Exploring Job Satisfaction in the Platform Economy: A Multidimensional Study of Mobile Digital Platform Workers in Chile

Abstract

This study examines job satisfaction among mobile digital platform workers in Chile, addressing the gap in understanding the interplay between autonomy, social support, and technology in the platform economy. It explores how these factors shape job satisfaction through the following research question: How do autonomy, social support, and technological factors influence job satisfaction in the platform economy? Using a quantitative methodology, 398 platform workers from three Chilean regions participated in a survey, with data analyzed via logistic regression models to evaluate the impact of sociodemographic and psychological variables. The findings highlight that autonomy and social support significantly enhance job satisfaction, while technological usability and transformation are pivotal for positive work experiences. The study concludes that fostering worker autonomy, robust support systems, and user-friendly technologies is critical for improving job satisfaction in the gig economy. These insights contribute to the academic literature and inform strategies for policymakers and platform operators, emphasizing the need for tailored approaches to enhance worker well-being and organizational outcomes.

1. Introduction

This article critically analyzes job satisfaction among mobile digital platform workers in Chile, addressing the need to explore the dynamic interplay of autonomy, social support, and technology in shaping work experiences within the platform economy. By integrating sociological, psychological, and technological perspectives, this study aims to uncover the unique dimensions of worker satisfaction in a rapidly evolving labor context. The research responds to calls for a deeper understanding of how gig work influences job satisfaction, offering a comprehensive examination grounded in both global and local contexts.

The platform economy [1,2,3] has significantly transformed traditional labor markets, creating opportunities for flexible work while simultaneously introducing new challenges [2,3,4]. In Chile, where platform work has become increasingly prevalent, there is limited empirical research examining the specific factors that affect worker satisfaction. This gap becomes even more critical considering the global shift towards digitalized work and the diverse conditions under which platform workers operate. This study aims to fill this void by addressing the nuanced dynamics of job satisfaction within the platform economy, focusing on a combination of intrinsic and extrinsic factors. What makes this research particularly novel is its context-specific approach, which highlights how global trends adapt to the local labor market and sociocultural realities of Chile.

Grounded in theories of job satisfaction, this study examines the roles of autonomy, social support, and the relationship with technology in predicting worker satisfaction. Methodologically, it employs a quantitative approach, using questionnaires to gather data from platform workers in three major regions of Chile: Metropolitan, Valparaíso, and Biobío. The data were analyzed using logistic regression models to explore the influence of sociodemographic and psychological variables on both extrinsic and intrinsic satisfaction levels. This methodology allows for a detailed examination of how key variables interact to shape worker experiences in the gig economy.

This research contributes to the academic literature by advancing theoretical frameworks for understanding job satisfaction in the platform economy, particularly within a Latin American context. It highlights the critical roles of autonomy and social support while addressing the often-overlooked impact of technological factors on satisfaction. In addition, the findings have practical implications for policymakers and platform operators. Enhancing worker satisfaction can lead to improved retention, productivity, and overall well-being, making it a priority for both regulatory and organizational strategies. This study also underscores the need for tailored approaches that account for the diverse needs of platform workers, given the varying sociodemographic factors influencing their experiences.

The structure of this paper unfolds as follows. The literature review situates the study within the broader discourse on the platform economy, labor markets, and job satisfaction, with a specific focus on the Chilean context. The methodology describes the data collection process, variables, and analytical tools used to evaluate the research questions. The results present key findings, exploring how autonomy, social support, and technological engagement influence satisfaction. The discussion contextualizes these findings within the existing literature, emphasizing their theoretical and practical implications. Finally, the conclusion highlights this study’s contributions and outlines directions for future research, particularly the importance of longitudinal and comparative analyses to further elucidate job satisfaction in platform economies.

2. Literature Review

2.1. Platform Economy in the World

Platform economies, as defined by the academic literature, refer to a technologically enabled socio-economic system where goods, services, or information are exchanged through digital platforms [1,2,3]. These platforms act as intermediaries connecting individuals or businesses for transactions, facilitating exchanges, shaping the institutional framework, and driving economic activities [2,3,4,5,6,7]. The distinctive feature of the platform economy lies in its emphasis on technologically mediated exchanges, highlighting the importance of optimizing resource use and sharing underutilized assets through models such as the sharing economy, the access economy, and collaborative consumption [1,3,8]. Furthermore, the platform economy has broader implications, influencing labor markets, regulatory frameworks, consumer behavior, and sustainability goals. By impacting consumption patterns, it contributes to sustainable development efforts [9,10,11,12].

Thus, the platform economy emerges as a dynamic ecosystem driven by digital platforms that not only facilitate transactions but also promote resource sharing and significantly influence economic and social interactions across various sectors [1,3]. Its impact extends beyond exchanges, reshaping regulatory frameworks, labor practices, and sustainability efforts to redefine contemporary economic activities and social structures.

The platform economy has established itself as a significant global influence, particularly in China, which hosts one of the world’s largest online platform economies [13]. This growth is fueled by digital technologies such as the internet, cloud computing, big data, and the Internet of Things, transforming traditional business models into platform-based structures [14]. Organizations within this economy have innovatively addressed challenges by redefining interactions between regulators and workers and introducing new business models, including platform cooperatives and guild-type organizations that collaborate with unions to meet platform workers’ specific needs [15]. Additionally, the influence of the platform economy extends to labor dynamics, affecting labor laws, job security, and the nature of employment, further demonstrating its transformative role [16,17].

