Determinants of Donation Intention in Live Streaming–Model Inventory

Abstract

The Determinants of Donation Intention in Live Streaming–Model Inventory is a multidimensional psychometric measurement instrument designed to evaluate the psychological, social, and economic factors governing viewers’ decisions to financially support content creators in synchronous interactive media environments. Grounded in the dedication-constraint framework, social exchange theory, and the psychology of parasocial interactions, the instrument operationalizes the complex transition from passive viewership to active monetary patronage. The inventory comprises 31 self-report items structured into nine theoretically distinct subscales: Social Interaction (4 items), Parasocial Interaction (4 items), Perceived Value (4 items), Trust in the Streamer (4 items), Affective Commitment (4 items), Continuance Intention (3 items), Attractive Alternatives (2 items), Switching Costs (3 items), and Donation Intention (3 items). Responses are captured using an authentic 7-point Likert scale ranging from 1 (Strongly disagree) to 7 (Strongly agree). Psychometric evaluations using structural equation modeling demonstrate robust construct validity, high internal consistency (Cronbach’s alpha values exceeding .70; Composite Reliability values ranging from .877 to .929), strong convergent validity (factor loadings exceeding .70; Average Variance Extracted values yielding square roots between .829 and .892), and verified discriminant validity confirmed through the Fornell-Larcker criterion and the Heterotrait-Monotrait ratio of correlations (all HTMT values below .85). By dissecting both dedication-based mechanisms (e.g., emotional attachment, parasocial bonding, trust) and constraint-based mechanisms (e.g., perceived switching costs, lack of attractive alternatives), the scale provides a rigorous empirical diagnostic for behavioral scientists, communication scholars, media psychologists, and platform strategists investigating digital consumer behavior and computer-mediated relationship maintenance.

Keywords

Live Streaming, Donation Intention, Dedication-Constraint Framework, Parasocial Interaction, Social Commerce, Affective Commitment, Switching Costs, Digital Gifting, Cyberpsychology, Structural Equation Modeling

Authors

The scale was developed and psychometrically validated by researchers investigating interactive media adoption, relationship marketing, and user engagement behaviors within digital platforms:

• Shih-Wei Chou — Department of Information Management, National Kaohsiung University of Science and Technology, Kaohsiung City, Taiwan. Email: [email protected]. Specializes in knowledge management, social media dynamics, consumer-platform relationship architecture, and digital behavioral modeling.
• Trieu Nguyen — International Master of Business Administration (IMBA) Program, College of Management, National Kaohsiung University of Science and Technology, Kaohsiung City, Taiwan. ORCID: 0000-0003-1114-4715. Email: [email protected]. Correspondence Address: No. 1, University Road, Yanchao District, Kaohsiung City, Taiwan. Specializes in live-streaming user psychology, post-adoption consumer commitment, and quantitative psychometrics.

Purpose

The primary purpose of the Determinants of Donation Intention in Live Streaming–Model Inventory is to provide a comprehensive, empirically validated diagnostic instrument for measuring the psychological, interpersonal, and structural forces that motivate digital audiences to voluntarily donate real money and virtual gifts to streamers. Live streaming has transformed digital entertainment from a static, one-way broadcasting model into a dynamic, synchronous, and co-created socio-technical ecosystem. Unlike subscription-based platforms or paywalled media services, live-streaming channels typically operate on an open-access model where content is freely viewable without mandatory financial compensation. Consequently, financial contributions—manifested through direct monetary tips, tokenized micro-transactions, and animated digital gifts—represent purely discretionary, prosocial economic actions.

Understanding why users freely allocate economic resources to creators in the absence of legal or contractual obligations represents an urgent inquiry at the intersection of media psychology, consumer research, and behavioral economics. The inventory addresses this need by operationalizing the dedication-constraint paradigm within synchronous interactive environments. In practical terms, donation decisions do not occur within an emotional or cognitive vacuum; they are driven by positive relational bonding (“dedication”—the desire to maintain a meaningful connection with the broadcaster and the community) alongside structural limitations and relational inertia (“constraints”—the perceived burdens of terminating engagement or replacing the broadcaster with an alternative channel).

