From Affective Satisfaction to Habit: Reinforcement Dynamics in Short-Form Drama Engagement

Summary

Short-form drama environments provide dense reinforcement conditions through ultra-short episodes and frequent re-entry. Using an S-O-R model and PLS-SEM, this study examined continuance intention among short-form drama users. The findings indicate that affective satisfaction was positively associated with habit, and habit was positively associated with continuance intention.

Abstract

Short-form drama environments have emerged as structurally distinct digital environments characterized by ultra-short episodes, high-frequency re-entry, and algorithmically reinforced exposure. While prior continuance research has examined satisfaction and habit as parallel predictors of sustained use, little is known about how compressed reinforcement cycles embedded in short-form ecosystems may strengthen the association between affective responses and habitual engagement under conditions of dense reinforcement. Drawing on the stimulus-organism-response (S-O-R) framework, this study proposes a structurally reinforced habit-formation model in which environmental stimuli (access convenience, passing time, and content quality) influence affective satisfaction, which, in turn, is associated with habitual engagement and continuance intention. Using cross-sectional survey data from 260 short-form drama users, analyzed using partial least squares structural equation modeling (PLS-SEM), the findings reveal that affective satisfaction is significantly associated with habit formation, and that habit emerges as a dominant driver of continuance intention. Quantitatively, content quality (β = 0.448, p < 0.001) and passing time (β = 0.385, p < 0.001) were the strongest predictors of affective satisfaction, which strongly predicted habit (β = 0.686, p < 0.001) and continuance intention (β = 0.496, p < 0.001). Habit further contributed to continuance intention (β = 0.393, p < 0.001), with the model explaining 61.0% and 68.9% of the variance in affective satisfaction and continuance intention, respectively. Theoretically, this study advances structurally compressed engagement research by conceptualizing short-form structural features as contextual reinforcement conditions that strengthen the association between satisfaction and habit. Practically, the results suggest that micro-episodic design and rapid re-entry mechanisms play a critical role in stabilizing behavioral continuity in short-form drama ecosystems. Because the study used a cross-sectional survey design, the findings should be interpreted as correlational evidence of associations among digital experience factors, affective satisfaction, habit, and continuance intention rather than evidence of causal or temporal change.

Introduction

The global digital entertainment landscape has undergone a profound transformation with the rapid expansion of short-form content platforms. Fueled by mobile-first consumption habits, algorithmic recommendation systems, and increasingly fragmented patterns of media use, short-form dramas have emerged as one of the fastest-growing entertainment formats worldwide^1,2. Unlike traditional streaming services that rely on extended viewing sessions, short-form drama platforms deliver ultra-short episodes designed for rapid consumption and frequent re-entry throughout the day. As these platforms become deeply embedded in everyday digital routines, understanding the psychological mechanisms that sustain user engagement has become an increasingly important research priority.

Despite the growing popularity of short-form dramas, existing continuance research remains largely grounded in traditional digital entertainment environments and information systems continuance research³ such as online streaming services, video-on-demand platforms, and social media applications. However, short-form drama ecosystems possess several distinctive structural characteristics—including ultra-short episodic formats, high-frequency re-entry opportunities, and algorithmically reinforced exposure—that create dense reinforcement cycles which repeatedly expose users to emotionally rewarding content within compressed interaction sequences^1,4,5. Consequently, psychological processes underlying engagement in short-form drama environments may differ substantially from those observed in traditional streaming contexts.

A further limitation of prior research is that satisfaction and habit (HB) have typically been conceptualized as parallel predictors of continuance intention (CI) in information systems continuance research⁶. Although both constructs have consistently demonstrated significant explanatory power, relatively little attention has been devoted to understanding how affective satisfaction (as) may gradually evolve into habitual engagement through repeated reinforcement experiences. In particular, it remains unclear how structurally compressed digital environments influence the affective-to-behavioral pathway linking emotional gratification to automatic usage patterns. As a result, the psychological process through which positive affective experiences become routinized into habitual engagement remains insufficiently understood.

To address this gap, the present study develops and tests a reinforcement-based habit formation model within the stimulus-organism-response (S-O-R) framework. Specifically, we propose that digital experience factors influence continuance intention through a sequential affective–behavioral process in which affective satisfaction is conducive to habit formation, which subsequently predicts continuance intention. By examining this mechanism in the context of short-form drama consumption, this study extends continuance theory beyond static post-adoption explanations and advances understanding of how structurally compressed digital environments shape sustained user engagement.

Based on the identified research gap, we hypothesize that continuance intention in structurally compressed short-form drama environments is shaped through a sequential affective–behavioral reinforcement process. Specifically, digital experience factors enhance affective satisfaction, affective satisfaction facilitates habit formation, and habit subsequently reinforces continuance intention. This central hypothesis guides the proposed research model and the empirical investigation presented in this study.

Unlike prior applications of the S-O-R framework, this study adopts the S-O-R framework⁷ to examine how environmental stimuli influence internal affective states and subsequent behavioral responses in structurally compressed media ecosystems. We propose that digital experience factors influence continuance intention through a sequential process in which affective satisfaction fosters habit formation, which, in turn, predicts continuance intention. The framework has been widely applied in digital commerce, social media, and online entertainment research to explain how environmental attributes shape user engagement and behavioral persistence^8,9. Consumer technology adoption research has similarly emphasized the role of experiential factors in shaping continued technology use¹⁰. In digital environments, experiential features function as environmental stimuli that trigger emotional reactions and shape post-adoption behavior.

