The double-edged sword effects of perceived algorithmic control on platform workers’ service performance

The gig economy, driven by online labor platforms, relies heavily on algorithmic control for worker management. Algorithmic control automates tasks like worker-customer matching, task allocation, and performance evaluation, reducing operational costs. However, the impact of perceived algorithmic control on platform workers' service performance remains under-researched. Existing studies primarily focus on continuance intention, workaround use, and well-being, neglecting the crucial role of service performance, which directly affects customer satisfaction and platform sustainability. This study addresses this gap by proposing a theoretical model based on the transactional theory of stress. This theory posits that individuals appraise stressors as either challenges or threats, influencing their coping mechanisms. Algorithmic control, when perceived as a challenge, can lead to job crafting (proactive job redesign) and improved service performance. Conversely, when perceived as a threat, it can trigger withdrawal behavior (avoidance strategies) and reduced service performance. Regulatory focus theory is integrated to explore how promotion (growth-oriented) and prevention (security-oriented) focus moderate these relationships. Promotion focus is expected to enhance the positive effect of algorithmic control via job crafting and reduce the negative effect via withdrawal, while prevention focus is expected to have the opposite effects. The study aims to reveal the double-edged sword effect of perceived algorithmic control and provide valuable insights for platform management.

Existing research on platform workers' responses to algorithmic control has primarily focused on their continuance intention, the use of workarounds to circumvent algorithmic constraints, and their overall well-being. Studies have shown mixed results, with some suggesting negative relationships between perceived algorithmic control and continuance intention or workaround use, mediated by factors such as legitimacy judgments and technostress. However, there is a significant lack of research investigating the direct impact of perceived algorithmic control on platform workers' service performance. This study aims to fill this gap by examining the mediating roles of job crafting and withdrawal behavior, and the moderating roles of promotion and prevention focus, providing a more nuanced understanding of the complex relationship between algorithmic control and worker performance.

Data was collected using an online questionnaire distributed through the Credamo platform to 391 platform workers (drivers and delivery staff) in China between May and August 2022. After removing invalid responses, 286 valid questionnaires were analyzed. The questionnaire measured perceived algorithmic control (using a 11-item scale from Pei et al., 2021), job crafting (4-item scale from Vogel et al., 2016), withdrawal behavior (4-item scale from Lehman and Simpson, 1992), promotion focus and prevention focus (scales from Koopmann et al., 2016), and service performance (based on self-reported customer ratings). Control variables included age, gender, education, average daily work hours, occupation type, and industry type. Confirmatory factor analysis (CFA) using MPLUS 8.3 was conducted to assess the validity and reliability of the measurement scales. Harman's single-factor test and an unmeasured latent methods factor model were used to assess common method variance (CMV). Hierarchical regression analysis in SPSS was used to test the hypothesized relationships, followed by bootstrapping analysis to assess mediating effects. Simple slope analysis was employed to interpret the moderating effects.

The study's findings support the proposed double-edged sword effect of perceived algorithmic control on service performance. Hierarchical regression analysis revealed that perceived algorithmic control was positively related to job crafting and negatively related to withdrawal behavior. Job crafting was positively related to service performance, while withdrawal behavior was negatively related. Bootstrapping analysis confirmed the mediating roles of both job crafting (positive indirect effect) and withdrawal behavior (negative indirect effect) in the relationship between perceived algorithmic control and service performance. The positive indirect effect of job crafting was significantly stronger than the negative indirect effect of withdrawal behavior. Furthermore, promotion focus positively moderated the relationship between perceived algorithmic control and job crafting, and negatively moderated the relationship between perceived algorithmic control and withdrawal behavior. Prevention focus showed the opposite moderating effects. Moderated mediation analysis revealed that the positive indirect effect of perceived algorithmic control on service performance through job crafting was stronger at high levels of promotion focus and weaker at high levels of prevention focus. Conversely, the negative indirect effect through withdrawal behavior was weaker at high levels of promotion focus and stronger at high levels of prevention focus.

The findings highlight the importance of considering the dual effects of perceived algorithmic control on platform workers. While algorithmic control can enhance performance through job crafting, especially when workers are promotion-focused, it can also lead to decreased performance through withdrawal behavior, particularly when workers are prevention-focused. These findings underscore the need for platform companies to design and implement algorithmic control systems that minimize the negative aspects and maximize the positive aspects. The moderating role of regulatory focus suggests that tailoring interventions to individual worker characteristics can significantly influence the effectiveness of algorithmic control strategies. Future research could explore how to cultivate a promotion focus among platform workers to harness the benefits of algorithmic control.

This study contributes to the literature by demonstrating the double-edged sword effect of perceived algorithmic control on platform workers' service performance, mediated by job crafting and withdrawal behavior, and moderated by promotion and prevention focus. The findings have significant implications for platform management, highlighting the importance of considering both the positive and negative consequences of algorithmic control and tailoring strategies to individual worker characteristics. Future research should investigate the impact of organizational-level factors, employ longitudinal designs, and use multiple data sources to enhance the robustness of the findings.

This study relies on cross-sectional data and self-reported measures, which may limit the generalizability of the findings and introduce potential biases such as common method variance and social desirability bias. The reliance on self-reported service performance also introduces potential measurement error. Future studies should address these limitations by employing longitudinal designs, multiple data sources, and objective performance metrics. The sample was limited to platform workers in China, limiting the generalizability to other cultural contexts. Future research should examine the generalizability across different cultural settings and platform types.