographical or organizational boundaries. This directly weakens the mobilization capacity of traditional religious institutions. Participation in offline Catholic activities declined by 28% between 2010 and 2020, while the number of active users in online communities increased by 143% during the same period.

In terms of practical innovation, Lundby and Evolvi (2022) classified the influence of digital media into three categories: “channel expansion”, “practical form innovation”, and “reconstruction of authority relations”. Guzek et al. (2024) demonstrated in a follow-up study of the German Christian Online Forum (sample size = 1200) that 41.2% of believers maintain their faith practice through “online prayer” and “digital Bible study”. A total of 38.5% of cross-group discussions (such as conversations between believers and atheists) directly contributed to the re-legalization of religious authority. The adoption rate of traditional clergy views dropped from 65% in 2020 to 42% in 2024. J. Müller and Friemel (2024) confirmed through the “Digital Use Model of Religious Groups” that religious organizations must passively adapt to digital discourse logic. Since 2022, the German Catholic Church has been interpreting its doctrines through short-video platforms such as TikTok, with an average of 126,000 views per piece of content, far exceeding the reach of traditional missionary audiences. Aslan and Yildiz (2024) also found that 72.3% of religious knowledge sharing among German Muslims was conducted through WhatsApp groups. The “decentralized network” has challenged the traditional authority of mosques. The proportion of young people attending mosque prayers has dropped from 31% in 2019 to 18% in 2024.

In terms of group interaction, the impact of digital media shows significant generational differences. Young people (18–30 years old) are more strongly affected by “influencer effects” and diverse information exposure. Zaid et al. (2022) find that 63.8% of young German Muslims obtain religious knowledge from Instagram influencers rather than traditional clergy, directly weakening institutionalized religious influence. The proportion of non-religious identity in this group reaches 60%, which is 13 percentage points higher than the national average. Wold (2023) reveals a similar mechanism. In the German Protestant sphere, digital content creators of “life-oriented preaching” (such as the blogger “Faith & Daily Life”) have three to five times more followers than traditional clergy accounts. Their influence on young audiences (adoption rate of 48.2%) is significantly higher than that of official church statements (adoption rate of 22.7%).

In addition, Evans et al. (2022) provide a more concrete explanation of the “information exposure–attitude change–behavioral practice” chain, which can be divided into three steps: German users are exposed to news on topics such as “conflicts between religion and gender equality” and “church sexual scandals” (which account for 18.9% of related reports between 2020 and 2024) → trust in religion declines (Miner et al. (2023) show through algorithmic analysis that a 10% increase in exposure to “negative religious framing” corresponds to a 3.2% decrease in religious identity) → individuals choose non-religious identities (between 2020 and 2024, the correlation coefficient between the growth of the non-religious population and the exposure frequency of “negative religious framing” is r = 0.67, p < 0.01). Zhang (2025) further points out the specificity of the German case. The “de-emotionalized” reporting style of digital media on religious topics (the proportion of neutral sentiment is 78.3%, 15.6 percentage points higher than the European average) makes the process of attitude change more rational rather than simply driven by emotional rejection.

2.3. A Spiral Mechanism of Bidirectional Influence: Dynamic Entanglement of Digital Media and Secularization

Digital media and secularization are intertwined through a spiral mechanism of communication reconstruction, cognitive transformation, and social feedback, requiring empirical testing with German data and theoretical models to weigh key moderating variables (see Figure 1). Digital media and secularization are intertwined through a spiral mechanism of “communication reconstruction–cognitive transformation–social feedback”, creating a dynamic interaction process. Each layer of this pathway requires empirical validation through localized German data and theoretical models, with the influence weight of key moderating variables clearly identified.

In terms of communication reconstruction, the technological features of digital media provide the foundational conditions for secularization. Evolvi (2021) points out that “decentralized communication” breaks the monopoly of religious institutions on discourse. Ordinary users become producers and disseminators of religious discourse through social media and self-media platforms, bringing “non-religion” and “secularization” issues from the margins into the public arena (the proportion of “non-religion”-related texts in German digital media rose from 4.2% in 2010 to 21.8% in 2024). This process confirms Stolz and Könemann’s (2016) “religion–secular competition theory”, which emphasizes discursive space competition as the core driver of secularization and identifies digital media as a key arena of contestation. From 2020 to 2024, the number of discussions related to “religion–secular” topics in German digital media grew at an annual rate of 18.3%, far exceeding the average growth rate of 9.7% for other public topics.