In academic and applied research environments, this measurement model serves several critical functions:

• Deconstructing Relational Value: It enables researchers to isolate the unique variance contributed by immediate conversational interactions (synchronous chat feedback) versus deep socio-emotional bonds (perceptions of the streamer as a reliable friend or peer).
• Analyzing Dual Commitment Mechanisms: It permits behavioral scientists to examine whether digital donations are sustained primarily through positive affective resonance (genuine admiration and perceived mutual value) or through defensive behavioral lock-in (high social switching costs and low availability of viable alternative creators).
• Evaluating Cyberpsychological Dependencies: It provides a structured foundation for investigating borderline compensatory behaviors, such as digital gifting driven by an excessive need for public social validation or compulsive identity affirmation within virtual peer groups.
• Informing Platform and Creator Strategies: For digital platform designers and content managers, the scale offers actionable insight into the service attributes (e.g., two-way responsiveness, perceived integrity, communal co-presence) that convert transient passive spectators into loyal, monetarily supportive brand advocates.

Psychological Construct

The Determinants of Donation Intention in Live Streaming–Model Inventory is conceptualized as an integrated, multi-construct measurement model. Rather than treating monetary gifting as an isolated behavioral reaction, the inventory operationalizes the cognitive, emotional, and structural antecedents that culminate in financial intent. The framework spans nine core dimensions:

1. Social Interaction (Items 1–4)

This construct measures the perceived quality, immediacy, and mutual effort present in the synchronous communication between the content creator and audience members. Rooted in the principles of computer-mediated communication, social interaction assesses the extent to which the broadcaster actively listens, demonstrates genuine interest in audience commentary, solicits audience sentiment regarding content preferences, and dedicates substantial personal effort toward maintaining real-time dialogue.

2. Parasocial Interaction (Items 5–8)

Originally formulated in mass communications theory, parasocial interaction in synchronous streaming captures an illusion of reciprocal intimacy that mimics authentic interpersonal friendships. This dimension measures whether the viewer perceives the streamer as a natural, unpretentious, “down-to-earth” companion, feels akin to an old acquaintance, experiences a sense of collective belonging within the streamer’s peer group, and experiences strong interpersonal attraction toward the creator’s persona.

3. Perceived Value (Items 9–12)

Perceived value reflects the consumer’s cognitive cost-benefit appraisal derived from consuming the stream. Drawing from consumer psychology and utility theory, this dimension evaluates whether the informational, entertainment, and social rewards acquired from watching the channel adequately compensate for the viewer’s non-monetary resource investments, specifically temporal expenditure and cognitive effort.

4. Trust in the Streamer (Items 13–16)

Trust captures the user’s psychological confidence in the moral integrity, benevolence, and authenticity of the broadcaster. Within digital commerce and peer-to-peer donation environments, trust serves as a critical uncertainty-reduction mechanism. The subscale gauges beliefs regarding the streamer’s honesty, benevolence toward viewers, general reliability, and overall trustworthiness.

5. Affective Commitment (Items 17–20)

Affective commitment embodies the emotional attachment, identification, and psychological involvement a viewer maintains with a specific streamer. In contrast to purely calculative engagement, this construct taps into deep social identity processes, evaluating whether the viewer experiences a strong sense of communal belonging, feels emotionally bonded, derives personal pride from being affiliated with the channel, and attributes high personal meaning to the streamer.

6. Continuance Intention (Items 21–23)

Continuance intention captures the viewer’s forward-looking behavioral commitment to repeatedly visit and consume the broadcaster’s content over time. Anchored in expectation-confirmation theory and post-adoption literature, this construct captures habituation and persistent engagement intentions, serving as a vital behavioral anchor preceding monetary expenditure.

7. Attractive Alternatives (Items 24–25)

Adapted from relationship dissolution theories and market competition models, this subscale captures the user’s perception of available substitute creators. It evaluates whether the viewer believes switching to an alternative streaming channel would produce an equivalent or superior level of satisfaction, reflecting the degree to which the current streamer enjoys a differentiated, irreplaceable relational monopoly.