2.2. Context of the Platform Economy in Latin America

Platform economies are becoming increasingly prominent in Latin America, standing out in various sectors and experiencing notable growth in the digital collaborative economy [18], although the informal sector remains fundamental [19]. This growth is accompanied by a surge in digital entrepreneurship, which strengthens an entrepreneurial ecosystem with prominent digital native companies in the region [20]. However, Latin American companies face challenges in fully capitalizing on the benefits of digitalization, which can lead to a decline in performance if not adequately adapted [21]. Digital transformation is gaining recognition, with an increase in the adoption of digital health technologies [22] and the growing use of digital platforms for work, which affects labor markets and poses regulatory challenges [23]. In education, technological initiatives such as online learning platforms have been implemented to promote autonomous learning [24,25]. During the COVID-19 pandemic, the adoption of digital solutions in sectors such as health and education intensified, highlighting the importance of telemedicine technologies, especially in rural areas [26,27,28]. While the potential of the platform economy for the development of nations in Latin America continues to be explored, policy in this regard is limited by the structural characteristics of the region’s economies [29].

2.3. Platform Economy in Chile

The platform economy in Chile has deeply integrated into the country’s socioeconomic fabric, reflecting both the adaptation to global work trends and unique local challenges. This integration is characterized by a diversity of perspectives ranging from labor conditions to resistance strategies, through to the implications of regulation and policy.

Research by Arriagada et al. [30] highlights the complexities of labor conditions within Chile’s platform economy, emphasizing precarity and the need for labor protections. This analysis is complemented by the work of Asenjo Cruz, Coddou Mc Manus, and Dhir [31], who explore the transformation of the world of work in Santiago, focusing on delivery platform workers and the influence of the socioeconomic context on the work experience.

Durri [32] provides broader context on the regulation of the platform economy, arguing the need to adapt regulations to new forms of work. He suggests that this regulatory framework would be crucial to address emerging challenges and ensure the protection of workers in an evolving labor market.

In terms of labor, Fielbaum and Tirachini [33] focus on the labor market in the shared economy, specifically on ride service drivers. Their study shows the preferences, concerns, and satisfaction levels of drivers, highlighting the importance of flexibility in working hours and the need for transparency in payment calculations.

García and Azócar [34] discuss Chile’s legislative solution for platform work, analyzing whether the proposed regulations are suitable for the dilemmas presented by this type of work. On the other hand, Gutierrez Crocco and Atzeni [35] examine the effects of the pandemic on platform economy messengers in Argentina and Chile, focusing on precarity, algorithmic control, and mobilization strategies. Also regarding employment conditions, Jirón et al. [36] observe the reality of digital mobile workers in Chile, highlighting the challenges of an unregulated work environment.

More critically, Leyton et al. [37] comment on the new regulation of platform work in Chile, suggesting that it currently represents a missed opportunity to address the problems associated with working conditions in the sector. Meirosu [38] examines informality and the platform economy in Chile, offering a perspective on the integration of technological advances and the growth of the digital economy.

Morales et al. [39,40,41] have explored organization and collective action among platform workers in Chile, Peru, and Spain, highlighting resistance strategies and collective action against the dynamics of precarization in the platform work context.

Tironi and Albornoz [42] have analyzed surveillance and frictions in platform urbanism, focusing on delivery workers in Santiago de Chile, while Johannsen and Gonzalez [43] discuss the impact of the platform economy on economic dependency in Chile, pointing out the need to address potential threats from new platform companies in the market and potential harm to various stakeholders.

These studies collectively underline the necessity of framing Chile’s platform economy within its broader economic and sociocultural context. For instance, the country’s longstanding economic policies, cultural perceptions of labor, and unique urban challenges shape the specific dynamics of platform work. By situating these analyses within the interplay of Chile’s neoliberal economic framework, persistent inequality, and the resilience of its workforce, readers can gain a deeper understanding of how global platform trends are locally adapted. Furthermore, examining regional disparities, historical labor movements, and the sociopolitical responses to emerging economic models enriches this exploration. The convergence of analyses on labor conditions, regulation, and resistance strategies underlines the complexity of managing and regulating the platform economy in a way that balances demand for economic flexibility with the protection and well-being of workers.

2.4. Job Satisfaction in the Platform Economy

The platform economy has revolutionized traditional employment relationships, shifting towards short-term, project-based work facilitated by digital platforms [30]. This transformation has reshaped job satisfaction, introducing a tension between objective conditions, such as material precarity and entrepreneurial opportunities, and subjective aspects, such as autonomy and fulfillment [20,44,45,46,47,48,49]. These dynamics challenge the traditional factors used to conceptualize and measure job satisfaction, particularly in areas such as leadership, teamwork, and systems of compensation

Job satisfaction remains a critical aspect of worker well-being and productivity [50,51,52,53,54]. It significantly influences organizational outcomes such as commitment, motivation, and quality of performance while reducing burnout and turnover [55,56,57,58,59,60,61,62,63,64,65,66]. Traditional models of job satisfaction often focus on aspects like physical workspace, recognition, and leadership [67,68,69,70]. However, platform work introduces unique characteristics—such as constant connectivity, virtualization, and flexibility—that demand a re-evaluation of these models [70,71].

Innovative models of job satisfaction have emerged to address these changing conditions. The sustainable job satisfaction model emphasizes work–life balance, professional development, and corporate social responsibility, acknowledging the broader social and individual impacts of work [71,72,73]. Meanwhile, the self-determined job satisfaction model centers on satisfying intrinsic psychological needs, such as autonomy, competence, and interpersonal relationships [74,75]. The dynamic job satisfaction model proposed by Zacher and Rudolph [76] adds a processual perspective, integrating situational and dispositional factors over time [77]. These frameworks provide valuable insights into understanding satisfaction in the platform economy.