In the present study, digital experiences—comprising access convenience (AC), passing time motivation (PT), and perceived content quality (CQ)—are conceptualized as environmental stimuli. Access convenience reflects the perceived effortlessness and fluency associated with initiating media use, which in turn is associated with experiential control and reduces cognitive load^11,12. Passing time represents leisure-oriented motivation within fragmented temporal contexts, a key driver of mobile media engagement^13,14. Perceived content quality captures narrative immersion, relevance, and entertainment value, which are critical determinants of user satisfaction in digital media contexts¹⁵. Together, these experiential stimuli shape users' affective evaluations during media interaction.

Affective satisfaction represents the organismic state in the S-O-R framework. Satisfaction reflects users' overall emotional appraisal of their experience and has consistently been identified as a primary predictor of continuance intention in post-adoption contexts^3,16. However, in high-frequency and repetitive usage environments, satisfaction alone may be insufficient to explain persistent engagement. Repeated positive affective experiences may strengthen the association between engagement and habitual behavior.

Habit formation provides a complementary explanation for behavioral persistence. Habit refers to learned automatic behavioral tendencies triggered by contextual cues through repeated prior actions^6,17. Within information systems research, habit has been identified as a critical determinant of continued technology use beyond conscious intention⁶. In short-form drama consumption contexts, frequent micro-engagements embedded in daily routines may facilitate the development of habitual repetition. As habit strengthens, behavior becomes less dependent on active cognitive evaluation and more resistant to situational variability, thereby reinforcing continuance intention.

By integrating digital experiences, affective satisfaction, and habit formation within a sequential mediation framework, this study advances theoretical understanding of sustained engagement in short-form drama contexts. Specifically, we propose that digital experiences influence continuance intention through a two-stage affective-behavioral pathway: experiential stimuli enhance affective satisfaction, which, in turn, fosters habit formation, ultimately reinforcing continuance intention.

Using survey data from 260 short-form drama users and PLS-SEM, this study empirically examines this sequential mediation mechanism. Theoretically, it extends the S-O-R framework by incorporating habit as a reinforcement mechanism and provides empirical evidence for the sequential affective–habitual pathway in short-form digital environments.

The S-O-R framework posits that environmental stimuli influence internal psychological states, which subsequently shape behavioral responses. In this study, access convenience, passing time, and content quality are conceptualized as stimuli; affective satisfaction and habit represent organismic processes; and continuance intention represents the behavioral response.

Affective satisfaction reflects users' overall emotional evaluation of their media experience and has been consistently associated with continuance intention. In high-frequency digital environments, repeated positive affective experiences may also contribute to habit formation. Prior research in digital media and streaming contexts suggests that affective satisfaction functions as a key organismic state linking platform stimuli to continuance-related responses¹⁸. Based on the proposed theoretical framework and prior literature, the following hypotheses are proposed.

Access convenience refers to the perceived ease of accessing digital content. According to the S-O-R framework, environmental stimuli influence users' affective evaluations. Service convenience research further suggests that reduced time and effort costs enhance perceived value and satisfaction in service encounters¹⁰. Accordingly, it is hypothesized that access convenience positively influences affective satisfaction (H1).

Passing time is a leisure-oriented motivation for consuming short-form content. Prior research suggests that time-oriented consumption motives generate positive emotional responses and increase satisfaction. Therefore, we hypothesize that passing time positively influences affective satisfaction (H2).

Content quality reflects users' perceptions of the relevance and entertainment value of media content. Prior studies consistently identify it as a major determinant of user satisfaction and continued engagement¹⁹. Accordingly, it is hypothesized that content quality positively influences affective satisfaction (H3).

Affective satisfaction and habit have consistently been identified as important antecedents of continuance intention in post-adoption contexts. Prior habit research further shows that repeated performance in stable contexts strengthens automaticity over time, while mobile media environments can intensify routine-based usage patterns^4,20. Hence, we propose that affective satisfaction positively influences continuance intention (H4).

Habit refers to automatic behavioral tendencies that develop through repeated past behavior in stable contexts^6,17. In post-adoption contexts, repeated positive experiences facilitate the formation of habitual usage patterns, as users increasingly rely on automated responses rather than conscious evaluation^6,17. From a reinforcement-learning perspective, repeated behaviors accompanied by positive affective outcomes strengthen associative learning processes and increase the likelihood of future automatic enactment²¹. Once a habit is formed, it becomes a key mechanism that translates satisfaction into behavioral persistence. Habitual users tend to continue engaging with content with less cognitive effort, which strengthens continuance intention⁶. Accordingly, habit is conceptualized in this study as a mediating mechanism that links affective satisfaction to continuance intention. Based on the above rationale, it is hypothesized that affective satisfaction positively influences habit (H5), habit positively influences continuance intention (H6), and habit mediates the relationship between affective satisfaction and continuance intention (H7). Furthermore, affective satisfaction and habit are proposed to sequentially mediate the relationships between access convenience and continuance intention (H8a), passing time and continuance intention (H8b), and content quality and continuance intention (H8c).

