Frontiers | The power of agency: how interactivity in data storytelling influences emotional engagement and perceived trustworthiness in news

The power of agency: how interactivity in data storytelling influences emotional engagement and perceived trustworthiness in news

• 1. Media Culture and Policy Lab, Faculty of Social Sciences, KU Leuven, Leuven, Belgium

• 2. Mediated Environments, LUCA School of Arts, Brussels, Belgium

Abstract

As digital news continues to evolve, data visualization has become an important instrument for enhancing storytelling and fostering emotional engagement. Interactivity is often assumed to deepen this engagement, yet its effects on emotional response and perceived trustworthiness remain underexplored. This study examines the impact of interactivity in data storytelling using The Uber Game published by the Financial Times. A between-subjects experiment was conducted with 60 participants, randomly assigned to either an interactive game condition or a non-interactive video condition. Data were collected through Likert-scale questionnaires and open-ended responses. Results show that the interactive version significantly increases emotional engagement. It also affects perceived trustworthiness, although more moderately, with some participants expressing skepticism toward specific elements of the experience. These findings suggest that while interactivity can enhance engagement, its influence on trustworthiness is more complex and context-dependent, with implications for the design of interactive news experiences.

1 Introduction

Data visualization is widely used in digital journalism to communicate complex social and economic issues to broad audiences. While visual representations are often assumed to enhance clarity and objectivity, visualization alone does not determine how data is interpreted, emotionally experienced, or trusted by readers. Without narrative structure and contextual guidance, visualizations may obscure meaning or distance audiences from the human implications of the data they present. To address these limitations, journalistic practice has increasingly adopted data storytelling approaches that integrate visualization with narrative structures. Rather than presenting data as isolated visual artifacts, data storytelling guides audiences through complex information by framing insights within coherent storylines. Prior research suggests that such approaches can support comprehension and foster emotional engagement, particularly for non-expert audiences (Kosara and Mackinlay, 2013; Shao et al., 2024). However, empirical findings on the audience effects of data storytelling remain mixed, indicating that its impact depends on how narrative and interactive elements are designed and combined (Boy et al., 2015).

Interactivity is a defining feature of data storytelling (Segel and Heer, 2010), typically positioned as a means of enabling users to explore and make sense of information. Yet research indicates that narrative-guided interactive elements do not necessarily motivate users to engage in deeper exploration as users often refrain from seeking insights beyond those explicitly presented within the narrative, even when the visualization affords opportunities for interaction (Boy et al., 2015). Interactivity may, however, support a different form of engagement by fostering emotional involvement through narrative agency—that is, the ability to make choices that influence how a story unfolds. Such agency can shape how audiences interpret data, the extent to which they emotionally connect with the narrative, and how they judge its credibility.

Understanding the role of agency is particularly important in news media, where emotional involvement must be balanced with credibility. Interactive narratives may deepen engagement by allowing audiences to actively participate in data-driven stories, yet they may also raise skepticism if the experience appears overly constructed or subjective. Despite its practical relevance, there is limited empirical research examining how agency in data stories affects emotional engagement and perceived trust in online journalism.

To address this gap, the present study investigates how interactive decision-making in data stories can foster a sense of agency among audiences and how this agency shapes emotional engagement and perceived trustworthiness in online news contexts. Using The Uber Game, a narrative and data-driven news game published by the Financial Times, the study employs a between-subjects experimental design comparing an interactive game-based format with a non-interactive video representation of the same content. This study makes three contributions. First, it isolates user agency as a mechanism shaping audience responses to interactive data storytelling. Second, it demonstrates that agency differentially influences affective (engagement) and epistemic (trust) responses. Third, it extends empirical research on data storytelling by examining a game-based journalistic format.

2 Background

2.1 Data storytelling in journalistic contexts

In digital journalism, data storytelling (DS) has emerged as a prominent approach for communicating complex societal issues to broad audiences. Rather than presenting data as isolated visual representations, DS integrates information visualization with narrative structures that guide interpretation and contextualize meaning derived from data. This approach reflects a shift from exploratory, analyst-oriented visualizations toward communicative formats designed to support public understanding. Conceptually, DS emphasizes the presentation of insights derived from data analysis rather than the display of data alone. By embedding data within coherent storylines, DS aims to reduce cognitive barriers and help audiences interpret the relevance and implications of quantitative information. Recent research suggests that such narrative-driven approaches can support comprehension and emotional engagement, particularly for non-expert audiences and socially complex topics (Bach et al., 2023; Shao et al., 2024). However, DS should not be understood as a singular or uniform practice. Journalistic data stories vary widely in their narrative structure, visual complexity, and degree of interactivity. Consequently, their effects on audiences depend less on the presence of storytelling per se than on the specific design choices through which narrative, visualization, and interaction are combined.

2.2 Interactivity as a design mechanism

Interactivity is frequently described as a defining characteristic of DS, yet it encompasses a broad range of design features. In journalistic visualizations, interactivity may include filtering, navigation, decision-making, or temporal progression, each affording different forms of user involvement. While interactive features are often assumed to enhance engagement, empirical findings indicate that their effects are contingent on design, context, and user characteristics. Research in information visualization suggests that interactivity can foster engagement by encouraging exploration, orienting attention, and supporting active sense-making (Boy et al., 2015; Jia and Sundar, 2023). Interactive features such as user control, contingent feedback, and real-time updates may increase perceived vividness and involvement, leading to deeper cognitive processing of information. At the same time, excessive or poorly structured interactivity can disrupt narrative flow, increase cognitive load, and undermine comprehension (Sundar et al., 2015; Liem et al., 2020).

While news games have been examined as a form of interactive journalism, much of this research focuses on their narrative, persuasive, or experiential dimensions (e.g., Bogost et al., 2010; Plewe and Fürsich, 2020; Gómez-García and de la Hera Conde-Pumpido, 2023; Sun, 2025). In parallel, research on data storytelling has largely investigated visual and interactive formats that are not game-based, emphasizing narrative structure, visualization techniques, and audience engagement (e.g., Segel and Heer, 2010; Ojo and Heravi, 2017; Huang et al., 2023). Consequently, news games specifically designed to communicate and facilitate interpretation of quantitative data remain comparatively underexplored at the intersection of these two traditions (Lin and Wu, 2020).

