In media psychology, we are often interested in whether a certain media exposure or use leads to media effects. Moreover, when we find effects, it is of great interest how long these effects last. Theoretical analyses and systematizations of effect patterns have increased scholarly attention on how effect patterns appear over time. While the question of the duration of media effects has been studied for a long time (e.g., the immediate and long-term effects of violent TV content), it is not until recently that advances in the methodological, such as experience sampling, and analytical realm, such as continuous time models, allow us to analyze the persistence media effects as unfolding across time.

To assess persistence, we conducted a one-week experience sampling study with six probes per day and average time intervals of 2.5 hours between probes. Drawing on 4,028 probes collected from 105 Dutch internet users, we examine how persistent the effects of different entertainment media (i.e., video, games, music) on affective well-being are. We assess persistence by analyzing time continuously using continuous time structural equation modeling (R-package ctsem). Specifically, we hypothesize that engaging with entertainment media is followed by an increase in positive affect, after which the strength of the effect fades. Based on previous research, we expect that entertainment media effects do not persist beyond 12 hours, that is, are no longer statistically different from zero. We exploratively assess at which time point (between 0 and 12 hours) the effect appears and disappears across individuals. Further, we expect that effects of watching videos and gaming are more persistent than that of listening to music.

Analyzing the onset and disappearance of media effects can help us understand both the practical relevance of media effects and provide an important basis for the theoretical understanding of entertainment processes over time.

In today’s digital landscape, individuals frequently engage in media multitasking – simultaneously using (entertainment) media alongside other media or non-media activities (Beuckels et al., 2021; Wang et al., 2015). While research has focused on the cognitive and affective demands of multitasking (e.g., Baumgartner & Wiradhany, 2022; Xu et al., 2022) or sought emotional gratifications broadly (e.g., Wang & Tchernev, 2012), little is known about how affective states influence the likelihood of multitasking behavior.

Drawing on Mood Management Theory (Zillmann, 1988), we propose that mood regulation, reaching a state of excitatory homeostasis with positive valence, may be an important but overlooked motivation for multitasking. Additionally, many studies have investigated media multitasking at the trait-level (e.g., Chang, 2017; Jeong & Fishbein, 2007), while digital media use is substantially explained by situational factors (Schnauber-Stockmann et al., 2024). Thus, an in-situ lens on multitasking can provide insights into the specific combination of activities in a situation without risking being too blurred by individuals’ recall bias. To address these gaps, we aim to explore how mood, defined by valence and arousal, predicts subsequent media multitasking in daily life (RQ1) and which (mediated) activities individuals multitask with (RQ2).

We conduct a secondary analysis of two intensive longitudinal datasets capturing young adults’ everyday entertainment media use. The first dataset examines music streaming on Spotify, typically a low-demand, background media activity (N = 135, T = 1,453 surveys), while the second focuses on Netflix use (N = 123, T = 331 surveys), a higher-demand media activity. Using multilevel modeling, we analyze the longitudinal within-person effects of mood on subsequent multitasking activity while controlling for between-person differences (Bell et al., 2019). For further information on the datasets, see Table 1 in the OSF: https://osf.io/n2e6a/?view_only=44bf35b8463c452d8612923976c1e6f0.

Parasocial relationships (PSR) are one-sided connections between media users and media personae (Horton & Wohl, 1956; Rubin et al., 1985) that extend beyond individual media use situations and can develop through multiple parasocial interactions (PSI; Schramm & Hartmann, 2008). Similar to interpersonal relationships, PSRs have a processual character: following initiation, they evolve through several stages with distinct cognitive, affective and behavioral manifestations (Tukachinsky & Stever, 2019). Despite some theory and research describing PSR development (for a review, see Walter et al., 2023), most studies on PSRs are cross-sectional. Consequently, PSR’s dynamic development and interactions with other influencing factors over time have received little attention (Liebers & Schramm, 2017).

The present study extends insights from longitudinal survey research (Bond, 2021; Tukachinsky Forster et al., 2023) through the unique advantages of computational methods (i.e., an externally valid reconstruction of the phenomenon over a long period of time; van Atteveldt & Peng, 2018; defining purposive samples; Margolin, 2019) by collecting and analyzing a large-scale data set of YouTube videos (n = 2,265 from 13 family vlog channels) and corresponding user comments (n = 280,427 from 35,062 users).

To investigate how users’ PSRs to vloggers develop over time through multiple PSIs (RQ1) and how this is impacted by the amount of exposure to (RQ2) and behavior of personas (RQ3), we assess the prevalence of nine established indicators for PSIs in comments (e.g., cognitive: evaluation of persona; affective: empathy with persona; Schramm & Hartmann, 2008) and three established predictors for PSIs in videos (verbal and visual address of the audience, self-disclosure of persona; Walter et al., 2023) through LLM-based zero-shot classification (e.g., Kroon et al., 2024). We comprehensively validate our approach and demonstrate the potential of computational methods to augment media psychological research by testing the external validity and dynamic nature of entertainment phenomena like parasocial engagement.

Over the past two decades, film and television have increasingly embraced complex storytelling, characterized by non-linear timelines, disrupted causality, and ambiguous logic. Cult classics [Memento (2001) and Mulholland Drive (2001)] paved the way for mainstream films and series such as Inception (2010), and Westworld (2016–), which challenge traditional narrative structures. Scholars have described these as “multiform” or “puzzle” films, highlighting their potential to demand active viewer engagement (Buckland, 2009; Elsaesser, 2008; Kiss & Willemsen, 2017). Despite growing scholarly attention, empirical research on the subjective experience of engaging with such narratives remains limited. Existing theories of narrative engagement emphasize ease of comprehension as a key mechanism (Bilandzic & Busselle, 2011), yet recent studies suggest that audiences might find meaning and enjoyment even in narratives that challenge comprehension through non-linear structures and ambiguity (Balint et al., 2016).

This study seeks to explore how viewers engage with narrational complexity, where comprehension may require active interpretation and cognitive flexibility. Specifically, we ask: (1) How do audiences navigate and engage with complex narratives? (2) What cognitive, emotional, and metacognitive responses are involved in this engagement? (3) What aspects of this experience are not covered by existing measures of narrative engagement? (4) What factors contribute to enjoyment despite—or because of—narrative complexity?

To investigate these questions, we conducted 20 in-depth, semi-structured interviews with participants selected for their interest in complex narratives. The interview structure followed a three-tiered approach, progressing from open-ended questions that allowed for unprompted responses to more topic-specific questions, ensuring various theoretical aspects of the experience were covered. Thematic analysis (Braun & Clarke, 2006) was used to identify patterns in how viewers experience and process complexity. Findings will inform the development of a theoretical framework for engagement with complex narratives and contribute to the design of a novel measurement tool for narrational complexity engagement.