Although prior studies have substantially advanced understanding of continuance behavior in digital media environments, several important gaps remain. Existing research has predominantly focused on traditional streaming, social media, or technology adoption contexts and has generally conceptualized satisfaction and habit as parallel determinants of continuance intention. Comparatively little attention has been devoted to understanding how structurally compressed short-form environments may shape the psychological process linking affective satisfaction and habit formation. Table 1^3,6,10,18,22summarizes representative prior studies and highlights the research gaps addressed by the present study.

Study	Research Context	Main Focus	Key Findings	Research Gap
Bhattacherjee (2001)3	Information Systems	Continuance Intention	Satisfaction predicts continuance intention	Habit formation process not examined
Limayem et al. (2007)6	Information Systems	Habit and Continuance	Habit influences continued system use	Satisfaction and habit treated as separate predictors
Venkatesh et al. (2012)10	Technology Acceptance	Post-adoption Behavior	Behavioral intention predicts continued use	Structural characteristics of digital environments not considered
Wu et al. (2025)18	Short-form Drama	Subscription Intention	Content quality affects continuance through satisfaction	Habit formation mechanism not examined
Zhang et al. (2024)22	Short-form Video	User Satisfaction	Satisfaction improves continued engagement	Satisfaction-to-habit conversion process unexplored
Recent Short-form Media Studies	Short-form Content Platforms	User Engagement	Algorithmic exposure increases engagement	Structural reinforcement mechanisms remain under-theorized
Present Study	Short-form Drama Ecosystem	Reinforcement-Based Habit Formation	Examines how affective satisfaction facilitates habit formation and continuance intention within structurally compressed environments	Addresses the identified gaps through a sequential affective–behavioral framework

Table 1: Summary of prior research and identified gaps. This table summarizes representative prior studies on digital continuance, satisfaction, habit, and short-form media engagement. Columns include: study author and year; research context; key constructs examined; the relationship between satisfaction and habit (parallel vs. sequential); and the specific research gap addressed by the present study.

This study empirically analyzes the relationships among independent variables (i.e., access convenience, passing time, and content quality), mediating variables (i.e., affective satisfaction and habit), and the dependent variable (i.e., CI). Gender, age, education, and income were included as control variables.

As shown in Figure 1, the model depicts a dense reinforcement cycle positioned between affective satisfaction and habit formation. This notation represents the structural context of short-form drama ecosystems—characterized by ultra-short episode duration, high re-entry frequency, and algorithmic content delivery—that creates abundant opportunities for cue-response pairing. The model proposes that environmental stimuli influence affective satisfaction, which subsequently facilitates habitual engagement under structurally compressed reinforcement cycles. Short-form ecosystems exhibit dense cue-response associations and repeated engagement opportunities that may strengthen the association between satisfaction and habit^23,24.

Figure 1: Reinforcement-based habit formation model in short-form drama environments. The 'Dense Reinforcement Cycle' denotes the high-frequency cue-response opportunities characteristic of short-form ecosystems (micro-episodes, rapid re-entry, algorithmic curation), theorized to intensify the satisfaction-to-habit relationship. This structural property is conceptualized as a contextual condition rather than an empirically measured temporal acceleration. Control variables (age, gender, education, and income) are excluded from the depicted pathways for visual clarity. Please click here to view a larger version of this figure.

Protocol

All procedures involving human participants were conducted in accordance with the ethical standards for human-subject research at aSSIST University and Dong-A University. As all responses were anonymized at the time of collection and no personally identifiable information was obtained at any stage, the Institutional Review Board of aSSIST University determined that the study qualified for exemption from full review under the set forth in Article 15 of the Bioethics and Safety Act of the Republic of Korea and the requirements for exemption from review set forth in Article 13 of its Enforcement Rule (https://public.irb.or.kr/pt/pt01/PT0103/PT0103R01.do). The exemption approval ID was obtained before beginning the survey; all participants reviewed an informed consent statement describing the study purpose, voluntary participation, anonymity, data use, and their right to discontinue participation. Only participants who provided electronic informed consent were allowed to proceed. No personally identifiable information was collected.

Data collection and analysis

Sample characteristics

Participants were recruited through Prolific from English-speaking countries, including the United States, the United Kingdom, Canada, Australia, and New Zealand, between July 5 and July 10, 2025. The questionnaire consisted of validated multi-item measures assessing access convenience, passing time, content quality, affective satisfaction, habit, and continuance intention, together with demographic questions. Participants were eligible if they were at least 18 years old, had prior experience watching short-form dramas, and actively used social networking services. Eligibility was confirmed using screening questions asking participants to confirm their age, prior short-form drama viewing experience, and current social networking service use. Participants who did not meet these criteria were excluded before completing the full questionnaire. Attention-check items were embedded in the survey to identify inattentive responses; participants who failed any attention-check item, submitted incomplete responses, or failed screening questions were excluded from analysis. A total of 292 responses were collected. After excluding 32 responses due to screening failure, incomplete responses, or failed attention-check items, 260 valid responses were retained for analysis. The survey required approximately 8–10 min to complete, and participants received approximately €2.00 ($3.440).