Despite these advancements, challenges persist in explaining job satisfaction within digital platform work. For example, algorithmic management and remote supervision have radically transformed traditional leadership practices [46]. The emotional intelligence model of job satisfaction, which emphasizes stress management and interpersonal relationships, has proven particularly relevant in addressing the psychological impacts of these changes [78,79].

Research also highlights the dual role of autonomy and flexibility in platform work. While these elements contribute positively to job satisfaction, the accompanying lack of job security, unpredictable work hours, and minimal social benefits undermine these gains [44,80]. Constructs like social support and technology-mediated relationships have emerged as critical dimensions, offering new ways to conceptualize job satisfaction in this context [81,82,83,84,85,86,87].

Efforts to adapt job satisfaction models to the digital age include integrating locally validated unifactorial global indices [81] with classic definitions, such as those proposed by Baxi and Atre [82], which focus on workers’ feelings and beliefs about their job. These approaches aim to bridge traditional and emerging perspectives on satisfaction by addressing the interplay of extrinsic and intrinsic conditions [83].

Moreover, platform economies demand new metrics for measuring satisfaction. For example, balancing flexibility with fair compensation and adequate social support remains a central challenge. Models such as the sustainable job satisfaction framework emphasize aligning professional aspirations with corporate social responsibility and fostering environments that mitigate financial precarity [71,72,84].

The implications of these findings extend beyond individual workers to organizational practices. High levels of job satisfaction correlate with greater citizenship behavior, loyalty, and productivity [55,62,63]. Recognizing and addressing the unique challenges of platform work can drive significant improvements in worker well-being and organizational outcomes [59,66].

To summarize, platform economies have disrupted traditional paradigms of work, necessitating innovative approaches to understanding job satisfaction. By integrating insights from traditional and emerging models, researchers can provide a more comprehensive framework that captures the complexity of satisfaction in the digital age [80,82,85,86,87,88,89,90,91,92,93].

2.5. Social Support, Autonomy, and Relationship with Technology

The proposed model assesses the variables of social support, autonomy, and the relationship with technology to explain job satisfaction in the platform economy. These elements were selected due to their critical roles in addressing the unique challenges faced by platform workers.

Social support refers to the perception of care, esteem, and belonging to a network of mutual assistance. Numerous studies have demonstrated its predictive role in job satisfaction across various contexts, including remote work, temporary jobs, and immigrant populations [94,95,96,97,98,99,100,101]. In the platform economy, where organizational supervision is often absent, alternative sources of support emerge—such as peer networks, family, and even client relationships. These networks can alleviate stress and reduce the risk of depersonalization and alienation [90,98]. For instance, the quality of client interactions may either empower workers or contribute to stress and dissatisfaction, highlighting the dual nature of these relationships [99,100,101,102,103,104,105,106,107,108,109].

In this study, social support is conceptualized as a mitigating factor against the stress, isolation, and instability inherent to platform work. Workers who perceive robust social support networks report higher satisfaction levels, as these networks provide both emotional and practical resources necessary for navigating precarious labor conditions.

Autonomy is another crucial dimension of job satisfaction, particularly valued by platform workers due to its implications for control over time, space, and task execution. The platform economy often requires workers to exercise self-determination in their roles, such as managing schedules, navigating resources, and making decisions without direct supervision [89,90,91,92,93,94,95,96,97,98,99,100,101,102,103,104,105,106,107,108,109,110,111,112,113,114,115]. While autonomy is frequently associated with increased satisfaction, its application within the platform economy is complex. For example, some studies describe autonomy in platform work as paradoxical: it offers flexibility and freedom while simultaneously imposing technological controls and organizational constraints [116,117,118].

In this study, autonomy is examined through its interplay with platform architecture and worker agency. Workers who perceive greater autonomy—such as flexibility in scheduling or decision-making—often report higher levels of satisfaction, particularly when supported by transparent processes and tools provided by platforms. These findings align with previous research indicating that environments promoting self-determination enhance worker satisfaction [114,115].

The relationship between workers and the technological platforms they use is a defining characteristic of the platform economy. This interaction encompasses usability, adaptability, and the extent to which technology facilitates or hinders task execution. Prior research [116,117,118,119,120,121,122] has emphasized that intuitive and efficient digital tools enhance workers’ sense of empowerment and productivity, which, in turn, positively influences job satisfaction [123,124,125]. Conversely, poorly designed platforms may lead to dissatisfaction by increasing cognitive load or reducing operational efficiency [126,127,128,129,130].

In this study, the relationship with technology is examined through multiple dimensions, including usability, transformation, and autonomy. For example, user-friendly interfaces and access to critical information can foster satisfaction by reducing friction in task execution. However, this study also acknowledges that the technological architecture of platforms can act as a double-edged sword, simultaneously enabling autonomy while embedding control mechanisms that may diminish the perceived freedom of workers [116,117,118,119,120,121,122].

By integrating social support, autonomy, and the relationship with technology, this study provides a multidimensional perspective on job satisfaction within the platform economy. Each dimension contributes uniquely to the overall satisfaction of platform workers, yet they are inter-related. For instance, robust social support may amplify the positive effects of autonomy, while effective technological tools can enhance the benefits of both autonomy and social support. The literature findings [126,127,128,129] emphasize the need for platform operators to adopt a holistic approach, addressing these interconnected factors to create more satisfying work environments.

This integrated approach ensures alignment with the study’s objectives, as it directly connects these dimensions to the unique conditions of platform work. It also addresses the limitations of traditional job satisfaction models by offering a nuanced framework tailored to the digital labor context.