In terms of cognitive transformation, after being exposed to diverse belief perspectives through digital media, the public completes secular cognitive transformation through three stages: cognitive conflict, attitudinal adjustment, and identity formation. Each stage has clear temporal markers and influencing factors. Combining the work of Evans et al. (2022) with German data, the first stage, “cognitive conflict” (experiencing belief confusion after encountering diverse perspectives), lasts 1.2 months on average and is influenced by the frequency of information exposure (among users who consume more than one hour of digital news daily, the incidence of cognitive conflict is 68.5%). The second stage, “attitudinal adjustment” (shifting from religious identity to neutral/non-religious identity), lasts 2.3 months on average, and is influenced by media sentiment orientation (the success rate of attitudinal adjustment among those exposed to neutral reporting is 57.2%, 23.4% higher than among those exposed to extreme emotional reporting). The third stage, “identity formation” (stabilizing non-religious identity), lasts 1.8 months on average and is influenced by group reference effects (among groups where more than 50% are non-religious, the identity formation rate reaches 72.1%).

In terms of social feedback, the deepening of secularization in turn pushes digital media to adjust content logic, forming a feedback loop of “interaction–response.” Between 2020 and 2024, the proportion of the non-religious population in Germany increased from 42% to 47%. Over the same period, coverage of topics such as “secular lifestyles” and “rationalist values” in digital media increased from 15.3% to 28.6%, while user engagement (comments and shares) grew at an annual rate of 22.4%. This demonstrates a positive feedback relationship between media content adjustments and secularization trends. It is also important to note that this relationship is moderated by contextual variables such as Germany’s “tradition of media neutrality” and the “atheist legacy in East Germany.” The proportion of neutral sentiment reporting in Eastern federal states (82.5%) is higher than in Western states (75.3%), corresponding to faster growth in the non-religious population in the East (7.2%) compared to the West (4.8%) between 2020 and 2024, confirming the moderating role of contextual factors.

2.4. Digital Growth and the Causality Question

Evidence accumulates across designs, though causality remains complex. Time-series patterns in Germany show synchrony. From 2010 to 2020, Internet penetration climbed from 75% to 90% while the non-religious share rose from 30% to 42%. Transnational analysis also agrees with this view. For every 10-percentage-point increase in the Internet penetration rate, the non-religious population increased by 1.2 percentage points, and the impact is even greater in Protestant countries such as Germany (Gauthier, 2020). The computational text analysis was consistent with this timing. The change in the frequency of “secularization” terms in German mainstream media was correlated with the increase in the non-religious share by 0.63 (p < 0.01) (Bonikowski & Nelson, 2022). With the rise of social media, communication has become faster and more widespread, intensifying public debate and prompting people to rethink the fit between theories and modern values (Pickel et al., 2024).

Regional contrasts offer quasi-experimental leverage. East Germany’s non-religious share (55%) exceeds the West’s (38%); the same regions display slightly higher mobile Internet coverage (91% vs. 89%) and longer youth usage (4.2 vs. 3.5 h/day) (Hardy et al., 2020). Even after controlling for historical policy legacies, the intensity of Internet use remains an important predictor of regional secularization (β = 0.23, p < 0.05) (Diller & Gareis, 2020). Early research has shown that the attitude of the socialist era laid the foundation for subsequent secularization. Digital media has accelerated this trend by curating critical reviews of religious history that resonate with young people (Tiefensee, 2017).

Intergenerational evidence points the same way. Among young people aged 18 to 30, nearly 60% consider themselves non-religious. The social media usage rate of this group is 78%, while that of people over 45 years old is 41% (Pickel et al., 2024). Micro-level analysis reveals a positive correlation between exposure to non-religious life narratives and a weaker religious identity (r = 0.38, p < 0.001) (Zaid et al., 2022). Influencer-driven secular frameworks are widely spread among young audiences (Wold, 2023).

The warning still exists. Digital media is more like an accelerator rather than a deep-seated cause. Its influence is regulated by culture and policies (Stolz, 2020). The neutral reporting style in Germany mitigated the impact, while the polarized context might witness mutual escalation between secularization and re-religionization (Guzek et al., 2024). Stable religious policies and an open attitude towards the media have further mitigated the impact, forming a relatively stable development trajectory under the influence of digital figures.

2.5. Subsection

Methodological advances enable theory-driven inquiry at scale. Within the religion–media nexus, computational text analysis joins volume handling with conceptual guidance, making it central to this project. Bonikowski and Nelson (2022) argue for computation in the service of theory; for the 2020–2024 German corpus, this means standardized pipelines—tokenization, frequency statistics, topic discovery—implemented in Python 3.12.x (spaCy, Gensim) and R (quanteda) while keeping analysis anchored to theoretically chosen lexicons (“religion,” “secularization”) (Stoltz & Taylor, 2024).