2.3 Engagement and empathy

Empirical research in information visualization indicates that engagement is influenced not only by visual appeal but also by users’ sense of purpose, autonomy, and meaningful outcome during interaction (Hung and Parsons, 2017). Integrating a narrative flow helps audiences to follow complex information and situate individual data points within a broader context (McKenna et al., 2017). In journalistic settings, storytelling-oriented visualizations have been shown to elicit higher engagement than purely informational formats (Huang et al., 2023). Empathy represents a specific dimension of engagement that is particularly relevant in journalism. Techniques such as character-driven narratives and anthropographic visualizations can humanize abstract data, enabling audiences to relate statistics to lived experiences (Morais et al., 2021). However, empathic responses are highly context-dependent and vary across users. While some studies report increased emotional involvement, others note neutral or even disengaged reactions, especially when visualizations are perceived as repetitive, manipulative, or overly complex (Liem et al., 2020).

2.4 Trust in journalistic visualizations

Trust in data is critical, especially in contexts such as online news, where misinformation can have significant consequences (Vosoughi et al., 2018). Trust in visualization is a complex concept that many researchers have applied several metrics to evaluate (Elhamdadi et al., 2022; Sacha et al., 2015; Xiong et al., 2019; Zhou et al., 2019). Mayr et al. (2019) highlight the tendency of users to depend on and interpret visualized information based on their trust, emphasizing the significance of reliability and comprehension. Narrative framing plays a crucial role in building trust by guiding users through complex datasets and presenting information within meaningful and relatable contexts. Narratives reduce ambiguity, making data more interpretable (Hullman and Diakopoulos, 2011). A well-structured narrative flow enables users to navigate complex data structures more effectively, leading to higher trust and comprehension (McKenna et al., 2017). Moreover, narratives that explicitly acknowledge uncertainty or potential biases in data interpretation can increase trust by demonstrating honesty and accountability (Hullman and Diakopoulos, 2011).

2.5 User agency as an underexplored mechanism

Across studies on DS, interactivity, engagement, and trust, user agency emerges implicitly yet is rarely examined explicitly. Agency refers to the extent to which users feel that their actions meaningfully influence the progression, pacing, and outcomes. In journalistic contexts, agency may shape how personally involved users feel with a story and how they evaluate the legitimacy of the information presented. Despite its conceptual relevance, agency is often treated as a by-product of interactivity rather than a distinct explanatory mechanism. Existing research frequently compares interactive and non-interactive formats without disentangling how specific forms of user control influence emotional engagement and trust. As a result, it remains unclear whether observed effects are driven by interactivity itself or by the sense of agency it affords.

3 The Uber game as case study

The Uber Game is a DS published by the Financial Times that places users in the role of an Uber driver and requires them to make a series of financial and operational decisions under time and resource constraints. Unlike static or linear data presentations, the game affords users direct control over decision-making and narrative progression, allowing choices to influence both short-term outcomes and cumulative financial results. The uniqueness of The Uber Game is the ability to customize data and story through interactivity and story immersion, which allows our study to address the following research questions:

RQ1: How does interactivity that supports agency in DS influence users’ emotional engagement in online news contexts?

RQ2: How does interactivity that supports agency in DS affect users’ perceptions of trustworthiness in the presented information?

Based on existing literature on engagement, interactivity, and narrative visualization, the study tests the following hypotheses:

H1: DS that provide users with a sense of agency will result in higher levels of emotional engagement than non-interactive visualizations.

H2: DS that offer user agency will lead to higher perceived trustworthiness compared to non-interactive formats.

4 Methodology

The study employed a between-subjects experimental design. Participants were randomly assigned to one of two conditions: an interactive game-based data story (Game condition) or a non-interactive video representation of the same content (Video condition). For this control video condition, a video walkthrough of the same content was created, preserving narrative structure, visual elements, and informational content while removing interactivity. The independent variable was the presence of interactivity affording user agency. The dependent variables were emotional engagement and perceived trustworthiness.

Agency was not directly measured as a self-reported construct but was instead operationalized through the experimental manipulation. Participants in the Game condition were afforded decision-making control over narrative progression, while those in the Video condition were not. This approach treats agency as a design-level variable rather than a psychological state (Dillman Carpentier et al., 2015).

4.1 Participants and procedure

A total of 60 participants participated in the study, with 30 participants assigned to each condition. Participants were recruited online and completed the study remotely. After providing informed consent, participants were exposed to the assigned condition with stimulus and subsequently completed an online survey measuring emotional engagement and perceived trustworthiness. Demographic information, including age, gender, education level, and prior experience with data visualizations, was collected at the end of the survey.

4.2 Measures

Emotional engagement was measured using an adapted version of the VisEngage questionnaire (Hung and Parsons, 2017). The scale consisted of 14 items capturing seven dimensions of engagement: captivation, discovery, enjoyment, empathy, exploration, control, and involvement. Responses were recorded on a seven-point Likert scale ranging from 1 (“strongly disagree”) to 7 (“strongly agree”). Minor wording adaptations were made to align items with the journalistic context of the study.

Perceived trustworthiness was measured using a five-item scale adapted from prior research on trust in visualizations (Pandey et al., 2023). The items assessed perceived credibility, clarity, reliability, familiarity, and transparency. Responses were recorded on the same seven-point Likert scale.

4.3 Data analysis

Quantitative analyses were conducted using SPSS. Internal consistency was assessed using Cronbach’s alpha. Construct validity was examined through exploratory factor analysis. Normality was assessed using the Shapiro–Wilk test. As normality assumptions were violated for several variables, non-parametric tests were applied for group comparisons (Tables 1–6).