2.6. Sociodemographic Variables and Their Influence on Job Satisfaction

Job satisfaction is a crucial aspect for employee well-being and organizational success. It has been found that various sociodemographic variables influence job satisfaction among different professional groups. The research conducted by Carrillo-García et al. [131] highlighted the impact of gender and age on job satisfaction among healthcare workers, emphasizing significant associations between these variables and the level of job satisfaction. Similarly, a study by Ebling and Carlotto [132] focused on burnout syndrome among health professionals and identified sociodemographic factors as key influencers of job satisfaction, using tools such as the Maslach Burnout Inventory and the Job Satisfaction Questionnaire to assess these relationships.

In the context of nursing care, Uchmanowicz et al. [133] conducted a study on the perception of implicit rationing of nursing care and found that sociodemographic variables, nurses’ evaluations of patient care quality, and overall job satisfaction were interconnected. Additionally, research by Chang et al. [134] delved into the associative stigma among mental health professionals, revealing that sociodemographic factors played a role in levels of job satisfaction, with high associative stigma linked to lower job satisfaction scores.

Educators also experience the impact of sociodemographic variables on job satisfaction, as evidenced by a study by Dicks et al. [135] on teachers in Germany during the SARS-CoV-2 pandemic. The research highlighted work-related variables as significant predictors of job satisfaction among teachers. Likewise, Maloney et al. [136] explored job satisfaction among pain medicine specialists in the United States, noting that factors such as age, gender, specialty of practice, salary, and workload could influence physicians’ job satisfaction.

In the health sector, Chen et al. [137] analyzed the relationship between effort–reward imbalance and job satisfaction among family doctors in China, identifying age, education, job rank, type of institution, years of work, and income as influential factors. Additionally, Alqarni et al. [138] studied stress, burnout, and job satisfaction among mental health professionals in Jeddah, Saudi Arabia, emphasizing the role of sociodemographic variables in the well-being of these professionals.

On the other hand, Dias et al. [139] explored the motivation and job satisfaction of individuals working with cancer, highlighting the complex interaction of motivational factors in job satisfaction. Akuffo et al. [140] investigated job satisfaction among opticians in Ghana, using logistic regression analysis to understand the association between sociodemographic characteristics and levels of job satisfaction. These studies collectively underscore the importance of considering sociodemographic variables to understand and improve job satisfaction across various professions.

Furthermore, Duah and Kofi [141] examined job satisfaction in organizations in Ghana, identifying multiple variables such as association with colleagues, recognition at work, salaries, workload, work environment, and working conditions as influencers of job satisfaction. Similarly, Ntopi et al. [142] focused on health surveillance assistants in Malawi, revealing the role of sociodemographic variables in shaping role stressors and job satisfaction among this group of healthcare workers.

Organizational and leadership factors also play a significant role in influencing job satisfaction. Karlita et al. [143] highlighted the effect of job characteristics and work–life balance on job satisfaction, emphasizing the positive impact of these factors. Additionally, Sánchez-Sellero and Sánchez-Sellero [144] analyzed job satisfaction in Spain, incorporating various organizational, work, and sociodemographic variables to understand the dynamics of job satisfaction in the context of economic crises.

The synthesis of these studies highlights the complex relationship between sociodemographic variables and job satisfaction in various professional settings. Understanding and addressing these factors is essential for promoting employee well-being, enhancing job satisfaction, and ultimately improving organizational performance.

3. Results

The following section is mainly focused on the methods used in the exploratory data analysis and regression model fitting sections ahead.

3.1. Correlation Analysis

To test correlation among variables, Pearson correlation coefficient is calculated according to (1):

$${\rho }_{X,Y}={\displaystyle \frac{cov\left(X,Y\right)}{{\sigma }_X{\sigma }_Y}}$$

(1)

This tests for linear dependence between two continuous variables. The value $\rho$ moves between −1 and 1, with −1 meaning a perfect inverse relationship (if one variable increases, the other decreases), and +1 meaning a perfect direct relationship (both variables increase or decrease in the same direction) [1,2].

3.2. ANOVA Test

To test if there exist differences in a variable with more than two groups, the ANOVA test is conducted, which fixes problems with multiple T-tests over the same variable (increasing p-value for multiple measurements). The hypothesis for this test is stated ahead, considering that for a variable there exist $m$ categories.

$$H_0:{\mu }_1={\mu }_2={\mu }_3=\cdots ={\mu }_m$$

$$H_0:\exists i:{\mu }_i\ne {\mu }_j\forall i,j\in \{1,\dots ,m\}$$

The null hypothesis states that the means for all groups are the same, while the alternative hypothesis states that there exists at least one group with significant differences among all groups tested.

3.3. F-Test for Regression Model (Overall Significance)

The F-test checks for overall significance in a regression model. In a similar way to the ANOVA test, it checks if any of the coefficients of the model are significant (at least one). The hypothesis for the test is shown ahead, considering the model has $p$ variables.

$$H_0:{\beta }_i=0\forall i\in \{1,\dots ,p\}$$

$$H_0:\exists i\in \left\{1,\dots ,p\right\}:{\beta }_i\ne 0$$

If the null hypothesis is rejected, then the model can be considered significant, and other analyses and validations can be applied to the model fitted.

3.4. Individual Significance of Parameters in Regression Analysis

The F-test is a tool that allows for testing overall significance in a regression analysis. If the F-test is rejected and there exists at least one significant parameter, a test is conducted for each one of the parameters of the model. The individual hypothesis is shown ahead.

$$H_0:{\beta }_i=0$$

$$H_1:{\beta }_i\ne 0$$

If the hypothesis is not rejected, the parameter $i$ (or variable $i$) is not significant in explaining the response variable in our model.