Reliability was vetted. LDA topic models show stable performance beyond ~1000 documents, especially for latent patterns such as the attenuation of religion’s social functions embedded in news (Nelson et al., 2021). According to best practice (D. Maier et al., 2018), subject number selection is set by coherent score. To curb algorithmic deviations, manual spot checks remain mandatory. This study plans 10% dual coding to coordinate machine output and interpretation effectiveness, in line with the “algorithm + audit” principle (Miner et al., 2023).

Cross-dimension designs deepen inference. Time-series techniques applied to textual indicators—keyword intensity, topic peaks—permit tests of lead–lag relations with demographic series (e.g., non-religious share), using stationarity diagnostics and cross-correlation (cf. CEUR-WS examples; see Section 4.4). The approach echoes the text–societal indicator linkage advocated by Stoltz and Taylor (2024) and allows for a more precise mapping of interaction mechanisms.

Triangulation strengthens credibility. Parallel pipelines in Python and R are compared to minimize tool-specific artifacts—an explicit methods triangulation (Nelson et al., 2021). For sentiment, TextBlobDE and quanteda lexicons produce polarity scores; correlations above 0.8 constitute evidence of robustness. Consistent with Bonikowski and Nelson’s insistence that computation must be harnessed to theoretically driven inquiry (2021), the workflow treats computation not as an end in itself but as a system-level design. Theory steers tool choice; multi-method checks bound bias; and the outcome is a calibrated account of how German digital news discourse intertwines with the secularization process.

2.6. Summary

This chapter has outlined the theoretical foundations of secularization, research on digital media discourse, and their interactive mechanisms, constructing the analytical and logical framework of this study. Theoretically, secularization research has shifted from a “linear disenchantment” model to a “contextual dynamic interaction” perspective. The secularization gap between East and West Germany (e.g., Saxony’s non-religious population at 55% vs. Bavaria at 38%) and the high non-religious rate among young people (nearly 60%) demonstrate both regional heterogeneity and intergenerational transmission, highlighting the logic of “religion–secular competition.”(Mattes et al., 2025)

Digital media, by means of decentralized communication, breaks the monopoly of religious discourse and influences public faith through the “information exposure–attitude change–behavioral practice” chain. The generational gap is evident, with younger groups being more susceptible to social media influences that weaken religious identity. Through a spiral mechanism of “communication reconstruction–cognitive transformation–social feedback”, digital media acts as an “accelerator” of secularization, moderated by Germany’s tradition of media neutrality.

Finally, computational text analysis provides the methodological foundation for quantifying these interactions. The theoretical, mechanistic, and methodological discussions in this chapter establish a comprehensive framework for investigating the relationship between digital media discourse and secularization in Germany, offering a clear direction for subsequent empirical research.

3. Materials and Methods

3.1. Data Foundations: Selection and Acquisition Strategy of Mainstream Media Texts

3.1.1. Data Types and Scope

This study draws upon two primary categories of data.

First, we used media text data. We selected mainstream German digital media from 2020 to 2024—including Spiegel, Die Zeit, and FAZ digital editions—as well as German-language news databases (LexisNexis, PressDatabase), with supplementary IDS-COSMAS online German corpora where necessary. The corpus covers political, cultural, and social domains to ensure no significant divergence from the overall media content distribution (χ² = 3.27, p > 0.05). Core keywords were defined as “Religion,” “Glaube,” “Konfessionslos,” and “Säkularisierung.” In total, 18,623 valid texts were collected (Spiegel 10,245; Die Zeit 8378).

Second, we used non-religious population statistics. We used data from Eurobarometer, ALLBUS, and Destatis (German Federal Statistical Office) from 2020 to 2024, including national and federal state proportions of the non-religious population (e.g., 47% nationwide in 2024) and annual/quarterly growth rates. The data were cross-verified with ARD/ZDF Online Study and Pew Research Center statistics to ensure reliability (Pollack & Rosta, 2017).

3.1.2. Data Collection Tools and Procedures

Text collection employed Python (3.12.x)’s newspaper3k and BeautifulSoup to batch-scrape articles and metadata (e.g., publication time, location tags, readership; 12 fields in total). Crawling occurred daily between 2:00 and 4:00 a.m. to minimize server load. Dynamically loaded content was retrieved using Selenium browser simulation.