8. Switching Costs (Items 26–28)

Switching costs represent the perceived friction, cognitive burden, lost communal status, and temporal investments required to terminate viewership with an incumbent streamer and establish comparable familiarity with another channel. This dimension captures the structural lock-in mechanism of the dedication-constraint paradigm, gauging whether switching channels is viewed as stressful, cumbersome, or excessively effortful.

9. Donation Intention (Items 29–31)

The ultimate focal outcome construct operationalizes the consumer’s self-reported behavioral inclination, explicit planning, and anticipated frequency of sending monetary tips, virtual gifts, or micro-transactions to the streamer during future broadcasts. It reflects the convergence of dedication and constraint pressures into concrete economic support.

Theoretical Framework

The conceptual architecture of the inventory is primarily anchored in the Dedication-Constraint Framework (Stanley & Markman, 1992; Bendapudi & Berry, 1997), a foundational paradigm in interpersonal relationship theory and relationship marketing. The framework asserts that human relational persistence is governed by two complementary yet structurally distinct psychological forces:

1. Dedication Mechanisms: Positive psychological forces that make an individual desire to maintain a relationship due to intrinsic rewards, deep affection, alignment of values, and personal meaning. In the context of live streaming, dedication is fostered through genuine real-time social interactions, authentic parasocial bonds, cognitive value appraisals, interpersonal trust, and affective commitment. Here, donation acts as an uncoerced gesture of reciprocity, gratitude, and communal bonding.
2. Constraint Mechanisms: Structural, economic, or psychological barriers that make an individual feel they must or have to maintain a relationship due to the punitive costs, perceived inconveniences, or relational voids associated with termination. Within digital streaming channels, constraints materialize as high switching costs (the social disruption of abandoning a familiar community and losing recognized chat status) and the absence of attractive alternatives.

The model additionally integrates Social Exchange Theory (Blau, 1964). According to social exchange paradigms, sustained human relationships rely on generalized reciprocity wherein benefits provided by one party (such as entertainment, emotional companionship, social validation, and responsive conversation provided by the streamer) generate an internal psychological tension or moral obligation in the recipient. Because traditional access to the broadcast is cost-free, voluntary financial donations serve as the viewer’s currency of equitable social balance, allowing them to reimburse the streamer for subjective utility received.

Furthermore, the instrument incorporates Parasocial Interaction Theory (Horton & Wohl, 1956), which explains how media consumers construct one-sided, affective attachments with televised and digital media personas. In synchronous live broadcasting, the legacy concept of parasocial interaction is transformed into a hybrid “synchronous parasocial relationship.” The streamer looks directly into the webcam, reads chat entries aloud, and addresses viewers by their usernames, blurring the boundary between an imaginary relationship and an authentic social exchange. The inventory captures this psychological illusion, measuring how perceived interpersonal warmth converts psychological attachment into tangible monetary patronage.

Validity

The psychometric validity of the Determinants of Donation Intention in Live Streaming–Model Inventory was rigorously established through standardized structural equation modeling (SEM) protocols administered to experienced live-streaming consumers across diverse digital platforms (e.g., Twitch, YouTube Live, TikTok Live, AfreecaTV).

Content and Translation Validity

The inventory’s constituent items were synthesized from validated measures in the information systems, human-computer interaction, and consumer psychology literature. Items were systematically contextualized to capture the idiosyncratic properties of synchronous live broadcasting. To ensure cross-cultural semantic equivalence, a rigorous translation and back-translation procedure was executed by bilingual academic experts in accordance with standard methodological frameworks (Schaffer & Riordan, 2003). A pre-test pilot administration was conducted with experienced live-streaming users to evaluate syntactic clarity, readability, and face validity before large-scale empirical deployment.

Convergent Validity

Convergent validity evaluates the extent to which measurement items adequately load on their designated theoretical construct. Confirmatory factor analysis demonstrated excellent convergent validity across all latent variables:

• Factor Loadings: All standardized item factor loadings exceeded the recommended threshold of .70 (ranging from .72 to .91) and were statistically significant at p < .001, confirming that each indicator accounts for substantial variance in its underlying latent construct.
• Average Variance Extracted (AVE): The Average Variance Extracted values for all constructs surpassed the conventional benchmark of .50. Furthermore, the square roots of the AVE values for each construct ranged between .829 and .892, indicating that the latent constructs capture substantially more variance from their assigned indicators than from extraneous measurement error.
• Composite Reliability (CR): All constructs achieved Composite Reliability scores ranging from .877 to .929, well above the .70 criterion, confirming robust internal convergence.