Measurement model

Responses were recorded on a 7-point Likert scale ranging from 1 (strongly disagree) to 7 (strongly agree). This study validated the research model using the PLS-SEM^25,26. PLS-SEM was selected because the study is prediction-oriented, examines a relatively complex mediation model, and focuses on explained variance in endogenous constructs.

Mediation analysis

Statistical significance was evaluated using bias-corrected bootstrapping with 5,000 resamples (Figure 2). Standardized path coefficients, indirect effects, effect sizes (f²), predictive relevance (Q²), and 95% confidence intervals were examined to evaluate the proposed hypotheses and mediation effects.

Figure 2: Research procedure and data analysis workflow. Eligible participants were active SNS users aged 18 years or older. The figure illustrates participant recruitment through the online survey platform, screening and attention-check procedures, survey data collection (N = 260), data cleaning, measurement model assessment, structural model assessment using PLS-SEM, and mediation testing through bootstrapping with 5,000 resamples. Please click here to view a larger version of this figure.

Structural model

Validated measurement scales for all study constructs were adapted from previously published studies. Minor wording modifications were made where necessary to ensure their suitability for the short-form drama context. The measurement items are presented in Table 2. Although the content quality scale was originally developed in broader streaming and digital entertainment contexts, it captures users' perceptions of the breadth and diversity of entertainment. In short-form drama contexts, entertainment formats may include hybrid or cross-genre elements that extend beyond traditional narrative episodes. Therefore, the item referring to entertainment variety reflects perceived content diversity rather than literal sports broadcasting and is conceptually consistent with the construct of content quality in this study.

Construct	Item	Question
Access convenience	AC1	I could watch short-form drama anytime I wanted.
	AC2	I could watch short-form drama wherever I am.
	AC3	I could use any digital device to access short-form drama content.
Passing time	PT1	I would watch short-form dramas because it passes the time away, particularly when I am bored
	PT2	I would watch short-form dramas when I have nothing better to do
	PT3	I would watch short-form dramas because it gives me something to occupy my time
Content quality	CQ1	Short-form drama content offers a wide range of content.
	CQ2	Short-form drama content offers exclusive and original content.
	CQ3	Short-form drama content provides a diverse range of entertaining episodes and genres.
	CQ4	Short-form drama content offers a wide variety of global content.
Affective media satisfaction	AS1	I feel happy after spending time with short-form drama
	AS2	Short-form drama gives me pleasure
	AS3	Short-form drama experience is enjoyable
	AS4	Feel good after watching short-form drama
Habit	HB1	I watch short-form drama out of habit.
	HB2	I access and use short-form drama contexts without thinking.
	HB3	I use short-form drama contexts naturally.
Continuance intention	CI1	I intend to use the short-form drama streaming in the near future.
	CI2	I am willing to continue using short-form drama content in the near future.
	CI3	I will maintain my engagement with short-form drama content in the near future.

Table 2: Measurement items of the research model. This table lists the measurement items, construct labels, and source references used to operationalize the research variables. All items were measured on a 7-point Likert scale ranging from 1 (strongly disagree) to 7 (strongly agree). The table presents each construct (access convenience, passing time, content quality, affective satisfaction, habit, and continuance intention), its associated indicator items, and the original study from which each scale was adapted.

Common method bias

Because all variables were collected using self-reported survey data, common method bias (CMB) was assessed using both procedural and statistical remedies. First, several procedural remedies were implemented to reduce potential method bias. Participation was voluntary and anonymous, and respondents were assured that there were no right or wrong answers. Measurement items were adapted from validated prior studies and were carefully reworded to fit the short-form drama context. The questionnaire was structured to minimize evaluation apprehension and reduce item ambiguity. Second, Harman's single-factor test was conducted to examine whether a single factor accounted for the majority of the variance. The unrotated exploratory factor analysis revealed that the first factor explained less than 50% of the total variance, suggesting that common method bias is unlikely to be a serious concern. Third, full collinearity variance inflation factors (VIFs) were examined following Kock and Mayfield²⁷. All VIF values were below the conservative threshold of 3.3 (Table 3), indicating that common method bias does not substantially affect the model estimates. Nevertheless, because all variables were collected from the same respondents within a single survey wave, residual common-method bias cannot be completely ruled out.

Inner Model Path	VIF
AC→AS	1.342
AS→HB	1.000
AS→CI	1.933
CQ→AS	1.410
HB→CI	1.939
PT→AS	1.154

Table 3: Collinearity statistics (VIF). This table presents the variance inflation factor (VIF) values for all predictor constructs used in the structural model. Rows list each predictor construct (access convenience, passing time, content quality, affective satisfaction, and habit), and the corresponding VIF values are reported in the adjacent column. VIF values were examined to assess potential multicollinearity among constructs and to evaluate the likelihood of common method bias.