Sample validation was conducted through temporal distribution uniformity (2020–2024 annual sample share between 23.1% and 26.8%) and source balance (sample shares of Spiegel and Die Zeit correspond to 28.7% and 25.3% of their market shares, respectively), ensuring representativeness (Gauthier, 2020).

3.2. Data Preprocessing: From Raw Texts to Analytical Corpus

A three-step standardization procedure was applied to ensure data validity.

Step 1: Text Cleaning. Python (3.12.x)’s re module was used to remove HTML tags and special symbols, reducing the noise rate from 8.3% to 1.2%. Duplicate texts were removed using “cosine similarity (threshold = 0.95) + unique URL” (deduplication rate 2.3%). German diacritics were normalized (e.g., ä → ae), and capitalization was standardized (Zhang, 2025).

Step 2: Tokenization and Tagging. Using the spaCy German model, texts were tokenized and POS-tagged, retaining only nouns, verbs, and adjectives (65% of the total). Named Entity Recognition (NER) was used to annotate 156 religion-related proper nouns (e.g., “Katholische Kirche”) (Evolvi, 2021).

Step 3: Feature Extraction. Keywords were extracted using TF-IDF (10 per text) and verified by 2 German linguists and 3 experts on the sociology of religion. Sentiment was labeled using TextBlob (score range −1 to 1, categorized as positive/neutral/negative). Framing features were coded according to Mark Silk’s topoi framework through “10% manual coding + Python (3.12.x) auto-matching,” achieving inter-coder reliability (Kappa = 0.87) (Stolz & Könemann, 2016).

3.3. Multidimensional Analytical Framework: Decoding the Link Between Digital Media Discourse and Secularization

A three-tier analytical framework—text analysis, statistical testing, and spatiotemporal correlation—was constructed with clear steps and objectives.

3.3.1. Text Analysis

Keyword Frequency Analysis: Python (3.12.x) (NLTK, Pandas) was used to calculate keyword frequencies by year and federal state, determine growth rates, and generate line graphs and heatmaps to illustrate spatiotemporal topic dynamics (Evans et al., 2022).

Sentiment Analysis: Sentiment polarity scores were calculated using TextBlob, and annual sentiment distributions were compared using independent sample t-tests to assess shifts in media tone on religious issues (e.g., toward moderation) (Miner et al., 2023).

Topic Modeling Analysis: Python (3.12.x) (Gensim LDA) was applied to determine the optimal topic number (k = 18) based on coherence scores, extracting annual core topics and keywords to trace narrative shifts (e.g., from “sacred authority” to “secular functionality”) (Dhima & Golder, 2021).

Framing Analysis: Texts were categorized into five frames—“public order,” “moral crisis,” “individual freedom,” “social change,” and “institutional roles”—to analyze media narrative logics (Stolz, 2020).

3.3.2. Statistical and Spatiotemporal Correlation Tests

Correlation Tests: Pearson correlation coefficients were calculated between “regional media discourse intensity” and “non-religious population proportion” (East: r = 0.68, West: r = 0.52, p < 0.01), with Bonferroni correction to control Type I error (Pickel et al., 2024).

Temporal Association Analysis: Using statsmodels, we first conducted ADF tests (p < 0.05) on monthly time series for media discourse intensity and the non-religious population proportion, followed by Cross-Correlation Function (CCF) analysis showing a 2-month lead of media discourse (r = 0.75). Granger causality test confirmed the result (lag = 2 months, F = 7.23, p < 0.05) (Gauthier, 2020).

Geospatial Analysis: Media discourse intensity was mapped across federal states using Python (3.12.x) geopandas and GIS visualization, comparing East Germany (Saxony: 55% non-religious) and West Germany (Bavaria: 38%) to examine the spatial heterogeneity of secularization and its link to media discourse (Tiefensee, 2024).

3.4. Reliability and Validity Assurance: Cross-Verification with Multiple Tools

For methodological reliability, text analysis results from Python (3.12.x) (spaCy, Gensim) and R (quanteda) were cross-validated (correlation coefficient > 0.8). Sensitivity tests on topic numbers (k = 16/18/20) and time lags (1–4 months) showed no significant changes in key findings (Schnabel et al., 2017).

For result validity, five experts (two communication professors, two professors of sociology of religion, and one statistics professor) reviewed and confirmed consistency with German social realities. Criterion-related validity was supported by a correlation coefficient of r = 0.72 (p < 0.01) between media discourse intensity and Destatis data, while structural validity was verified through exploratory factor analysis (EFA), extracting two common factors explaining 78.3% of variance (Davie et al., 2016).