3.5. R-Squared

To test the performance of a regression model, R-squared (also known as the coefficient of determination) and adjusted R-squared are computed considering the error terms of the model. Let $S{S}_{res}$ be the sum of squares of residuals computed by (2):

$$S{S}_{res}=\sum _{i=1}^n{\left(y_i-f\left(\beta ,x\right)\right)}^2=\sum _{i=1}^n{e_i}^2$$

(2)

Let $S{S}_{tot}$ be the total sum of squares computed according to (3):

$$S{S}_{tot}=\sum _{i=1}^n{\left(y_i-\overline{y}\right)}^2$$

(3)

We can compute the coefficient of determination with (4):

$$R^2=1-{\displaystyle \frac{S{S}_{res}}{S{S}_{tot}}}$$

(4)

This corresponds to the percentage of variance or variability explained with our model from the response variable. The range of movement of $R^2$ is between 0 and 1, with 0 meaning the worst performance in a regression model, and 1 a perfect fit for the model (100% of variance explained through linear regression).

For multiple regression models, it is common to analyze the adjusted $R^2$, which penalizes the addition of variables into the model and avoids increasing $R^2$ artificially. Adjusted $R^2$ is computed as follows in (5):

$$\overline{R^2}=1-\left(1-R^2\right){\displaystyle \frac{n-1}{n-p-1}}$$

(5)

3.6. Variance Inflation Factors

Consider a model in the form shown in (6). If the coefficients ${\beta }_i$ are computed through the ordinary least squares (OLS) method, then the correlation and collinearity among variables should be addressed.

$$y_i={\beta }_0+{\beta }_1{x}_1+{\beta }_2{x}_2+\cdots +{\beta }_p{x}_p+e_i$$

(6)

The collinearity (mentioned before) must be addressed while applying the OLS method because the solution of the equation that gives the values of $\beta$ can be expressed according to (7) in matrix form:

$$\widehat{\beta }={\left(X^{T}X\right)}^{-1}{X}^{T}Y$$

(7)

If some variables $x_i$ and $x_j$ (with $i\ne j$) are perfectly correlated (one can be expressed as a transformation of the other), then the equation shown in (7) has no solution, and the coefficients $\beta$ cannot be computed. For testing this potential issue, the Variance Inflation Factors (VIFs) are computed according to (8):

$$\mathrm{VI}{\mathrm{F}}_i={\displaystyle \frac{1}{1-R_i^2}}$$

(8)

If some of the $\mathrm{VI}{\mathrm{F}}_i$ values are above 10, we say that our model has issues with collinearity among explanatory variables, and the coefficients $\beta$ cannot be interpreted, so a reformulation should be addressed to fix this issue (adding more variables or changing the variables used as explanatory features).

The previous results hold for numerical values (either continuous or discrete). For cases in which the variables are categorical, the GVIF_i is applied, which corresponds to a generalized collinearity diagnostic for cases with categorical variables, multi-level responses, or polynomial terms in single variables.

3.7. Durbin–Watson Test for Autocorrelation

Once a regression model is fitted, one of the assumptions made over the sample of data is that it is $iid$, which means independent and identically distributed; to test the independence of the data, the error terms (residuals) are tested through the Durbin–Watson test, which addresses potential issues with autocorrelation in the residuals of a model.

Consider $e_t$ as the residual associated with the observation at time $t$. The statistic for the test is shown in (9):

$$d={\displaystyle \frac{\sum _{t=1}^T{(e}_t-e_{t-1}{)}^2}{\sum _{t=1}^T{e_t}^2}}\in [\mathrm{0,4}]$$

(9)

In general terms, $d~2(1-R^2)$, and the following hypothesis is stated:

$$H_0:\rho =0$$

$$H_1:\rho \ne 0$$

If the null hypothesis is not rejected, it can be assumed that there is no presence of serial autocorrelation within the error terms of the model.

3.8. Breusch–Pagan Test for Equal Variance

As stated before, in a regression framework, the sample is assumed to be $iid$. In Section 3.7, it is shown how to test for independence on the regression model, and the other half (identically distributed) is, in part, addressed with the Breusch–Pagan test, which measures if the variance of the residuals is constant in the model. In the case that we reject the null hypothesis of constant variance, it can be stated that the model lacks homoscedasticity and some transformations should be applied to fix it:

$$H_0:\sigma =cte$$

$$H_1:\sigma ={\sigma }_t$$

3.9. Exploratory Data Analysis

The dataset consists of 398 observations and 18 variables organized as

Table 1. Definition and typology of variables.

Variable	Type	Levels
City	Categorical	RM Valparaíso Biobío
Age	Categorical	Between 18 and 24 years old Between 24 and 34 years old 35 years old or older
Genre	Categorical	Male Female Other
Labor old	Categorical	Six months or less Between six months and two years More than two years
Immigrant	Categorical	Yes No Other
Incomes	Numerical continuous
Extrinsic Satisfaction	Numerical continuous
Intrinsic Satisfaction	Numerical continuous
Overall Job Satisfaction	Categorical	Very unsatisfied Unsatisfied Neutral Satisfied Very satisfied
Technology—Usability	Numerical continuous
Technology—Interest	Numerical continuous
Technology—Transformation	Numerical continuous
Technology—Fear	Numerical continuous
Technology—Autonomy	Numerical continuous
Support	Numerical continuous
Users support	Categorical	Totally disagree Disagree Neutral Agree Totally agree
Family support	Categorical	Totally disagree Disagree Neutral Agree Totally agree

.