Factors	Items	Game version	Video version
Mean	SD	Mean	SD
Captivation	EG_1_ca	5.63	1.402	3.27	1.413
EG_2_ca	5.63	1.299	3.27	1.461
Control	EG_3_co	5.73	1.202	3.50	1.676
EG_4_co	5.63	1.217	3.80	1.789
Discovery	EG_5_di	5.67	1.269	4.23	1.716
EG_6_di	5.93	1.015	3.30	1.725
Exploration	EG_7_ex	5.07	1.413	3.30	1.896
EG_8_ex	5.33	1.398	3.23	1.775
Attention	EG_9_at	5.80	0.997	4.47	1.655
EG_10_at	5.80	1.324	4.43	1.695
Interest	EG_11_in	5.47	1.042	3.67	1.647
EG_12_in	5.97	0.999	4.30	1.643
Empathy	EG_13_em	5.50	1.280	4.13	1.634
EG_14_em	5.00	1.438	3.27	1.530
Familiarity	TR_1_fam	4.77	1.165	4.47	1.655
Clarity	TR_2_cla	5.53	0.819	4.63	1.426
Credibility	TR_3_cre	5.83	1.085	4.57	1.165
Reliability	TR_4_reli	5.13	1.306	4.67	1.539
Confidence	TR_5_con	5.00	1.438	4.00	1.597

Descriptive statistics.

Variable	Reliability statistics
Cronbach’s Alpha	N of Items
Trustworthiness	0.744	5
Engagement	0.951	14

Reliability statistics.

Test	statistic	value
Kaiser-Meyer-Olkin measure of sampling adequacy	0.858
Bartlett’s Test of Sphericity	Approx. Chi-Square	1,011.486
df	171
Sig.	<0.001

KMO and Bartlett’s test.

Item	Component
1	2
EG_10_at	0.822	0.092
EG_4_co	0.820	0.105
EG_7_ex	0.809	0.189
EG_3_co	0.801	0.293
EG_6_di	0.792	0.303
EG_12_in	0.788	0.284
EG_8_ex	0.751	0.199
TR_2_cla	0.748	0.044
EG_1_ca	0.736	0.446
EG_9_at	0.665	0.333
EG_5_di	0.641	0.253
EG_11_in	0.637	0.533
EG_14_em	0.599	0.483
EG_13_em	0.580	0.559
TR_1_fam	0.407	0.058
TR_4_reli	−0.028	0.901
TR_5_con	0.100	0.880
TR_3_cre	0.317	0.810
EG_2_ca	0.482	0.503

Rotated component matrix.

Group	Kolmogorov-Smirnova	Shapiro–Wilk
Statistic	df	Sig.	Statistic	df	Sig.
EG_1_ca	Game	0.270	30	<0.001	0.830	30	<0.001
Video	0.198	30	0.004	0.926	30	0.038
EG_2_ca	Game	0.311	30	<0.001	0.825	30	<0.001
Video	0.206	30	0.002	0.936	30	0.072
EG_3_co	Game	0.221	30	<0.001	0.857	30	<0.001
Video	0.151	30	0.081	0.942	30	0.103
EG_4_co	Game	0.218	30	<0.001	0.858	30	<0.001
Video	0.157	30	0.056	0.919	30	0.025
EG_5_di	Game	0.304	30	<0.001	0.828	30	<0.001
Video	0.239	30	<0.001	0.887	30	0.004
EG_6_di	Game	0.293	30	<0.001	0.827	30	<0.001
Video	0.208	30	0.002	0.881	30	0.003
EG_7_ex	Game	0.212	30	0.001	0.885	30	0.004
Video	0.187	30	0.009	0.894	30	0.006
EG_8_ex	Game	0.283	30	<0.001	0.849	30	<0.001
Video	0.256	30	<0.001	0.870	30	0.002
EG_9_at	Game	0.246	30	<0.001	0.862	30	0.001
Video	0.226	30	<0.001	0.920	30	0.027
EG_10_at	Game	0.227	30	<0.001	0.820	30	<0.001
Video	0.199	30	0.004	0.917	30	0.022
EG_11_in	Game	0.229	30	<0.001	0.870	30	0.002
Video	0.158	30	0.055	0.945	30	0.127
EG_12_in	Game	0.247	30	<0.001	0.839	30	<0.001
Video	0.168	30	0.030	0.934	30	0.064
EG_13_em	Game	0.185	30	0.010	0.886	30	0.004
Video	0.202	30	0.003	0.912	30	0.016
EG_14_em	Game	0.190	30	0.007	0.882	30	0.003
Video	0.136	30	0.165	0.945	30	0.120
TR_1_fam	Game	0.188	30	0.008	0.897	30	0.007
Video	0.226	30	<0.001	0.909	30	0.014
TR_2_cla	Game	0.249	30	<0.001	0.870	30	0.002
Video	0.201	30	0.003	0.885	30	0.004
TR_3_cre	Game	0.225	30	<0.001	0.843	30	<0.001
Video	0.212	30	0.001	0.892	30	0.005
TR_4_reli	Game	0.159	30	0.050	0.930	30	0.049
Video	0.186	30	0.010	0.888	30	0.004
TR_5_con	Game	0.190	30	0.007	0.919	30	0.025
Video	0.168	30	0.031	0.948	30	0.152

Tests of normality (Shapiro–Wilk) for all variables.

Item	Mann–Whitney U	Wilcoxon W	Z	Asymp. sig. (2-tailed)
EG_1_ca	109.000	574.000	−5.105	<0.001
EG_2_ca	113.500	578.500	−5.055	<0.001
EG_3_co	130.500	595.500	−4.794	<0.001
EG_4_co	190.500	655.500	−3.908	<0.001
EG_5_di	218.000	683.000	−3.516	<0.001
EG_6_di	88.000	553.000	−5.456	<0.001
EG_7_ex	217.500	682.500	−3.489	<0.001
EG_8_ex	166.000	631.000	−4.275	<0.001
EG_9_at	233.000	698.000	−3.298	<0.001
EG_10_at	234.000	699.000	−3.261	0.001
EG_11_in	173.500	638.500	−4.160	<0.001
EG_12_in	175.500	640.500	−4.151	<0.001
EG_13_em	234.500	699.500	−3.256	0.001
EG_14_em	192.000	657.000	−3.877	<0.001
TR_1_fam	420.000	885.000	−0.455	0.649
TR_2_cla	297.500	762.500	−2.358	0.018
TR_3_cre	204.500	669.500	−3.731	<0.001
TR_4_reli	383.500	848.500	−1.008	0.314
TR_5_con	287.000	752.000	−2.454	0.014

5 Results

The final sample consisted of 60 participants (30 per condition) (see Table 7). Most participants were between 25 and 34 years old and held at least an undergraduate degree. A majority of participants reported prior interest in or familiarity with data visualizations. Gender distribution was comparable across conditions.