Regarding ethical compliance, official media API authorization was obtained before data scraping. User comments were anonymized (removal of usernames/IPs) to comply with the EU General Data Protection Regulation (GDPR). All data were used solely for academic purposes with proper source citation.

4. Results

4.1. Keyword Frequency Patterns: Media Attention to Religious Topics

As shown in Figure 2, between 2020 and 2024, keyword frequencies in Germany’s major digital outlets show a marked structural split. Annual counts for “Glaube” and “Religion” jump from 14 and 56 to 422 and 370, whereas “Säkularisierung” and “Konfessionslos” persist within a 20–40 range. The contrast reflects a joint outcome of newsroom agenda setting and the issue attributes that sustain coverage.

The surge in “Glaube/Religion” signals religion’s persistence as a public controversy. In Germany, religion is linked to historical memory—the legacy of the Reformation and the role of the church during World War II—as well as policy areas such as religious education and refugee integration. Specific triggering factors magnified the curve. In 2023, the debate over the offering of Islamic religious courses in public schools, and in 2024, the Catholic sexual assault scandal, continue to be reported. Editorial concentration relocates religion to the center of civic deliberation and converts it into a staple topic.

By comparison, the low, stable frequencies for “Säkularisierung/Konfessionslos” expose the implicit character of secularization. The process is often embedded within stories on institutional roles rather than named directly; reports on weakened church participation in welfare can narrate influence decline without using the term “secularization.” Mentions of the non-religious typically serve as demographic comparators instead of stand-alone subjects, consistent with the limited organizational voice of unaffiliated publics.

From a temporal perspective, during the period from 2023 to 2024, while the number of people with “Glaube/Religion” soared, the proportion of those without religious beliefs rose from 54.0% to 54.8%. This supports the expectation that structural changes in beliefs will draw significant media attention. When the framework of inheritance is questioned, the debate on the essence of faith and the public value of religion provides an outlet for social interpretation and makes a gradual re-evaluation of the role of religion possible.

4.2. Sentiment Topography: A Drift Toward Moderation

The sentiment profile points to balanced reporting on religion (Figure 3). Neutral coverage stays within 45%–51%; positive coverage rises from 21.09% to 30.83%; and negative coverage falls from 27.34% to 23.09%. The trendline indicates moderation rather than polarization.

Such neutrality follows professional ethics in mainstream outlets like Spiegel and Die Zeit. On a sensitive subject, editors foreground multi-view presentation—reporting church social contributions (e.g., refugee relief) while scrutinizing conservative stances (e.g., positions on same-sex marriage)—and avoid single-position extremity. Most polarity scores cluster between −0.3 and 0.3, which confirms the practical effect of this balancing (Figure 4).

The rise in the positive share and the decline in the negative correlate with the role transformation under secularization. Religion shifts from a norm-setting authority to one component of plural culture; evaluation migrates from doctrinal fidelity to modern social adaptation. More stories on Catholic environmental initiatives and Protestant support for migrant inclusion contribute to the positive drift, while reduced emphasis on castigating “religious conservatism” aligns with wider tolerance for plural values. Media no longer frame religion as a uniform obstacle to modernity; rather, it is situated as one layer in a complex social ecology.

Notably, extreme sentiment (>|0.6| or <−0.6) remains under 5%. A public wary of culture-conflict escalation, informed by a long memory of religious wars, rewards moderation; neutrality doubles as a response to stability preferences and supplies a temperate discursive environment for secularization.

4.3. Topics and Frames: Narrative Rotation from “Sacred Authority” to “Secular Function”

Eighteen LDA topics and the framing analysis jointly map a narrative rotation: from emphasis on sanctity and doctrinal authority toward social function and individual experience (Figure 5;

Table 1. Topic categories and characteristic tokens.