Troubleshooting and modifications

If a substantial number of responses failed the screening or attention-check criteria, the eligibility criteria were refined and the survey instructions clarified to ensure participants had genuine prior experience with short-form drama. If internal consistency reliability fell below recommended thresholds (e.g., composite reliability < 0.70) or average variance extracted (AVE) fell below 0.50, item loadings were examined, and poorly performing indicators were removed while maintaining theoretical coherence. If multicollinearity was detected (e.g., variance inflation factor values exceeded recommended limits), construct specifications were reassessed, and redundant indicators were removed. If mediation or moderation effects were non-significant, the sample size was increased and/or construct operationalization was re-examined to ensure adequate statistical power. Bootstrapping was conducted using a sufficient number of subsamples (e.g., 5,000) to obtain stable parameter estimates.

Results

Sample characteristics

Table 4 shows a well-balanced sample of 260 respondents across key demographic categories. By age, the largest group is participants in their 20s or younger (38.8%), followed by those in their 30s (32.3%). Respondents in their 40s (14.2%) and 50s (14.6%) are represented in smaller but similar proportions. Gender is almost evenly distributed, with 49.2% male and 50.8% female participants. In terms of education, university graduates make up the largest group (50.0%), followed by those with graduate school education (26.5%). Smaller shares reported high school (12.7%) and college-level education (10.8%). For annual household income, the sample is relatively spread across income brackets, though the largest proportion is under $30,000 (23.1%), while the smallest is $100,000 or more (16.5%). Overall, the income distribution appears fairly diverse without heavy concentration in a single category.

Category	Item	Frequency	Ratio
Age	20s or younger	101	38.80%
	30s	84	32.30%
	40s	37	14.20%
	50s	38	14.60%
Gender	Male	128	49.20%
	Female	132	50.80%
Education	High school	33	12.70%
	College	28	10.80%
	University	130	50.00%
	Graduate school	69	26.50%
Annual household	Under $30,000	60	23.10%
income	$30,000–$49,999	52	20.00%
	$50,000–$74,999	57	21.90%
	$75,000–$100,000	48	18.50%
	$100,000 or more	43	16.50%

Table 4: Demographic characteristics of respondents (N = 260). This table presents the demographic distribution of the final analytical sample. Rows list the demographic categories (gender, age, education, and income), and columns present the frequency (n) and percentage (%) for each category.

Measurement model

As shown in Table 5, the external loadings of all constructs are well above the minimum threshold. CR and Cronbach's α values are above 0.7, indicating high internal consistency. The AVE value exceeds 0.5, which satisfies convergent validity²⁶.

Construct	Cronbach's alpha	CR (rho_a)	CR (rho_c)	AVE
AC	0.830	0.833	0.898	0.747
AS	0.958	0.959	0.970	0.889
CQ	0.790	0.821	0.861	0.610
HB	0.899	0.915	0.937	0.831
PT	0.850	0.896	0.907	0.765
CI	0.923	0.923	0.951	0.866

Table 5: Reliability analysis. This table presents indicator loadings, Cronbach's alpha, composite reliability (CR), and average variance extracted (AVE) for all constructs. Rows list each construct (access convenience, passing time, content quality, affective satisfaction, habit, and continuance intention) and its indicator items, while columns present the standardized factor loadings, Cronbach's alpha, CR, and AVE values. These metrics were used to assess internal consistency, reliability, and convergent validity.

Discriminant validity was tested using the Fornell–Larcker and the heterotrait–monotrait ratio (HTMT) criteria. Table 6 shows that the square roots of the AVEs are greater than the correlations with other components for each component. Likewise, Table 7 shows that all HTMT ratios are significantly below the 0.85 threshold²⁸.

	AC	AS	CQ	HB	PT	CI
AC	0.864
AS	0.497	0.943
CQ	0.493	0.666	0.781
HB	0.461	0.686	0.612	0.912
PT	0.274	0.587	0.346	0.433	0.875
CI	0.552	0.777	0.646	0.744	0.491	0.931

Table 6: Fornell-Larcker criterion. This table presents the Fornell-Larcker discriminant validity assessment. Rows and columns represent the study constructs (access convenience, passing time, content quality, affective satisfaction, habit, and continuance intention). Diagonal elements represent the square root of the AVE for each construct, whereas off-diagonal elements represent inter-construct correlations.

	AC	AS	CQ	HB	PT	CI
AC
AS	0.557
CQ	0.592	0.738
HB	0.523	0.73	0.724
PT	0.3	0.627	0.372	0.47
CI	0.63	0.825	0.742	0.806	0.519

Table 7: HTMT ratio. This table presents HTMT values for all construct pairs. Rows and columns represent the study constructs (access convenience, passing time, content quality, affective satisfaction, habit, and continuance intention). HTMT values below the recommended threshold indicate satisfactory discriminant validity between constructs.

Common method bias was additionally assessed using Harman's single-factor test and full collinearity VIF analysis. The first factor accounted for less than 50% of the total variance, and all VIF values were below 3.3, suggesting that common method bias was unlikely to threaten the validity of the findings.