Characteristic	Response	Video version	Game version
Frequency	Percent (%)	Frequency	Percent (%)
Interested in news’ visualizations	Yes	27	90%	27	90%
No	3	10%	3	10%
Age	From 18 to 24 years old	7	23%	6	20%
From 25 to 34 years old	20	67%	21	70%
From 35 to 44 years old	3	10%	3	10%
From 45 to 54 years old	0	0%	0	0%
Above 55 years old	0	0%	0	0%
Gender	Man	6	20%	9	30%
Woman	22	73%	21	70%
Non-binary	2	7%	0	0%
Nationality	Vietnamese	16	53%	19	63%
Indonesia	2	7%	2	7%
Turkish	2	7%	1	3%
Finnish	0	0%	2	7%
Greek	0	0%	1	3%
Australian	0	0%	1	3%
Taiwan	0	0%	1	3%
Ecuadorian	0	0%	1	3%
Belgian	4	13%	1	3%
Chinese	2	7%	1	3%
Philippine	1	3%	0	0%
Chilean	1	3%	0	0%
Italian	1	3%	0	0%
American	1	3%	0	0%
First language	Vietnamese	17	57%	19	63%
Turkish	2	7%	1	3%
Indonesian	2	7%	2	7%
Finnish	0	0%	2	7%
Greek	0	0%	1	3%
Hindi	0	0%	1	3%
English	0	0%	1	3%
Spanish	1	3%	1	3%
Dutch	1	3%	1	3%
Chinese	2	7%	1	3%
Filipino	1	3%	0	0%
Hindi	0	0%	0	0%
French	2	7%	0	0%
Nederlands	1	3%	0	0%
Italian	1	3%	0	0%
Occupation	Student	14	47%	8	27%
Office worker	15	50%	20	67%
Researcher	0	0%	2	7%
Lab assistant	1	3%	0	0%
Highest level of education completed	High school	2	7%	2	7%
Undergraduate (Bachelor’s degree)	18	60%	17	57%
Postgraduate (Master’s/ PhD’s degree)	10	33%	11	37%

Demographic of respondents.

5.1 Quantitative results

Our results show that participants in the Game condition reported higher mean scores for emotional engagement across all seven engagement dimensions, with the largest differences observed for discovery, control, captivation, and exploration (see Table 1). These descriptive patterns were supported by Mann–Whitney U tests, which revealed significant differences across all emotional engagement dimensions (all p ≤ 0.001), consistently favoring the Game condition. Effect sizes were large (r = 0.42 to 0.70), with the strongest effects for discovery (EG_6_di: Z = −5.456, p < 0.001, r = 0.70), captivation (EG_1_ca: Z = −5.105, p < 0.001, r = 0.66), and captivation over time (EG_2_ca: Z = −5.055, p < 0.001, r = 0.65). Mean scores for perceived trustworthiness were also higher in the Game condition across all five dimensions, although differences were smaller than those observed for engagement. Inferential tests showed that credibility was significantly higher in the Game condition, with a large effect (Z = −3.73, p < 0.001, r = 0.48), and clarity showed a smaller but significant difference (Z = −2.36, p = 0.018). No significant differences were found for familiarity or reliability, suggesting that interactivity enhanced some aspects of trustworthiness more than others (Table 8).

Item code	Item text	Dimension	Scale	Source	Adapted?
EG_1_ca	While playing the game, I felt immersed and absorbed in its content.	Captivation	1–7 Likert	Hung and Parsons (2017)	Yes — original item referenced “interactive chart”; adapted to game context
EG_2_ca	While playing the game, time seemed to pass quickly.	Captivation	1–7 Likert	Hung and Parsons (2017)	Yes — wording aligned to game format
EG_3_co	I felt that I could make decisions and interact with the game in a way that suited my strategy.	Control	1–7 Likert	Hung and Parsons (2017)	Yes — extended to include decision-making and strategy framing
EG_4_co	I felt in control of my choices and how I navigated through the game.	Control	1–7 Likert	Hung and Parsons (2017)	Yes — adapted from general control item to game navigation
EG_5_di	While playing this game, I learned new information or insights that I had not known before.	Discovery	1–7 Likert	Hung and Parsons (2017)	Yes — adapted from “interactive chart” to game context
EG_6_di	I figured out how to interact with the game and make decisions as I progressed.	Discovery	1–7 Likert	Hung and Parsons (2017)	Yes — adapted to reflect progressive decision-making in game
EG_7_ex	I actively explored different choices and scenarios to understand their impact in the game.	Exploration	1–7 Likert	Hung and Parsons (2017)	Yes — adapted to game scenario structure
EG_8_ex	I experimented with different decisions to see how they influenced the game’s outcome.	Exploration	1–7 Likert	Hung and Parsons (2017)	Yes — adapted to reflect outcome-driven experimentation
EG_9_at	While playing this game, I focused on specific choices and their potential consequences.	Attention	1–7 Likert	Hung and Parsons (2017)	Yes — adapted from general attention to consequence-focused attention
EG_10_at	I had to pay attention to multiple aspects (e.g., financial status, driving conditions, customer ratings) to make informed decisions.	Attention	1–7 Likert	Hung and Parsons (2017)	Yes — contextualized to Uber Game-specific variables
EG_11_in	The challenges and scenarios presented in this game were engaging and relevant to me.	Interest	1–7 Likert	Hung and Parsons (2017)	Yes — adapted from general interest to challenge-relevance framing
EG_12_in	The interactive elements of this game made the experience more interesting and thought-provoking.	Interest	1–7 Likert	Hung and Parsons (2017)	Yes — adapted to foreground interactivity as a driver of involvement
EG_13_em	I felt that I really empathized/felt for the characters or situations in the game.	Empathy	1–7 Likert	Jennett et al. (2008)	Yes — adapted from “empathized with the game” to include characters and situations
EG_14_em	I felt deeply connected to the game emotionally.	Empathy	1–7 Likert	Jennett et al. (2008)	Yes — adapted to emotional connection framing
TR_1_fam	I am familiar with the topic or data presented in this game.	Familiarity	1–7 Likert	Pandey et al. (2023)	Minimal — “visualization” replaced with “game”
TR_2_cla	The game clearly communicates the difficulties and trade-offs involved in being a gig worker.	Clarity	1–7 Likert	Pandey et al. (2023)	Yes — contextualized to gig economy content of The Uber Game
TR_3_cre	I believe the game presents an accurate and credible representation of real-life gig economy work.	Credibility	1–7 Likert	Pandey et al. (2023)	Yes — adapted to journalistic credibility in game format
TR_4_reli	I would trust the facts and insights presented in this game.	Reliability	1–7 Likert	Pandey et al. (2023)	Yes — adapted from general reliability to fact-based trust in game narrative
TR_5_con	I would feel confident using the information in this game to form an opinion or make decisions.	Confidence	1–7 Likert	Pandey et al. (2023)	Minimal — “visualization” replaced with “game”; phrasing slightly extended