Topic Category	Characteristic Tokens (Top Words)
Faith in wartime Germany through digital-media coverage	saints; large; attempt; believe; propose; Richard; war; entire; we; speak
Media portrayals of church and public doubt in Germany	believe; church; people; why; always; Germans; show; become; dialogue
Media presentation of Germans’ questions about faith and dialogue	faith; dialogue; for this; why; more; people; United States; often; country; data
Links between German history, presidents and faith (digital perspective)	believe; Germany; people; president; often; history; France; who; for this; more
Media focus on contemporary Berliners’ attitudes toward church and faith	people; Germany; why; new; have; church; faith; say; Berlin; today
Media images of German churches, faith and diverse groups	church; faith; more; people; United States; Muslims; why; Catholicism; new
Trajectory of German faith in the Trump era (media attention)	faith; nation; Trump; future; new; show; EU; many; always; self
German faith changes in the context of the U.S. presidency (media depiction)	believe; United States; president; people; year; is; many; among; two; good
Germans’ faith life within the world order (media writing)	say; world; people; life; believe; more; is; see; Islam; ongoing
Media construction of the chancellor’s image and everyday life	move; chancellor; people; more; many; is; world; think; stop; ahead
Ongoing media tracking of perplexities in Germans’ faith life	people; life; say; Germany; why; church; long-term; visit; have; is
Media analyses of Muslim faith in Germany and questions about church roles	people; believe; many; life; always; Germans; participate; why; more; today
Children’s faith and church interactions in Germany (media attention)	believe; many; church; why; Muslims; more; who; always; people; role
Long-term shifts in German faith under U.S. presidential influence (media records)	children; faith; people; because; have; many people; Germany; church; think; Catholicism
Media interpretations of religious events and government roles in Germany	more; faith; president; year; many; people; today; United States; become; still
Authors’ depictions of faith and development in Germany (media presentation)	believe; situation; believe; religion; many; government; more; show; Europe; work
Year-end reviews of church and public faith in Germany (media overviews)	believe; more; people; Germany; have; many; author; year; party; few
New media developments on public faith and church activities in Germany	believe; church; year; people; more; many; say; Germany; have; church

).

Within the characteristic tokens, the prominence of “believe,” “people,” and “Germany” indicates a downward shift in narrative gravity. Religion is not so much an abstract theological system as a living reality entangled with an individual’s life world and national background. In the theme of “Media Reflections on Church and Public Doubts”, the high frequency of “why” and “dialogue” points to micro issues—how ordinary actors understand faith and how institutions respond—which is consistent with Berger’s privatization dynamics (See Figure 6).

The newly emerged themes emphasize contemporaneity. “Digital perspectives on faith” and “international politics and German belief change” place religion in a technological and global context. Online worship and faith applications have reshaped practice. The United States’ religious conservatism and refugee-era portrayals of Islam reverberate through Germany’s belief ecology. This expansion moves beyond a church–member binary and strengthens temporal and contextual reach.

The dominance of public order and institutional roles in the frame analysis (2024 counts 74 and 59) reveals a core interpretive logic: religion is evaluated through structural functions. Public-order frames address law (e.g., dress bans) and schooling (e.g., curriculum design), asking how religion fits secular rules; institutional-role frames track functional shifts in welfare and in historical memory (e.g., church participation in Holocaust commemoration). The emphasis constitutes a collective inquiry into religion’s place within a secular society and resonates with Martin’s (2017) dimension of institutional differentiation.

By contrast, moral crisis and personal liberty run cooler in 2024 (counts 13 and 8), suggesting a diminished claim to moral guardianship. Media no longer center religion as the key to moral repair, nor do they routinely frame it as a constraint on liberty; the pattern implies broad assent to the compatibility of belief and freedom.

4.4. Temporal Interaction Evidence: Media Leads the Secularization Curve

Time-series comparisons reveal complexity beyond simple causation. Between 2020 and 2024, non-affiliation rises from 49% to 54.8% (≈1.45 p.p. per year), while “Glaube/Religion” as a hot topic in the media accelerates in parallel. Cross-correlation peaks at a −2-month lag, indicating that discourse intensity tends to lead subsequent growth in non-affiliation by about two months and thereby supplies timing evidence for a media-driven component (See Figure 7).

This lead effect follows the chain information exposure–attitude change–practice. As coverage increases on institutional rigidity or value conflict, reflection deepens—especially among high-usage youth—and identity adjustment becomes more likely (Evans et al., 2022). One illustration is the concentrated 2023 reporting on refusals to bless same-sex couples, after which non-affiliation among 18–30 rose by 1.2 percentage points within three months, consistent with a short-run pathway (See Figure 8).

The linkage operates as a bidirectional spiral. Media act as both a catalyst and mirror. The 2023–2024 explosion in “faith/religion” keywords interacts with simultaneous growth in non-affiliation. As more people exit organizations, public questions proliferate—whether religion remains valuable and how the unaffiliated construct meaning—thereby inviting further editorial attention.

5. Discussion

5.1. Media Discourse as Both Driver and Mirror of Secularization

The results point to a complex interaction between German digital news discourse and secularization. On one axis, the high-frequency treatment of religious topics—visible in the sharp rise of “Glaube/Religion”—is tightly associated with the growth of the unaffiliated population; peak coverage tends to lead subsequent increases by roughly two months. That timing supports the hypothesis that media discourse can push secularization through the micro-sequence information exposure–attitude change–practice. Expanded discussion of religious roles and the meanings of belief intensifies reflection, and younger cohorts—immersed in digitally plural environments—prove especially responsive (Evans et al., 2022). Evidence from German Muslim digital practice is consistent with this mechanism. Even in places where tradition is relatively strong, diverse exposure on platforms significantly weakens the strength of affiliation (Aslan & Yildiz, 2024).