Structural model

As shown in Figure 3 and Table 8, the structural model results indicate that affective satisfaction (AS) was significantly predicted by access convenience (β = 0.171, p < 0.01), passing time (β = 0.385, p < 0.001), and content quality (β = 0.448, p < 0.001), explaining 61.0% of its variance (R² = 0.610), with content quality exerting the strongest influence. Affective satisfaction strongly predicted Habit (β = 0.686, p < 0.001; R² = 0.470) and also had a significant direct effect on continuance Intention (β = 0.496, p < 0.001). Habit further contributed significantly to continuance Intention (β = 0.393, p < 0.001), which was explained at a substantial level (R² = 0.689). These findings suggest a partial mediation pattern in which affective satisfaction is linked to continuance intention both directly and indirectly through habit formation. The control variables (age, gender, education, and income) did not show significant effects on continuance intention.

Predictive relevance (Q²) was assessed using Stone-Geisser's Q². All endogenous constructs exhibited Q² values above zero (AS = 0.412, HB = 0.356, CI = 0.523), indicating satisfactory predictive relevance.

In addition to examining path significance, the effect size (f²) results provide insight into the substantive contribution of each predictor to the explained variance of the endogenous constructs. Regarding antecedents of AS, CQ exhibits a large effect (f² = 0.365), indicating that it is a primary driver of AS in the model. PT also demonstrates a substantial contribution (f² = 0.329), approaching the threshold for a large effect and suggesting strong practical relevance. In contrast, AC shows only a small effect (f² = 0.056), implying a comparatively limited incremental contribution to AS. These findings suggest that while all three antecedents contribute to AS, CQ and PT are the dominant explanatory factors, whereas AC plays a more supportive role.

For the downstream relationships, AS exerts an exceptionally large effect on HB (f² = 0.887), underscoring its critical role in explaining habitual behavior. Furthermore, AS shows a large effect on CI (f² = 0.409), indicating a substantial contribution to the model's explanatory power for continuance outcomes. HB also contributes meaningfully to CI (f² = 0.256), with a medium-to-large effect size, though its incremental impact is smaller than that of AS. Overall, the effect size analysis highlights AS as the central mechanism in the structural model, exerting dominant influence on both HB and CI, while CQ and PT emerge as key upstream determinants of AS.

These findings reinforce the theoretical proposition that affective reinforcement plays a dominant role in shaping habitual engagement, which subsequently stabilizes continuance intention in short-form digital media contexts.

Structural model results (Figure 3) illustrate the standardized path coefficients (β) and coefficients of determination (R²) for the proposed research model. Control variables (age, gender, education, income) were included in the model but are omitted from the figure for visual clarity.

Figure 3: Structural model results. The results show the standardized path coefficients (β), significance levels, and coefficients of determination (R²) for the proposed research model. AC = Access Convenience; PT = Passing Time; CQ = Content Quality; AS = Affective Satisfaction; HB = Habit; CI = Continuance Intention. β represents standardized path coefficients, and R² represents explained variance for endogenous constructs. Significance levels are indicated as follows: *p < 0.05, **p < 0.01, and ***p < 0.001. Control variables (age, gender, education, and income) were included in the analysis but are omitted from the figure for visual clarity. Please click here to view a larger version of this figure.

Path coefficients and hypothesis testing results (Table 8) present the standardized path coefficients (β), standard errors (SE), t-values, p-values, and effect sizes (f²) for all hypothesized relationships in the structural model.

Path	β	Mean	STD	t value	p value
AC→AS	0.171	0.173	0.061	2.815	0.005
AS→HB	0.686	0.687	0.03	22.864	0.000
AS→CI	0.496	0.496	0.052	9.558	0.000
CQ→AS	0.448	0.449	0.052	8.56	0.000
HB→CI	0.393	0.393	0.054	7.284	0.000
PT→AS	0.385	0.380	0.059	6.492	0.000

Table 8: Path coefficients. This table presents the results of the structural model analysis. Rows list each hypothesized path in the structural model (e.g., access convenience → affective satisfaction, affective satisfaction → habit), while columns present the standardized path coefficients (β), standard errors (SE), t-values, p-values, and effect sizes (f²) for all hypothesized relationships. Significance levels: p < 0.001, p < 0.01, p < 0.05.

Mediation analysis

To examine the sequential mediation mechanism, we analyzed the indirect paths from environmental attributes to CI through AS and HB. The results of the specific indirect-effect analysis (Table 9) revealed that all sequential paths were statistically significant. The mediation analysis based on bootstrapping indicates that AS and HB transmit significant indirect effects across the model. Specifically, AS significantly mediates the effects of the antecedents on HB: the indirect effects are positive and statistically significant for AC → AS → HB (β = 0.117, t = 2.835, p = 0.005), CQ → AS → HB (β = 0.307, t = 7.822, p < 0.001), and PT → AS → HB (β = 0.264, t = 5.993, p < 0.001). Among these, the magnitude of the indirect effect is largest for CQ, followed by PT, with AC showing a smaller yet reliable mediated influence. This pattern suggests that improvements in AC, CQ, and PT primarily translate into stronger habitual outcomes through increases in AS, with CQ operating as the most influential upstream pathway.