Survey items for emotional engagement and perceived trustworthiness.

5.1.1 Reliability and validity

The emotional engagement scale demonstrated excellent internal consistency (Cronbach’s α = 0.95). The perceived trustworthiness scale showed acceptable internal consistency (Cronbach’s α = 0.74). Exploratory factor analysis supported a two-factor structure corresponding to emotional engagement and perceived trustworthiness. The Kaiser–Meyer–Olkin measure indicated adequate sampling adequacy (KMO = 0.86), and Bartlett’s test of sphericity was significant [χ²(171) = 1011.49, p < 0.001].

5.2 Qualitative results

5.2.1 Empathy and perspective-taking

Participants in both the game and video conditions reported empathic responses toward Uber drivers, frequently referring to emotional strain, financial insecurity, and limited worker protections. Across both conditions, 22 participants mentioned daily challenges faced by drivers, including stress, pressure, and mental exhaustion. In the game condition, eight participants explicitly described the experience as allowing them to “step into the shoes” of Uber drivers. Player 23 stated that the experience “helped me to play the role of a real Uber driver to understand and sympathize with the difficulties of drivers,” while Player 29 referred to poor treatment by app-based platforms and a lack of benefits. Five game participants mentioned drivers’ stress or mental exhaustion, compared to one participant in the video condition (e.g., Player 16; Viewer 10). Financial concerns were also frequently mentioned. Three participants in the game condition and three in the video condition referred to expenses such as fuel and maintenance. Game participants more often mentioned trade-offs between work and personal life, including missed family or social events, as described by Player 30. Video participants expressed sympathy more briefly, for example stating “sympathetic, sometimes angry” (Viewer 10) or “I felt a bit sorry” (Viewer 5).

5.2.2 Emotional responses and engagement

Participants in both conditions reported emotional responses related to financial pressure and work uncertainty. Seven participants in the game condition and six in the video condition used terms such as “stress,” “anxiety,” “frustration,” or “worry.” Game participants more frequently described emotional intensity, using expressions such as “high pressure” (Player 1), “helplessness” (Player 8), and “intense feelings” (Player 20). In contrast, six video participants qualified their emotions using attenuated language, such as “a bit depressed” (Viewer 6), “a bit sad” (Viewer 7), or “a little anxious” (Viewer 29). Beyond discomfort, 10 game participants reported excitement, curiosity, immersion, or motivation to continue the experience. Six participants mentioned curiosity about alternative outcomes or endings. Game 30 noted that “the more I played, the more situations opened up, making me curious,” while Game 19 described “the desire to play again and try to make more rational decisions”. In the video condition, two participants described the content as thought-provoking (Viewer 16; Viewer 25), and none reported interest in further exploration.

5.2.3 Interactivity and decision-making

Participants in the game condition frequently commented on decision-making and interaction. Two participants in the game group did not complete the experience, while 10 participants in the video group reported not watching the video until the end. Both groups included three participants who reported low interest in news visualizations. Among participants who did not complete the assigned format, reported interaction times differed. Player 29 and Player 30 reported playing the game for 15 min and 3 min respectively, while the average reported viewing time in the video condition was 1.44 min. Ten game participants described the experience as personalized and decision-driven. Player 8 stated that “making decisions yourself helps you understand the implications,” while Player 28 reported needing “to think carefully” before each choice. Several participants mentioned replaying or wanting to replay the game (e.g., Player 19). Four game participants stated that the experience helped them remember drivers’ challenges more clearly than watching a video, as noted by Player 8. One video participant explicitly compared the formats, stating that “the video factor probably made the experience less effective than me playing the actual game” (Viewer 26). Seventeen participants across both conditions related the experience to their own lives. Game participants referred to family-related prompts or personal encounters with Uber drivers (e.g., Player 8; Player 18). In the video condition, three participants noted that emotional impact was strongest during the final scenes involving financial calculations.

5.3 Repetitiveness and prior knowledge

Eight participants in the game condition described decision-making as repetitive or predictable, such as “the decisions looked the same and repetitive” (Player 9). In the video condition, repetitiveness led to disengagement, with Viewer 2 reporting watching “another 30 seconds to see if there were any differences (turned out no)”. Participants in both conditions referred to prior knowledge of the gig economy. Some game participants reported gaining limited new insight due to familiarity with the topic. In the video condition, six participants described difficulty extracting meaningful insights or reported neutral emotional responses, including statements such as “no feelings” (Viewer 20) or “neutral, no special feelings” (Viewer 19). Participants in the game condition more frequently used first-person language when describing their experience [e.g., “I felt annoyed” (Player 10); “I was frustrated” (Player 3)], whereas participants in the video condition more often used third-person or generalized phrasing, such as “he tried his best but…” (Viewer 6).