3.10. Review per Category

In terms of satisfaction levels (extrinsic and intrinsic), there are no significant differences when reviewing them per city. The ANOVA test does not reject the null hypothesis for both variables, showing equal means per city (as a category) (

Table 2. Satisfaction levels per city.

City	Observations	Extrinsic Satisfaction	Intrinsic Satisfaction	Conclusions
Biobío	48	27.0	26.7	No significant differences
RM	282	26.7	26.0
Valparaíso	68	26.5	26.1

).

Reviewing the data by age shows the same results as the previous analysis. The following table summarizes both variables (extrinsic and intrinsic) per categories of age (

Table 3. Summary by age categories (extrinsic and intrinsic).

Age	Observations	Extrinsic Satisfaction	Intrinsic Satisfaction	Conclusions
Between 18 and 24 years old	81	27.4	26.9	No significant differences
Between 25 and 34 years old	196	26.3	25.7
35 years old or older	121	26.8	26.3

); the ANOVA test does not reject the null hypothesis for both variables, so the means per group—category are equal.

Grouping the data by gender shows differences in the groups, particularly with the groups corresponding to each other. The difference in means is shown with the ANOVA test, which rejects the null hypothesis in both variables per category with p-values less than 10% (

Table 4. Summary by satisfaction levels.

Gender	Observations	Extrinsic Satisfaction	Intrinsic Satisfaction	Conclusions
Male	328	26.9	26.4	Significant differences
Female	65	26.3	25.1
Other	5	20.6	19.8

).

Satisfaction levels (intrinsic and extrinsic) grouped according to length of employment are shown in the following table. The ANOVA test does not reject the null hypothesis of equal means so there are no significant differences in the mean levels of satisfaction per category (

Table 5. Comparison of satisfaction levels by length of employment.

Length of Employment	Observations	Extrinsic Satisfaction	Intrinsic Satisfaction	Conclusions
Six months or less	140	27.0	26.1	No significant differences
Between six months and two years	66	27.3	26.7
More than two years	192	26.2	26.0

).

The following table shows the satisfaction levels for both variables in terms of the categories of the immigration variable. The ANOVA test shows no significant differences, so the means of groups are equal for intrinsic and extrinsic satisfaction (

Table 6. Satisfaction levels by categories of immigration variable.

Immigrant	Observations	Extrinsic satisfaction	Intrinsic Satisfaction	Conclusions
No	84	27.4	25.9	No significant differences
Other	5	24.4	24.8
Yes	309	26.5	26.2

).

The overall job satisfaction and both satisfaction levels are shown in

Table 7. Overall job satisfaction.

Overall Job Satisfaction	Observations	Extrinsic Satisfaction	Intrinsic Satisfaction	Conclusions
Very unsatisfied	20	15.6	15.9	Significant differences
Unsatisfied	49	21.1	20.1
Neutral	96	24.4	23.1
Satisfied	170	28.7	28.3
Very satisfied	63	32.7	32.9

. The ANOVA test rejects the null hypothesis, finding significant differences in the means of satisfaction per level of job satisfaction; the more satisfied with the job, the higher the satisfaction levels for both types of variables. Previous results are similar to what is observed reviewing the variables of user support and family support (

Table 8. Comparison of satisfaction levels by user support.

Users Support	Observations	Extrinsic Satisfaction	Intrinsic Satisfaction	Conclusions
Totally disagree	36	19.4	17.8	Significant differences
Disagree	60	25.6	24.7
Neutral	141	26.0	25.4
Agree	117	28.2	27.8
Totally agree	44	32.4	32.8

and

Table 9. Comparison of satisfaction levels by family support.

Family Support	Observations	Extrinsic Satisfaction	Intrinsic Satisfaction	Conclusions
Totally disagree	15	19.5	18.3	Significant differences
Disagree	23	25.1	24.1
Neutral	78	25.5	25.5
Agree	166	26.9	26.1
Totally agree	116	28.4	28

), with rejections of the null hypothesis in the ANOVA test.

3.11. Response Variable

Visualizing both satisfaction variables in histograms and QQ-Plots (Figure 1), it can be seen that both follow a symmetrical distribution similar to a normal distribution. The QQ-Plots show the adjustment to normal theoretical quantiles with the main differences in the extreme values of both distributions.

3.12. Correlation Analysis

Correlation analysis (Figure 2) is conducted considering the continuous variables in the dataset. For both variables of satisfaction, there are strong linear correlations, showing that usability, transformation, autonomy, and support factor have correlations above 0.5, with all of them positive, suggesting direct relations among dependent and independent variables. It is important to note that some of the variables that will be used as explanatory for the analysis have strong linear correlation coefficients; for instance, usability and interest show a linear relationship of 0.72, which could imply the existence of potential multicollinearity in the model with the data used to generate it.

3.13. Multiple Linear Regression Models

The results of the multiple linear regression model for extrinsic satisfaction are shown in

Table 10. Results of multiple linear regression model for extrinsic satisfaction.