Discriminant Validity

Discriminant validity assesses whether the empirical indicators differentiate between conceptually distinct constructs. This was established through three stringent empirical standards:

• Fornell-Larcker Criterion: In compliance with the Fornell & Larcker (1981) criterion, the square root of the AVE for every single construct (.829 to .892) was consistently greater than the highest bivariate inter-construct correlation involving that latent variable.
• Heterotrait-Monotrait Ratio (HTMT): All calculated HTMT ratios fell safely below the conservative threshold of .85 recommended by Henseler et al. (2015), providing rigorous proof that the nine dimensions represent empirically discrete behavioral phenomena.
• Cross-Loading Analysis: In accordance with Hair et al. (2021), every individual measurement item loaded significantly higher on its assigned latent factor than on any cross-construct, verifying that no construct cross-contamination obscured the measurement model.

Predictive and Nomological Validity

Nomological validity was demonstrated through partial least squares structural equation modeling (PLS-SEM). Both dedication factors (Parasocial Interaction, Perceived Value, Trust, and Affective Commitment) and constraint factors (Switching Costs) accounted for significant explained variance (R²) in Continuance Intention and subsequent Donation Intention, confirming the predictive utility of the dedication-constraint architecture in interactive media monetization.

Reliability

The inventory demonstrates high empirical reliability and measurement precision across all subscales, fulfilling international psychometric benchmarks for behavioral and psychological survey instruments.

Internal Consistency

Internal consistency across the nine subscales was systematically confirmed through two distinct statistical indices: Cronbach’s alpha (α) and Composite Reliability (CR). All estimated Cronbach’s alpha values exceeded the standard threshold of .70 (Hair et al., 2021), reflecting homogeneous indicator pools within each operationalized construct. Composite Reliability values, which avoid the assumption of tau-equivalence and provide a more precise calculation of scale reliability, ranged from .877 to .929 across the nine constructs.

Factor Analysis

The structural dimensionality of the inventory was tested through rigorous factor analytic methods, including testing for common method variance and assessing global structural model fit.

Common Method Bias Evaluation

Because self-report electronic surveys carry the potential for common method bias (CMB), researchers conducted Harman’s single-factor test as an initial diagnostic (Podsakoff et al., 2003, 2012). An exploratory unrotated factor analysis incorporating all 31 items revealed that the first extracted factor accounted for only 41.15% of the total cumulative variance. Because this value falls well below the conservative empirical threshold of 50%, common method bias is not a pervasive threat to the validity of the findings.

Confirmatory Factor Analysis and Structural Modeling

Confirmatory factor analysis (CFA) using covariance-based and partial least squares structural equation modeling indicated excellent overall model fit to the observed data. The hypothesized nine-factor structure yielded fit indices that satisfied standard econometric and psychometric criteria:

• Chi-Square / Degrees of Freedom (χ²/df): Observed values ranged between 1.50 and 2.40, falling safely within the recommended benchmark of < 3.0.
• Comparative Fit Index (CFI) & Tucker-Lewis Index (TLI): Both CFI and TLI values exceeded the .95 cutoff, demonstrating strong confirmatory baseline fit.
• Root Mean Square Error of Approximation (RMSEA): RMSEA was observed below .06 (with 90% confidence intervals between .045 and .065), demonstrating low residual approximation error.
• Standardized Root Mean Square Residual (SRMR): SRMR remained below the .08 ceiling, confirming that the discrepancy between the observed and implied correlation matrices was negligible.

Collectively, the factor analysis substantiates that the nine constructs operate as distinct structural components, providing empirical support for the dedication-constraint framework in interactive streaming environments.