In addition, the results provide consistent evidence that continuance-related outcomes (CI) are shaped through both single-step mediation via AS and serial mediation via AS and HB. The indirect effects of AC → AS → CI (β = 0.085, t = 2.598, p = 0.009), CQ → AS → CI (β = 0.222, t = 6.774, p < 0.001), and PT → AS → CI (β = 0.191, t = 5.091, p < 0.001) are all positive and significant, indicating that AS serves as a key mechanism linking the antecedents to CI. Moreover, the serial indirect effects through both mediators are also significant: AC → AS → HB → CI (β = 0.046, t = 2.636, p = 0.008), CQ → AS → HB → CI (β = 0.121, t = 5.261, p < 0.001), and PT → AS → HB → CI (β = 0.104, t = 4.638, p < 0.001). Together, these findings indicate a robust process in which antecedents first increase AS, which then strengthens HB and ultimately is associated with CI, while AS also exerts a direct mediating pathway to CI independent of HB. Overall, CQ shows the strongest mediated influence on both HB and CI, PT exhibits comparably strong indirect effects, and AC demonstrates smaller but statistically meaningful indirect effects, supporting the proposed mediation and serial mediation mechanisms in the model. Since the confidence intervals do not include zero, these findings are consistent with the proposed 'satisfaction-habit-continuance' chain in the context of short-form drama consumption.

Path	β	Mean	STD	t value	p value
AC→AS→HB	0.117	0.119	0.041	2.835	0.005
AS→HB→CI	0.270	0.270	0.040	6.776	0.000
AC→AS→CI	0.085	0.086	0.033	2.598	0.009
CQ→AS→HB	0.307	0.308	0.039	7.822	0.000
AC→AS→HB→CI	0.046	0.047	0.017	2.636	0.008
CQ→AS→CI	0.222	0.222	0.033	6.774	0.000
PT→AS→HB	0.264	0.261	0.044	5.993	0.000
CQ→AS→HB→CI	0.121	0.121	0.023	5.261	0.000
PT→AS→CI	0.191	0.189	0.038	5.091	0.000
PT→AS→HB→CI	0.104	0.103	0.022	4.638	0.000

Table 9: Indirect effects. This table presents the results of the bootstrapped mediation analysis. Rows list each hypothesized indirect path (e.g., AC → AS → HB, CQ → AS → HB → CI), while columns present the indirect effect estimates, standard errors, t-values, p-values, and 95% bias-corrected confidence intervals for the hypothesized mediation pathways.

Table 10 summarizes the hypothesis testing results. As shown in Table 10, all proposed hypotheses (H1–H8c) were supported.

Hypothesis	Path	Result
H1	AC→AS	Supported
H2	PT→AS	Supported
H3	CQ→AS	Supported
H4	AS→CI	Supported
H5	AS→HB	Supported
H6	HB→CI	Supported
H7	AS→HB→CI	Supported
H8a	AC→AS→HB→CI	Supported
H8b	PT→AS→HB→CI	Supported
H8c	CQ→AS→HB→CI	Supported

Table 10: Hypothesis support summary. This table summarizes the outcomes of hypothesis testing and indicates whether each proposed hypothesis was supported based on the structural model and mediation analyses. Rows list each hypothesis (H1–H8c), and columns present the hypothesized path, the standardized path coefficient or indirect effect, and the support status (supported vs. not supported).

DATA AVAILABILITY

All anonymized raw survey data supporting the findings of this study are provided in Supplementary File 1 (CSV format). The file contains the complete de-identified participant responses used in all analyses reported in this manuscript. The SmartPLS project file, model settings, and analysis outputs are provided in Supplementary File 2. No personally identifiable information is included in any shared dataset.

Supplementary File 1: All anonymized raw survey data supporting the findings of this study.Please click here to download this file.

Supplementary File 2: The SmartPLS project file, model settings, and analysis outputs.Please click here to download this file.

Discussion

This study aimed to explain continuance intention in short-form drama contexts by integrating digital experience factors, affective satisfaction, and habit formation within the S-O-R framework. The findings provide strong empirical support for the proposed sequential mechanism and suggest that digital continuance is best understood as a dynamic affective–behavioral reinforcement process. Digital experience factors—access convenience, passing time, and content quality—significantly influence affective satisfaction, with content quality exerting the strongest effect, followed by passing time, while access convenience plays a comparatively smaller yet significant role. This pattern suggests that while frictionless accessibility is necessary in fragmented viewing contexts, emotionally engaging and immersive content remains the primary determinant of users' positive affective evaluations. In short-form drama environments, where narrative compression and rapid engagement are central, perceived content richness and creativity appear to be the most powerful emotional stimuli. Affective satisfaction also emerges as the central mechanism in the structural model, contributing to habit formation and continuance intention through the "satisfaction-habit-continuance" chain.

These findings extend prior digital continuance research by reframing continuance intention as a sequential affective-behavioral process rather than a static post-adoption outcome. Traditional continuance models, including expectation-confirmation theory, primarily emphasize cognitive evaluation and satisfaction as direct predictors of sustained use^29,30. In contrast, the present findings demonstrate that in high-frequency, structurally compressed digital environments, affective satisfaction may function not only as an evaluative outcome but also as a psychological correlate of habitual engagement. This interpretation is consistent with habit theory, which suggests that repeated behavior under stable contextual cues gradually strengthens automatic behavioral tendencies. However, because the present study does not directly test temporal acceleration or causal habit formation, short-form drama environments may be understood as conditions under which habitual engagement is more likely to be reinforced.