Variable	Level	Transformation	Estimate	p-Value
Intercept			−2.16	0.47
Genre	Female		−0.13	0.80
	Male		−0.02	0.98
Age	<35 years old		−0.86	0.03 ***
Incomes			0.01	0.14
Overall Job Satisfaction	Unsatisfied		3.14	0.01 ***
	Neutral		5.41	~0.0 ***
	Satisfied		7.63	~0.0 ***
	Very Satisfied		10.14	~0.0 ***
Technology Usability		Log-transform	2.35	0.09 ***
Technology Interest			0.08	0.46
Technology Transformation			0.23	0.04 ***
Technology Fear			0.01	0.94
Technology Autonomy		Log-transform	1.46	0.09 ***
Support Factor			0.27	~0.0 ***
Users Support	Disagree		0.53	0.52
	Neutral		0.91	0.23
	Agree		0.46	0.59
	Totally Agree		0.73	0.48
Family Support	Disagree		−1.35	0.29
	Neutral		−1.96	0.08 ***
	Agree		−1.21	0.27
	Totally Agree		−1.95	0.08 ***

*** p < 0.01: The result is statistically significant at the 99% confidence level.

. The Fisher test for the overall significance of the regression model rejects the null hypothesis (p-value < 0.1), so some of the variables give significance for the model fitted, and the R2 is 68% with an adjusted R2 of 66%. In terms of significance per variable, age, overall job satisfaction (all levels), usability, transformation, autonomy, support factor, and family support (in two levels) show significance in terms of its parameters.

The test for normality of residuals rejects the null hypothesis, implying that they do not follow the distribution adjusted; this validation will be addressed later. The Durbin–Watson test does not reject null hypothesis, so there is no presence of autocorrelation in the error terms of the model, and the Breusch–Pagan test does not reject the null hypothesis of equal variance for errors, meaning there is presence of homoscedasticity in the model.

The results of the multicollinearity check are shown in

Table 11. Results of VIF and GVIF per variable in extrinsic satisfaction model.

Variable	Technology Usability	Technology Interest	Technology Transformation	Technology Fear	Technology Autonomy	Support Factor
VIF	2.76	2.37	2.26	1.31	1.99	2.24
Variable	Gender	Age	Overall Job Satisfaction	Users Support	Family Support
GVIF	1.07	1.03	1.13	1.14	1.09

with respective VIF values per variable. All the values are below 10 for the VIF and GVIF values per variable, showing no issues with collinearity among the model’s features.

The results of the multiple linear regression model for intrinsic satisfaction are shown in

Table 12. Results of multiple linear regression model for intrinsic satisfaction.

Variable	Level	Transformation	Estimate	p-Value
Intercept			2.11	0.48
Genre	Female		−0.56	0.27
	Male		−0.59	0.71
Age	<35 years old		−1.05	0.01 ***
Incomes			0.018	0.02 ***
Overall Job Satisfaction	Unsatisfied		1.95	0.05 ***
	Neutral		4.10	~0.0 ***
	Satisfied		7.19	~0.0 ***
	Very Satisfied		9.74	~0.0 ***
Technology Usability		Log-transform	0.12	0.93
Technology Interest			0.08	0.46
Technology Transformation			0.44	~0.0 ***
Technology Fear			−0.01	0.99
Technology Autonomy		Log-transform	0.41	0.63
Support Factor			0.29	~0.0 ***
User Support	Disagree		1.54	0.06 ***
	Neutral		1.99	0.01 ***
	Agree		1.50	0.07 ***
	Totally Agree		2.33	0.02 ***
Family Support	Disagree		−0.67	0.59
	Neutral		−0.70	0.53
	Agree		−0.76	0.48
	Totally Agree		−1.20	0.28

*** p < 0.01: The result is statistically significant at the 99% confidence level.

. The Fisher test rejects the null hypothesis (p-value < 0.1), meaning significance for the overall regression model. R2 is 71%, while the adjusted R2 equals 69%. In terms of significance of variables, age, income, overall job satisfaction (in all levels), transformation, support factor, and user support (in all levels) show significance with p-values less than 10%.

The Durbin–Watson test shows a p-value of 0.28, not rejecting the null hypothesis of autocorrelation, which equals zero in error terms, while the Breusch–Pagan test gives a p-value of 0.12 greater than 10%, not rejecting the null hypothesis of equal variances in residuals from the model (homoscedasticity).

The results of the multicollinearity check for the intrinsic satisfaction model are shown in

Table 13. Results of VIF and GVIF per variable in intrinsic satisfaction model.

Variable	Technology Usability	Technology Interest	Technology Transformation	Technology Fear	Technology Autonomy	Support Factor
VIF	3.41	2.37	2.26	1.31	1.99	2.24
Variable	Gender	Age	Overall Job Satisfaction	Users Support	Family Support
GVIF	1.07	1.03	1.14	1.15	1.09

with respective VIF values per variable. The values of VIF and GVIF per variable are below 10, meaning there are no issues with multicollinearity among the features used for the model.

3.14. Normality Check

To test normality assumptions for the model developed, the Shapiro–Wilk test was conducted over the residuals of both models. The model for extrinsic satisfaction rejects the null hypothesis of normally distributed residuals, while the model for intrinsic satisfaction does not reject the null hypothesis, confirming normality in the distribution of error terms. The following Figure 3 shows QQ-Plots for the residuals of both models (extrinsic and intrinsic). For both cases, the error terms fit in the curves corresponding to the theoretical normal quantiles (diagonal curves in charts). Moreover, the skewness of error terms is −0.51 and −0.24 for the extrinsic and intrinsic models, respectively, while kurtosis is about 3.3 for both models, being near to the theoretical moments of a normal distribution.

4. Discussion

The findings of this study provide a nuanced understanding of job satisfaction among mobile digital platform workers in Chile, emphasizing the multidimensional influences of autonomy, social support, and the relationship with technology. These results address the need for empirical evidence in the context of the platform economy, particularly within the Latin American labor market.

