From Narrative Transportation to Algorithmic Cultivation: The Impact of Exposure to Wealth-Flaunting Short Videos on Materialism in Emerging Adulthood and Interventions

Cultivation theory posits that prolonged media exposure gradually shapes individuals' social cognition and value orientations. With short-video platforms becoming the dominant digital medium in China—reaching 1.04 billion users by the end of 2025—wealth-flaunting content showcasing lavish lifestyles has proliferated, raising concerns about its potential cultivation of materialistic values among college students in emerging adulthood. However, existing empirical literature suffers from three key limitations: most studies rely on cross-sectional designs that cannot establish temporal precedence; few have directly compared the differential persuasive power of short videos versus other media formats presenting identical content; and virtually no research has examined whether disrupting algorithmic personalization—a core feature distinguishing short-video platforms from traditional media—can attenuate cultivation effects. Grounded in cultivation theory and its recent extension to algorithmic cultivation, the present research addressed these gaps through four progressive studies examining whether wealth-flaunting short videos possess medium-specific persuasive advantages, whether long-term exposure cultivates both explicit and implicit materialism, and whether disabling personalized recommendations can mitigate such effects.Study 1 employed a between-participants experimental design (N = 124 college students) comparing wealth-flaunting content presented as short videos versus social media image-text posts (Weibo), with content held constant across conditions. Narrative transportation and state materialism were measured post-exposure. Study 2 adopted a four-wave longitudinal design (N = 232, monthly intervals) measuring wealth-flaunting short-video exposure, affluent population estimates, and explicit materialism (Material Values Scale and Forest Management Game) at each wave. Latent growth modeling was used to examine cross-variable growth trajectories. Study 3 (N = 91) used a two-timepoint correlational design: participants first reported their frequency of exposure to wealth-flaunting short videos over the past three months, then completed two implicit materialism measures one week later—a Single-Category Implicit Association Test (SC-IAT) and a collage paradigm in which participants selected items that "make me happy" from 100 options across five categories. Study 4 was a three-week, two-arm randomized controlled trial (N = 110 high-exposure college students) in which the experimental group disabled personalized recommendation on their short-video apps while the control group maintained normal use. Weekly measures of wealth-flaunting content exposure, upward social comparison, state materialism, and actual Douyin usage time were collected across four timepoints (baseline plus three weekly assessments). Generalized linear mixed models and latent growth models were used for analysis.In Study 1, the short-video group reported significantly higher state materialism than the image-text group. Narrative transportation fully mediated this effect. In Study 2, latent growth modeling revealed that the initial level of wealth-flaunting short-video exposure positively predicted the initial levels of both affluent population estimates and explicit materialism. Critically, the rate of change in exposure positively predicted the rate of change in materialistic values, supporting a dose-response pattern. However, the rate of change in affluent population estimates negatively predicted the rate of change in materialism, contrary to the indirect cultivation hypothesis. In Study 3, wealth-flaunting short-video exposure significantly predicted the proportion of material items selected in both rounds of the collage paradigm after controlling for daily short-video use, but did not predict SC-IAT D-scores. In Study 4, GLMM analyses showed significant group × time interactions for both material-flaunting content exposure and actual Douyin usage time, with the experimental group showing significantly steeper declines. Longitudinal mediation analysis using latent growth modeling revealed a significant serial indirect effect: disabling personalized recommendations → reduced wealth-flaunting content exposure → weakened upward social comparison → decreased materialism.The findings demonstrated that wealth-flaunting short videos exerted a robust cultivation effect on college students' materialism. The unique immersive modality of short videos and their algorithmic recommendation mechanisms significantly amplified this negative impact. Importantly, our ecological intervention proves that disabling personalized recommendations effectively breaks the algorithmic feedback loop and mitigates materialism. These findings extend the traditional cultivation theory into the algorithmic media era and provide crucial empirical evidence for policymakers and platform administrators to implement "one-click-off" recommendation features.