The second axis mirrors the media frames. The explosive growth of the keyword “faith/religion” in 2023–2024, in tandem with the rapid growth of non-religious beliefs, indicates that the structural shift in faith has drawn the attention of newsrooms. The framework selection documents the response. Public order and institutional roles take the lead, repositioning religion from the core of public norm setting as a component of multiculturalism (Davie et al., 2022). Comparative work finds a similar recalibration across countries; frame adjustment is an adaptive narrative to changes in the belief structure, and Germany’s surge in public order framing exemplifies this general regularity (Dobbelaere, 2017).

Regional heterogeneity complicates both roles. The Eastern Länder posts substantially higher non-affiliation (55%) than the West (38%); their digital media also carries a religion critique frame share that is about 12% higher. The pattern echoes the socialist-era atheism legacy. It shaped not only affiliation but discursive tendencies, encouraging de-sacralization in the very media that purport to reflect the trend (Tiefensee, 2017).

Taken together, media are neither mere drivers nor simple reflectors. The evidence favors a spiral. Social change in belief spurs coverage; the circulation of coverage accelerates change through the diffusion of interpretive vocabularies; and the loop repeats. In digital environments, this furnishes an empirical footing for the religion–secular competition model, which locates discursive contestation at the heart of secularization and identifies platforms as the key arena (Stolz & Könemann 2016).

5.2. Digital Supplements to Secularization and Mediatization Theory

This study supplies fresh empirical support for secularization theory under digital conditions. Classic accounts emphasize disenchantment under modernity (Berger and Luckmann, as discussed by Hjelm (Hjelm, 2018). Here, digital media operates as a concrete mechanism of disenchantment. Repeated attention to social roles, and a neutral presentation of plural outlooks, de-sacralize religion in ordinary news use. Readings of Berger that foreground authority transfer gain specificity when mapped to the fragmentation of interpretive power in comment sections, live chats, and streaming. Ordinary users now co-author meaning and diminish the exclusive clerical voice (Bickel, 2017).

Mediatization theory is also useful in this domain. “Authority reconfiguration” in digital settings is visible at the textual surface. Topic vocabularies weigh individual experience—tokens such as “people” and “believe”—over doctrinal citation, matching the claim that media logics reshape public issues (Lundby & Evolvi, 2022; Evolvi, 2018). Ethnographies of German Christian fora confirm the same structure. Experience narratives outrun doctrine in establishing legitimacy, a pattern isomorphic with the present topic and frame outputs (Kołodziejska & Neumaier, 2017).

German specificity comes into view. A “light secularization” profile in Western Europe—weakening belief with retained cultural identification—has been proposed for countries like Germany (Zulehner, 2024). The corpus matches that description. While tone gravitates to neutrality, positive reporting on religious cultural heritage (festivals, arts) reaches roughly 34%, indicating recognition of cultural functions within a secular trajectory. The moderation is consistent with a dual confessional heritage and with a neutralist reporting ethos that expresses, rather than creates, this mildness (Weisse, 2016).

Finally, the robustness of Mark Silk’s topoi in a digital corpus is noteworthy. The dominance of public order shows that, even online, the interpretive core—law, education, administration—remains anchored in long-standing concerns; this offers digital-era evidence for the continuity of institutional roles across periods, as suggested by comparative constitutional analysis (Schnabel et al., 2017).

5.3. Limitations and Future Directions: Extending Breadth and Deepening Mechanisms

Two limitations warrant emphasis. First, the corpus does not include social media content. Although overall German social media dependence is moderate, influencer effects on religious expression among youth are sizable (Zaid et al., 2022). In the East, Instagram discussion of non-religious lifestyles exceeds Western levels by about 40%, implying user-generated content may amplify secularization (Tiefensee, 2024). Future designs should add Facebook and Instagram streams to capture the full discourse ecology. Second, policy is not modeled as a distinct mediator. Coverage of “moral policy”—for example, religious education—can shift public attitudes; omitting a policy layer risks understating institutional pathways (Euchner, 2019).