The significant role of job autonomy as a predictor of satisfaction aligns with established theories, demonstrating how the ability to control work schedules and processes enhances workers’ experiences. This study’s findings regarding social support emphasize its critical role in mitigating the precarious conditions often associated with platform work. The support provided by peers, family, and the platform itself contributes substantially to intrinsic satisfaction levels, reflecting the importance of fostering robust interpersonal and systemic networks. Technological usability and transformation also emerged as significant factors, revealing that intuitive and efficient technological tools enhance workers’ sense of empowerment and productivity. These findings corroborate the growing body of literature that highlights the centrality of user-friendly digital interfaces in improving job satisfaction.

Contrary to expectations, variables such as technological fear and gender did not exhibit significant associations with satisfaction levels. This lack of significance suggests that other contextual or individual factors may mediate these relationships. Additionally, the observed age-related differences in satisfaction highlight the nuanced ways in which demographic variables influence worker experiences. Younger workers reported lower satisfaction levels, which may reflect generational differences in expectations and adaptability within the platform economy.

This study’s results align closely with its objectives, providing empirical evidence to support the theoretical framework. By demonstrating the importance of autonomy, social support, and technology in shaping satisfaction, the findings contribute to the understanding of how platform work conditions impact well-being. The integration of these dimensions into a cohesive model advances the discourse on job satisfaction in digital labor markets, particularly within the unique sociocultural and economic context of Chile.

These insights underscore the need for targeted strategies to improve job satisfaction among platform workers. Policies that enhance autonomy, foster supportive environments, and prioritize the development of advanced technological tools are essential. Such measures not only benefit workers by improving their well-being but also support platform operators by enhancing productivity and retention. Furthermore, the findings highlight the potential for broader societal impacts, as improved satisfaction among platform workers can contribute to the sustainability and equity of the digital economy.

This study’s limitations, including its cross-sectional design and regional focus, necessitate caution in generalizing the findings. Future research should employ longitudinal designs to explore the dynamic nature of job satisfaction and examine how these factors evolve over time. Comparative studies across different cultural and economic contexts would also provide valuable insights, helping to refine and expand the proposed model. By addressing these gaps, future research can build on the foundational contributions of this study to advance the understanding of job satisfaction in the platform economy.

5. Conclusions

This study provides a comprehensive examination of job satisfaction among platform workers in Chile, combining sociological, psychological, and technological perspectives to uncover the multifaceted factors influencing worker satisfaction. By integrating these dimensions, the research offers a nuanced understanding of how autonomy, social support, and interactions with technology contribute to the overall experience of platform work. Among the findings, job autonomy emerged as a critical determinant, with workers who have greater control over their schedules reporting higher satisfaction levels. This finding aligns with the existing literature, underscoring the value of flexibility in the gig economy, where control over work processes is often a primary motivator.

Social support also plays a pivotal role in enhancing job satisfaction. Platform workers who benefit from peer networks, effective communication channels, and supportive environments created by platforms exhibit higher levels of well-being. These findings highlight the importance of fostering collaborative and inclusive digital ecosystems to sustain worker morale and satisfaction. Furthermore, technological factors, particularly usability and transformation, significantly influence satisfaction, with positive experiences with user-friendly and reliable tools correlating with greater extrinsic and intrinsic satisfaction. Platforms that invest in intuitive, high-quality technological solutions stand to enhance the work experience for their users.

The research additionally identifies the impact of sociodemographic and psychological variables on job satisfaction. Younger workers under the age of 35 reported lower satisfaction levels, while income positively influenced intrinsic satisfaction, albeit modestly. These insights suggest the necessity of tailored approaches to address the diverse needs of platform workers, ensuring that interventions are inclusive and responsive to varying demographic profiles. The statistical significance of technology-related variables, including usability and transformation, emphasizes the growing centrality of digital tools in shaping work experiences, particularly in the context of extrinsic satisfaction. These findings offer a robust empirical foundation for understanding the dynamic interplay of personal, social, and technological factors in the platform economy.

Despite its contributions, this study has limitations that warrant further exploration. The cross-sectional design precludes an understanding of how job satisfaction evolves over time, limiting insights into the long-term impacts of the identified determinants. Additionally, the focus on a single cultural and geographic context restricts the generalizability of the findings to other regions with differing economic and social dynamics. Addressing these limitations through longitudinal and comparative studies would provide a more comprehensive understanding of job satisfaction in the platform economy.

The findings have substantial implications for policymakers and platform operators. Policies that enhance worker autonomy, provide structured social support, and ensure seamless technological interactions can significantly improve job satisfaction. For platform operators, prioritizing these factors could lead to a more satisfied, productive, and loyal workforce, ultimately contributing to the sustainable growth of their business models. Legislative efforts aimed at ensuring fair working conditions and protecting worker rights will further bolster the well-being of platform workers.

Future research should investigate the long-term effects of autonomy, support, and technological usability on job satisfaction, particularly through longitudinal studies. Comparative research across diverse regions and types of platform work could reveal how cultural and contextual factors mediate these relationships, leading to more tailored and effective strategies. Understanding these nuances is crucial for the development of comprehensive approaches that benefit both workers and platform operators.

As the platform economy continues to evolve, ongoing research is needed to monitor emerging trends and challenges. The dynamic nature of platform work necessitates adaptive policies and practices to maintain relevance and effectiveness. By addressing the persistent and emerging issues faced by platform workers, businesses and policymakers can contribute to a more equitable and sustainable digital economy. Ultimately, improving job satisfaction is not only beneficial for workers but also serves as a strategic advantage for platforms seeking to enhance performance and competitiveness in an increasingly complex economic landscape.