5.4 Participants’ assessment of visualization trustworthiness

Four game participants described the interactive structure and presentation as realistic or credible. Player 2 noted that “the graphics is interactive and quite representing the actual condition,” while Player 6 commented that the format appeared “well-invested.” Eight game participants referred to the simulation as reflecting real-life decision-making processes (e.g., Player 22). Video participants commented less frequently on realism, though two referred to everyday challenges such as helping a child with homework (Viewer 24). One participant questioned the relevance of the content due to its age (Viewer 3). Participants in both conditions commented on data presentation. Game participants frequently referred to structured presentation and trade-off comparisons, with Player 22 describing the information as “segmented and presented in a timely, relevant, and intuitive manner”. Video participants occasionally highlighted the inclusion of statistics, such as “very specific stats” (Viewer 1). Five participants expressed concerns about the transparency of calculations, particularly regarding earnings. Eight participants mentioned the Financial Times as influencing their trust, including five participants in the video condition. Player 8 stated that the association with the Financial Times “gives it some credibility,” while Viewer 6 referred to the article’s journalistic origin. Two participants expressed doubt when sources were not immediately visible (e.g., Player 14). Participants with prior familiarity with Uber or similar reporting more frequently described the information as credible (e.g., Viewer 5; Viewer 9). Participants without prior experience expressed uncertainty about the accuracy of specific information (e.g., Player 27; Viewer 23).

6 Discussion

The results indicate that interactivity enhances emotional engagement (supporting H1), while its effects on perceived trustworthiness are more modest and context-dependent (partially supporting H2). These findings suggest that agency may differentially influence affective outcomes (such as engagement) and epistemic evaluations (such as trust), highlighting the importance of distinguishing these response dimensions in human–media interaction research.

6.1 User agency and emotional engagement

The most consistent finding across both quantitative and qualitative data is the strong relationship between interactivity and emotional engagement. Participants who interacted with the game reported significantly higher levels of captivation, control, discovery, and exploration compared to those who viewed the non-interactive video. Facilitating agency through decision-making prompted participants to consider their perspectives and promoted emotional engagement. Participants experienced how their decisions shaped the narrative outcome, which appeared to intensify emotional responses such as stress, frustration, curiosity, and empathy. Rather than simply learning about the challenges faced by gig workers, participants described experiencing these challenges through decision-making. This supports prior research suggesting that engagement in data visualization is not driven solely by visual appeal, but by users’ sense of autonomy and meaningful action (Hung and Parsons, 2017; Jia and Sundar, 2023). Notably, this reported sense of agency does not necessarily imply genuine control over the message or interpretive space. Interactivity in DS frequently takes a paternalistic form (Anderson and Borges-Rey, 2019; Appelgren, 2018), where curated choices guide users through an authored pathway that enables co-creative sensemaking (Anderson, 2021).

These findings indicate that emotional engagement emerged not merely from narrative content or emotional framing alone, but from interactive participation fostering identification with the narrative subject. While both formats elicited some degree of empathic response, the game encouraged deeper perspective-taking, reflected in participants’ use of first-person language and references to personal responsibility. Still, a minority of the participants reported only neutral or limited emotional engagement. This aligns with previous findings that empathic effects in DS are highly context-dependent and influenced by individual differences, prior knowledge, and interaction preferences (Liem et al., 2020). Agency is therefore better viewed as a facilitator than a universal driver of engagement.

6.2 Trustworthiness through interactivity

The perceived trustworthiness tended to be higher among participants who interacted with the DS, although this effect varied across dimensions. The sense of agency created by letting users decide solutions or write their own story allows them to explore or dig deeper into the story. Hence, users feel more in control and the content seems more transparent, which is in line with the work of Hullman and Diakopoulos (2011). Trust also tended to grow when users felt the story was logically structured and thoughtfully developed, which aligns with findings from McKenna et al. (2017). The qualitative responses further supported this connection. Some participants mentioned that feeling emotionally engaged with the drivers’ experiences made the story and data seem more real and logical. On the other hand, a few expressed doubts about the credibility of the information, especially if the format felt to create emotional connection with users. Beyond the interactivity, trust could be influenced by various factors, including the formats, the perceived credibility of the news source, or how emotionally compelling the experience was. This pattern resonates with Li et al. (2018), who found that interactivity alone did not enhance perceived data credibility, while source attribution and design quality played more decisive roles. Both studies suggest that trust in journalistic data presentations is not primarily driven by the interactive features of the format, but by contextual cues such as perceived source credibility and presentation clarity. In our study, participants frequently cited the Financial Times association as a trust anchor, which aligns with Li et al.’s (2018) finding that peripheral source cues function as cognitive shortcuts when users lack sufficient domain knowledge to evaluate content directly.

In addition, qualitative insights remain ambiguous whether participants’ trust is directed toward the data visualization itself or the narrative message communicated through the experience. Many participants expressed that they perceived trust in general, but their explanations were a mix between the trust toward visual presentation of the data, the realism of the narrative, or the overal experience. Therefore, when interpreting trust in DS in a journalistic context, it is essential to consider that users may be responding not just to how data is represented, but also to the broader message and emotional environment in the experience.

Beyond interactivity, individual differences in prior familiarity likely influenced results. Qualitatively, participants familiar with the gig economy more readily accepted the game’s information as credible, while unfamiliar participants expressed uncertainty, which is consistent with Pandey et al. (2023), who identified topic familiarity as one of the strongest participant-level predictors of perceived trust in visualizations, independent of design features. Familiarity with interactive formats may similarly have moderated engagement responses. The sample’s relative homogeneity limits the extent to which these effects can be fully disentangled here; future research should explicitly measure topic familiarity and visualization literacy as covariates.

Overall, these insights offer a deeper understanding of how interactivity in data storytelling can influence not just engagement, but also how credible the information feels to the audience. This expands on traditional views that trust mainly depends on the authority of the

6.3 Implications for data storytelling in journalism

The findings suggest that agency plays a differentiated role in journalistic DS. From a theoretical perspective, the study contributes to research on DS by reframing agency as a core explanatory mechanism rather than a secondary attribute of interactivity. By demonstrating that agency primarily affects affective engagement while exerting weaker influence on trust, the findings help clarify why prior empirical results on DS have been mixed. Differences in how agency is designed and experienced may account for variability across studies.

For journalistic practice, the results suggest that interactive narrative formats are particularly effective when the goal is to foster emotional connection and audience involvement. However, designers should be cautious in assuming that greater agency automatically leads to higher trust. Transparency in data sources, clarity in calculation logic, and visible journalistic accountability remain essential, especially when interactive formats resemble games or simulations.