Instrument / Measurement Tool

The operational specifications of the measurement tool are summarized below:

• Test Type: Multi-dimensional self-report behavioral questionnaire / Psychometric inventory.
• Format: Electronic questionnaire (administered via web-based survey interfaces, computer terminals, or mobile platforms).
• Number of Items: 31 items organized across 9 latent subscales.
• Authentic Response Scale: 7-point Likert scale ranging from 1 (Strongly disagree) to 7 (Strongly agree).
• Subscale Item Allocation:
  • Social Interaction: Items 1–4
  • Parasocial Interaction: Items 5–8
  • Perceived Value: Items 9–12
  • Trust in the Streamer: Items 13–16
  • Affective Commitment: Items 17–20
  • Continuance Intention: Items 21–23
  • Attractive Alternatives: Items 24–25
  • Switching Costs: Items 26–28
  • Donation Intention: Items 29–31
• Scoring Rules: Subscale scores are calculated by averaging the individual item ratings within each designated construct, yielding continuous subscale means ranging from 1.00 to 7.00. No items require reverse coding. Higher scores in each subscale indicate higher levels of that specific construct (e.g., higher perceived value, deeper affective commitment, or greater donation intention).
• Target Population: Human adults (aged 18 and older) who actively participate in or observe synchronous live streaming channels.

Permissions & Fee and Test Year

The Determinants of Donation Intention in Live Streaming–Model Inventory was formally established and published in 2023 in peer-reviewed academic literature. The authors—Shih-Wei Chou and Trieu Nguyen—have made the measurement model accessible for non-commercial, educational, scientific, and scholarly research purposes without assessment fees. Researchers and institutional investigators may utilize and adapt the scale items provided that appropriate bibliographic credit and formal academic citations are granted to the original 2023 publication in Computers in Human Behavior. Commercial utilization, monetization, or proprietary integration into commercial software tools requires prior authorization from the copyright holders.

References

• Bendapudi, N., & Berry, L. L. (1997). Customers’ motivations to maintain relationships with service providers. Journal of Retailing, 73(1), 15–37. https://doi.org/10.1016/S0022-4359(97)90013-0
• Blau, P. M. (1964). Exchange and power in social life. John Wiley & Sons.
• Chou, S.-W., & Nguyen, T. (2023). Understanding donation intention in live streaming: A dedication-constraint approach. Computers in Human Behavior, 144, Article 107757. https://doi.org/10.1016/j.chb.2023.107757
• Fornell, C., & Larcker, D. F. (1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 18(1), 39–50. https://doi.org/10.1177/002224378101800104
• Hair, J. F., Hult, G. T. M., Ringle, C. M., & Sarstedt, M. (2021). A primer on partial least squares structural equation modeling (PLS-SEM) (3rd ed.). SAGE Publications.
• Henseler, J., Ringle, C. M., & Sarstedt, M. (2015). A new criterion for assessing discriminant validity in variance-based structural equation modeling. Journal of the Academy of Marketing Science, 43(1), 115–135. https://doi.org/10.1007/s11747-014-0403-8
• Horton, D., & Wohl, R. R. (1956). Mass communication and para-social interaction: Observations on intimacy at a distance. Psychiatry, 19(3), 215–229. https://doi.org/10.1080/00332747.1956.11023049
• Podsakoff, P. M., MacKenzie, S. B., Lee, J.-Y., & Podsakoff, N. P. (2003). Common method biases in behavioral research: A critical review of the literature and recommended remedies. Journal of Applied Psychology, 88(5), 879–903. https://doi.org/10.1037/0021-9010.88.5.879
• Podsakoff, P. M., MacKenzie, S. B., & Podsakoff, N. P. (2012). Sources of method bias in social science research and recommendations on how to control it. Annual Review of Psychology, 63(1), 539–569. https://doi.org/10.1146/annurev-psych-120710-100452
• Schaffer, B. S., & Riordan, C. M. (2003). A review of cross-cultural methodologies for organizational research: A best-practices approach. Organizational Research Methods, 6(2), 169–215. https://doi.org/10.1177/1094428103251542
• Stanley, S. M., & Markman, H. J. (1992). Assessing commitment in personal relationships. Journal of Marriage and the Family, 54(3), 595–608. https://doi.org/10.2307/353245

Items of the Scale