The study also extends the S-O-R framework by integrating habit as a sequential reinforcement mechanism linking affective states to behavioral persistence. While prior S-O-R applications often position satisfaction and habit as parallel predictors of continuance, the present model demonstrates that habit emerges as a downstream consequence of affective reinforcement. This sequential affective-habitual pathway provides a more temporally coherent explanation of post-adoption behavior, particularly in mobile-first entertainment contexts characterized by rapid and repetitive engagement cycles. The findings further introduce structurally compressed engagement environments as contextual conditions that may strengthen the association between affective responses and habitual behavior. Although structural features such as micro-episodes, rapid re-entry opportunities, and algorithmic prompts were not directly operationalized in this model, the theoretical framing highlights how dense reinforcement cycles embedded in short-form ecosystems can intensify affect-driven behavioral stabilization. Importantly, this mechanism reflects normative routinization rather than pathological dependence, highlighting the distinction between habitual engagement and compulsive usage.

The findings offer practical implications for short-form drama platforms and content providers. Understanding the central role of content quality in affective satisfaction suggests that narrative originality, emotional storytelling, and immersive production elements are key environmental stimuli that generate positive affective responses, subsequently supporting repeated engagement and sustained use. Technical optimization alone is insufficient without content that generates genuine emotional resonance. User experience design should also consider how environmental features reinforce positive affective states during repeated interactions. Features such as seamless transitions, emotionally engaging previews, and personalized recommendations can amplify affective satisfaction and facilitate habitual engagement. Although access convenience exerts a smaller effect on affective satisfaction compared to content quality, reduced friction in digital environments remains important because it preserves experiential fluency and prevents disruptions in the affective reinforcement cycle. Overall, the findings suggest that sustainable digital media engagement can be supported by combining emotional engagement principles with behavioral design while monitoring both satisfaction and habitual usage patterns.

Despite its contributions, this study has several limitations that offer directions for future research. First, the cross-sectional survey design limits causal inference and temporal ordering among the proposed constructs. Although this study conceptualized short-form drama environments as dense reinforcement contexts, the structural characteristics underlying reinforcement density were not directly measured. Therefore, the findings should be interpreted as evidence of association rather than temporal acceleration or causal habit formation. Future research should employ longitudinal or experimental designs and explicitly operationalize reinforcement-density-related variables to validate the proposed mechanism. Second, the data were collected in a single short-form drama context and relied exclusively on self-reported measures, which may limit generalizability and introduce common-method bias despite statistical controls. Future research should examine different cultural settings, age groups, and platform contexts while incorporating behavioral tracking data to complement self-reported responses. In addition, the present study did not collect platform-specific usage information (e.g., ReelShort, DramaBox, ShortMax, TikTok-based drama platforms). Consequently, the findings should not be generalized uniformly across all short-form drama ecosystems. Future studies should compare users across platforms and examine whether platform characteristics and usage intensity moderate the relationships among affective satisfaction, habit, and continuance intention. Third, although this study distinguishes habitual engagement from the pathological dependence conceptually, future research could incorporate measures of problematic use or self-control to further clarify the boundary conditions between healthy routinization and excessive consumption. This study also focused specifically on short-form drama content; therefore, generalizability to other forms of short-form content, such as general short videos or social media reels, requires further empirical testing. Additional psychological constructs, such as emotional attachment, media trust, and perceived personalization, may further refine the understanding of continuance processes in short-form structural ecosystems. In summary, this study demonstrates that continuance intention in short-form drama contexts is driven by a sequential affective-behavioral reinforcement process in which digital experiences shape affective satisfaction, satisfaction fosters habit formation, and habit stabilizes behavioral persistence. By integrating emotional and automatic mechanisms within the S-O-R framework, the findings provide a comprehensive explanation of sustained engagement in fragmented short-form digital ecosystems.

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
Chatgpt 5.2	OpenAI	RRID:SCR_023775	Enhancing readability and grammar check,  https://www.chatgpt.com/
Prolific (Online Participant Recruitment Platform)	Prolific Academic Ltd	Not available / Not identified	Online platform used to recruit survey participants. Participants were screened based on eligibility criteria (e.g., current employment in an organization with AI adoption experience) and compensated according to Prolific's fair-pay guidelines. https://www.prolific.com/
Qualtrics (Online Survey Platform)	Qualtrics, LLC	RRID:SCR_016728	Web-based survey platform used for questionnaire design, distribution, and response data collection. All measurement items were administered via Qualtrics using a 7-point Likert scale (1 = strongly disagree, 7 = strongly agree). https://www.qualtrics.com/
SmartPLS 4.0 (PLS-SEM Software)	SmartPLS GmbH	RRID:SCR_022040	Used for all Partial Least Squares Structural Equation Modeling (PLS-SEM) analyses, including outer model assessment (composite reliability, AVE, HTMT), inner model path analysis (bootstrapping, 5,000 samples), MICOM 3-step compositional invariance testing (permutation, 1,000 draws), and PLS-MGA between-group difference tests. https://smartpls.com/