6.4 Limitations and future directions

While the findings provide valuable insights, they should be interpreted in light of several limitations. The sample was relatively homogeneous and skewed toward younger, highly educated participants, which may limit generalizability. In addition, emotional engagement and trust were measured using self-reported questionnaires, which may not fully capture dynamic or unconscious responses during interaction. A further limitation is that individual variation in perceived agency within the Game condition could not be accounted for, as agency was not measured at the individual level. The sample size (N = 60) is sufficient to detect large differences between conditions, which is consistent with the strong effects observed for emotional engagement. However, smaller differences in perceived trustworthiness may be harder to detect with this sample, which could explain why some trust dimensions did not reach statistical significance. Some engagement subdimensions (e.g., empathy) should be interpreted cautiously due to scale adaptation.

Future research may adopt different types of interactivities (e.g., gamification vs. scrolling or filtering tools) to determine which design elements are most effective in enhancing comprehension, empathy, and credibility in journalistic storytelling, and could incorporate validated agency scales to examine how perceived agency moderates these effects. Longitudinal designs and real-time measurement techniques may also provide deeper insight into how emotional and cognitive responses evolve throughout interactive news experiences.

7 Conclusion

This study examined how interactivity and user agency shape audience experiences with data stories in online journalism. By comparing an interactive, game-based data story with a non-interactive video of the data story, the findings show that interactive formats are associated with higher levels of emotional engagement, while effects on perceived trustworthiness are more limited and vary across dimensions. Quantitative results indicate that participants in the game condition reported stronger emotional engagement across multiple engagement dimensions, whereas differences in perceived trustworthiness were present only for certain aspects, such as clarity and credibility. Qualitative findings complement these results by showing that participants in the game condition more frequently described decision-making, personal involvement, and emotionally intense responses, while participants in the video condition more often described observational and reflective experiences. From a human–media interaction perspective, the findings highlight user agency as a key interaction characteristic that shapes how audiences experience data-driven narratives. However, agency alone does not uniformly enhance trust, as participants also relied on factors such as data transparency, source credibility, and prior knowledge when evaluating the information presented.

The findings offer practical implications for designers and data journalists by suggesting that interactive features can support affective involvement, but should be carefully balanced with clear data presentation and transparency to support informed trust judgments.

Statements

Data availability statement

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

Ethics statement

The studies involving humans were approved by Sociaal-Maatschappelijke Ethische Commissie (SMEC) – Social and Societal Ethics Committee KU Leuven (approval number M-2025 02 149). The studies were conducted in accordance with the local legislation and institutional requirements. The participants provided their written informed consent to participate in this study. Written informed consent was obtained from the individual(s) for the publication of any potentially identifiable images or data included in this article.

TH: Investigation, Writing – original draft. GI: Supervision, Writing – review & editing. SC: Conceptualization, Supervision, Writing – review & editing.

Funding

The author(s) declared that financial support was received for this work and/or its publication. The authors acknowledge internal research funding provided by KU Leuven under project number C2M/24/021, as well as by LUCA School of Arts under project name Stories-In-Between.

Conflict of interest

The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Generative AI statement

The author(s) declared that Generative AI was used in the creation of this manuscript. The authors used a generative AI tool to assist with language editing and improving clarity of expression. The authors reviewed and take full responsibility for the final content of the manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

References

• AndersonB. (2021). The relational UX: constructing repertoires of audience agency in pioneer journalism practice. Med. Stud. (Med. Stud.)15, 167–189.

• AndersonB.Borges-ReyE. (2019). Encoding the UX: user interface as a site of encounter between data journalists and their constructed audiences. Digit. Journal.7, 1253–1269. doi: 10.1080/21670811.2019.1607520

  • CrossRef
  • Google Scholar

• AppelgrenE. (2018). An illusion of interactivity: the paternalistic side of data journalism. Journal. Pract.12, 308–325. doi: 10.1080/17512786.2017.1299032,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• BachB.FreemanE.Abdul-RahmanA.TurkayC.KhanS.FanY.et al. (2023). Dashboard design patterns. IEEE Trans. Vis. Comput. Graph.29, 1–11. doi: 10.1109/TVCG.2022.3209448,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• BogostI.FerrariS.SchweizerB. (2010). Newsgames: Journalism at Play. Cambridge, MA: Mit Press.

• BoyJ.DetienneF.FeketeJ. D. (2015). Storytelling in Information Visualizations: Does it Engage Users to Explore Data? In: Proceedings of the 33rd Annual ACM Conference on Human Factors in Computing Systems (CHI ‘15), Association for Computing Machinery, pp. 1449–1458.

• BoyJ.RensinkR. A.BertiniE.FeketeJ. (2014). A principled way of assessing visualization literacy. IEEE Trans. Vis. Comput. Graph.20, 1963–1972. doi: 10.1109/tvcg.2014.2346984,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• Dillman CarpentierF. R.RogersR. P.BarnardL. (2015). Eliciting behavior from interactive narratives: isolating the role of agency in connecting with and modeling characters. J. Broadcast. Electron. Media59, 76–93. doi: 10.1080/08838151.2014.998222

  • CrossRef
  • Google Scholar

• ElhamdadiH.GabaA.KimY. S.XiongC. (2022). How do we Measure Trust in Visual Data Communication? 2022 IEEE Evaluation and beyond – Methodological Approaches for Visualization (BELIV), pp. 85–92.

• Financial Times. (2017). The Uber Game. Available online at: https://ig.ft.com/uber-game/ (Accessed March, 2025).

• Gómez-GarcíaS.de la Hera Conde-PumpidoT. (2023). Newsgames: the use of digital games by mass-media outlets to convey journalistic messages. Games Cult.18, 449–474. doi: 10.1177/15554120221105461

  • CrossRef
  • Google Scholar

• HuangX.LiX.YaoJ. (2023). Investigating the role of visual storytelling in enhancing audience engagement in digital journalism. MEDAAD1, 10–17. doi: 10.70470/medaad/2023/002

  • CrossRef
  • Google Scholar

• HullmanJ.DiakopoulosN. (2011). Visualization rhetoric: framing effects in narrative visualization. IEEE Trans. Vis. Comput. Graph.17, 2231–2240. doi: 10.1109/TVCG.2011.255,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• HungY. H.ParsonsP. (2017). Assessing User Engagement in Information Visualization. CHI EA’17: Proceedings of the 2017 CHI Conference Extended Abstracts on Human Factors in Computing Systems, pp. 1708–1717.