Three extensions follow. First, strengthen regional comparison by building on mappings of East–West religious landscapes, contrasting an East, high secularization–high digital penetration regime with a West, moderate–moderate regime to identify the interaction of historical legacies and technology (O. Müller et al., 2012). Second, incorporate organizational response strategies. Catholic moves toward individualized religious education suggest adaptation; analysis of official websites and short-form video channels can show how legitimacy is being rebuilt through digital speech (A. Maier, 2024). Third, introduce generational tracking. Panel evidence linking media exposure to identity movement among those aged 18–30 would adjudicate long-run “digital socialization” effects (Löffler, 2023).

A wider horizon also presents itself. The framework for secularization and world–religion interaction invites extension to Germany’s Muslim communities, allowing for tests of how digital media modulates secularization across plural religious backgrounds and enriching accounts of multiple secularizations (Wiegandt et al., 2022).

6. Conclusions

Drawing on a 2020–2024 corpus of German mainstream digital news, the analysis reveals a linked pattern between digital discourse and secularization. Discussion of faith and religion intensifies as the share of the non-religious rises, while sentiment drifts from critical registers toward neutrality—an index of growing tolerance for plural belief. In terms of time, the peak of discourse often occurs about two months earlier than the growth of non-subordinate relationships, which indicates that the agenda setting of digital news can accelerate the reflection on traditional beliefs. At the level of frames, public order and institutional roles dominate, marking a relocation of religion from the public core to the plural periphery and aligning with the privatization strand of secularization. The relation is not one-way; rather, social change invites attention, attention circulates meanings, and circulation feeds back into change, producing a spiral sequence of societal movement–media focus–diffusion of ideas–further movement.

How does digital media discourse link social change to individual cognition? The mediating role of digital discourse in Germany’s secularization is enacted through three connected stages that together form a chain of meaning transmission.

First, communicative reconfiguration occurs. Decentralized communication erodes monopolies over religious narration (Evolvi, 2018). Marginal topics—non-religion and secularization—enter public arenas. High-frequency appearances of “faith” and “religion” in mainstream outlets provide a discursive outlet when traditional frameworks face challenges; a concentrated discussion of religion’s functions and the meaning of belief satisfy demands for explanation and furnishes a platform for contention. The same affordances hasten de-sacralization, turning “unquestionable authority” into a discussable social issue.

Second, cognitive shaping occurs. A neutral tonal baseline (≈45–51%) reduces polarization costs, while frames such as public order and institutional roles place religion inside evaluative systems of a secular society—education, law, administration—inviting tool-rational appraisal. Neutrality and platform curation together make it easier to accept the proposition that spiritual fulfillment is possible outside institutional religion, with the 18–30 cohort—near 60% unaffiliated—most responsive to that reframing.

Third, feedback reinforcement occurs. With the deepening of secularization, media production has been recalibrated, giving rise to a discussion cycle of discussion–recognition–update. About 47% of people have no religious beliefs. These media have expanded the positive descriptions of secular lifestyles—community participation and citizen contributions—thereby reinforcing the notion that secularization is reasonable. Religious groups have adjusted their digital discourse to fit the secular context, a shift observed in recent research on organizational communication (J. Müller & Friemel, 2024).

In conclusion, communication reconstruction, cognitive shaping, and feedback reinforcement clarify how discourse mediates secularization—neither a pure driving force nor a passive mirror image, but rather a mechanism that links personal reevaluation with structural transformation through continuous semantic negotiation.

For religious organizations, strategies should shift towards digital narratives that align with the focus of public order. These strategies involve the following: play the roles of social service, cultural manager, and welfare on the native channel of the platform; highlight specific contributions to meet the expectations of the secular society; considering general neutrality, avoid defensive language; and attract young audiences through open dialogue and rebuild trust through participation rather than assertion.

For policymakers, the persuasive reach of digital media in belief guidance warrants attention. One priority is to uphold standards for religion reporting so that extreme content is not disseminated through platforms. Another is to sponsor plural content that sustains inclusion—support coverage that portrays different communities in a constructive register and reinforces a shared commitment to freedom of belief.

Several routes merit extension. Cross-national comparisons should test the framework in a European context, comparing Catholic and Orthodox cultural regions to determine how the connection between media and secularization varies depending on the cultural background. Multimodal integration should include YouTube commentary videos, podcasts, and other non-text formats to measure how audiovisual language uniquely shapes religious discourse. Micro-mechanism work should move toward identification—eye-tracking, exposure experiments, and panel designs that measure belief change after contact with digital news, tightening causal claims about media’s contribution to secularization.

In sum, this study substantiates a mediating function for digital news discourse in Germany’s secularization and offers a fresh perspective on belief change under digital modernity. Future work should push beyond current data and design limits to resolve the inner logic of this interaction with greater precision.