• JennettC.CoxA. L.CairnsP.DhopareeS.EppsA.TijsT.et al. (2008). Measuring and defining the experience of immersion in games. Int. J. Hum.-Comput. Stud.66, 641–661. doi: 10.1016/j.ijhcs.2008.04.004

  • CrossRef
  • Google Scholar

• JiaH.SundarS. S. (2023). Vivid and engaging: effects of interactive data visualization on perceptions and attitudes about social issues. Digit. Journal.12, 1205–1229. doi: 10.1080/21670811.2023.2250815

  • CrossRef
  • Google Scholar

• KosaraR.MackinlayJ. (2013). Storytelling: the next step for visualization. Computer46, 44–50. doi: 10.1109/mc.2013.36

  • CrossRef
  • Google Scholar

• LiN.BrossardD.ScheufeleD. A.WilsonP. H.RoseK. M. (2018). Communicating data: interactive infographics, scientific data and credibility. J. Sci. Commun.17:A06. doi: 10.22323/2.17020206

  • CrossRef
  • Google Scholar

• LiemJ.PermC.WoodJ. (2020). Structure and empathy in visual data storytelling: evaluating their influence on attitude. Comput. Graph. Forum39:13980. doi: 10.1111/cgf.13980

  • CrossRef
  • Google Scholar

• LinJ.-H. T.WuD.-Y. (2020). Newsgames for the greater good: the effects of graphic realism and geographic proximity on knowledge acquisition and willingness to help. Journal. Mass Commun. Q.97, 30–51. doi: 10.1177/1077699018820315

  • CrossRef
  • Google Scholar

• MayrE.HynekN.SalisuS.WindhagerF. (2019). Trust in Information Visualization. TrustVis@EuroVis.

• McKennaS.RicheN. H.LeeB.BoyJ.MeyerM. (2017). Visual narrative flow: exploring factors shaping data visualization story reading experiences. Comput. Graph. Forum36, 377–387. doi: 10.1111/cgf.13195

  • CrossRef
  • Google Scholar

• MetzgerM. J.FlanaginA. J. (2013). Credibility and trust of information in online environments: the use of cognitive heuristics. J. Pragmat.59, 210–220. doi: 10.1016/j.pragma.2013.07.012

  • CrossRef
  • Google Scholar

• MoraisL.JansenY.AndradeN.DragicevicP. (2021). Can Anthropographics Promote Prosociality? A Review and Large-Sample Study. In: Proceedings of the 2021 CHI Conference on Human Factors in Computing Systems, pp. 1–18.

• OjoA.HeraviB. (2017). Patterns in award-winning data storytelling. Digit. Journal.6, 693–718. doi: 10.1080/21670811.2017.1403291

  • CrossRef
  • Google Scholar

• PandeyS.McKinleyO.CrouserR.OttleyA. (2023). Do you trust what you see? Toward a multidimensional measure of trust in visualization. arXiv2023:14. doi: 10.1109/VIS54172.2023.00014

  • CrossRef
  • Google Scholar

• PleweC.FürsichE. (2020). Producing Newsgames beyond boundaries: journalists, game developers, and the news business. Convergence26, 486–502. doi: 10.1177/13548565209180

  • CrossRef
  • Google Scholar

• SachaD.SenaratneH.KwonB. C.EllisG.KeimD. A. (2015). The role of uncertainty, awareness, and trust in visual analytics. IEEE Trans. Vis. Comput. Graph.22, 240–249. doi: 10.1109/TVCG.2015.2467591,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• SegelE.HeerJ. (2010). Narrative visualization: telling stories with data. IEEE Trans. Vis. Comput. Graph.16, 1139–1148. doi: 10.1109/TVCG.2010.179,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• ShaoH.Martinez-MaldonadoR.EcheverriaV.YanL.GasevicD. (2024). Data Storytelling in Data Visualization: Does it Enhance the Efficiency and Effectiveness of Information Retrieval and Insights Comprehension? CHI’24: Proceedings of the 2024 CHI Conference on Human Factors in Computing Systems, Association for Computing Machinery, No. 195, pp. 1–21.

• SunH. (2025). The negotiation of truth claims in Newsgames: the tension between “fact” and “fiction”. Journal. Stud.26, 220–239. doi: 10.1080/1461670X.2024.2418039

  • CrossRef
  • Google Scholar

• SundarS. S.JiaH.WaddellT. F.HuangY. (2015). “Toward a Theory of Interactive Media Effects (TIME): Four Models for Explaining how Interface Features Affect User Psychology,” in The Handbook of the Psychology of Communication Technology, ed. SundarS. S. (Hoboken, NJ: Wiley Blackwell), 47–86.

• VosoughiS.RoyD.AralS. (2018). The spread of true and false news online. Science359, 1146–1151. doi: 10.1126/science.aap9559,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• XiongC.PadillaL.GraysonK.FranconeriS. (2019). Examining the Components of Trust in Map-Based Visualizations. 1st EuroVis Workshop on Trustworthy Visualization (TrustVis), pp. 19–23.

• ZhouJ.LiZ.HuH.YuK.ChenF.LiZ.et al. (2019). Effects of Influence on User Trust in Predictive Decision Making. Extended Abstracts of the 2019 CHI Conference on Human Factors in Computing Systems, pp. 1–6.

Summary

data storytelling, data visualization, engagement, interactive visualizations, online news, trust

Hoang TH, Ipers G and Claes S (2026) The power of agency: how interactivity in data storytelling influences emotional engagement and perceived trustworthiness in news. Front. Comput. Sci. 8:1843009. doi: 10.3389/fcomp.2026.1843009

Mariel Garcia-Hernandez, University of Monterrey, Mexico

Paulo Couraceiro, University of Minho, Portugal

Caroline Guedes, Universidade Federal da Bahia Faculdade de Comunicacao, Brazil

This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.