期刊追踪 | Information Systems Research: 2025年第36卷2期

摘要：尽管区域性的数据保护与隐私制度常常被视为跨境数字贸易的重要障碍，但通过监管来缓解消费者的隐私担忧，反而可能会提升他们对外国数字产品或服务的需求。本文通过实证研究对此进行了探讨，考察了《通用数据保护条例》（GDPR）对全球移动应用市场的影响。我们构建了一个涵盖 GDPR 实施前后 26 个月期间、由苹果 App Store 发布的应用程序的综合数据集，并运用计量经济学模型来分析该法规对国家间应用交易的影响。与“区域性数据保护和隐私法规阻碍数字贸易、加剧市场碎片化”的观点相反，实证结果表明，在 GDPR 实施后，欧盟国家中表现最优的外国应用数量显著增加，相较于本土应用更具优势。进一步的系列分析探讨了供给端与需求端可能推动这一效果的潜在机制。研究结果支持需求端的解释：GDPR 有助于缓解消费者的隐私顾虑，并增强他们在采纳外国数字产品时的信心。

Abstract: Although regional data protection and privacy regimes are often cited as major barriers to crossborder digital trade, mitigating consumer privacy concerns through regulations can potentially increase the demand for foreign digital products or services. This study presents empirical evidence on this issue by examining the impact of the General Data Protection Regulation (GDPR) on the global mobile app market. We construct a comprehensive data set of apps distributed by Apple’s App Store over the 26-month period covering the enactment of the GDPR and employ econometric models to analyze the regulation’s effects on app trade between country pairs. Contrary to assertions that regional data protection and privacy regulations impede digital trade and aggravate fragmentation, the empirical results demonstrate a significant increase in top-performing foreign apps compared with native ones in the European Union countries post-GDPR. We further conduct a series of analyses to explore the underlying mechanisms potentially driving these effects from both the supply and demand sides. Our findings lend support to the demand-side mechanism, whereby the GDPR helps alleviate consumer privacy concerns and provides reassurance in adopting foreign digital products.

摘要： 尽管在线问诊正在快速普及，但医生必须积极参与医患互动平台，才能充分发挥其潜力。基于社会影响的视角，本研究以实证方法考察了在在线健康社区（OHC）扩散过程中，跨区域的医生采纳行为。我们重点分析了地理邻近性和社会邻近性采纳者的影响，并探讨了邻近性影响与竞争在采纳过程中的交互作用。研究数据来自中国某大型在线健康社区，涵盖了三个省份32个城市的21,654名医生的面板数据。结果表明：在本地竞争较低的情况下，地理邻近性和社会邻近性均能促进医生采纳。然而，随着本地竞争的加剧，社会邻近采纳者的影响增强，而地理邻近采纳者的影响则减弱。这一模式在职称较低的医生群体中表现得更为明显。

Abstract: Although online doctor consultations are rapidly gaining popularity, physicians must actively participate in physician–patient interaction platforms to fully unlock their potential. Through a social influence lens, this study empirically investigates physician adoption behavior over time across regions in the diffusion of an online health community (OHC). We examined the impacts of geographically and socially close adopters and investigated the interaction of proximity influences and competition in adoption. We collected panel data on 21,654 physicians in 32 cities in three provinces in China from a large Chinese OHC. The results demonstrate that both geographic and social proximity facilitate adoption when local competition is low. However, as local competition increases, the impact of socially close prior adopters increases, whereas that of geographically close prior adopters decreases. This pattern becomes stronger for physicians with lower titles.

摘要：阴谋论指控强势群体策划恶意阴谋，正日益被视为对当代社会的威胁。虽然已有研究承认信息技术（IT）在阴谋论传播中的作用，但对于与技术相关的阴谋信念、其形成过程及其影响的理解仍然有限。基于阴谋信念的理论洞见以及关于个人对技术认知影响的信息系统（IS）研究，本文提出了“技术阴谋信念”的概念。我们将其定义为：个体对一种未经证实的叙事的认可，该叙事声称某项技术的发布组织正在利用该技术秘密追求邪恶目标。在此基础上，我们开发了TECONMIND（technology-conspiracy-mindset）模型。该模型认为，个体对技术特征（绩效期望、努力期望和感知风险）以及技术发布方特征（感知的恶意与感知的权力）的认知，会促成技术阴谋信念的形成。此外，该模型提出，技术阴谋信念与更广泛的阴谋思维之间存在互为强化的关系：技术阴谋信念会助长阴谋思维，而阴谋思维又会进一步促进新的技术阴谋信念。我们通过实证研究证明了技术阴谋信念的普遍存在，并在两项研究中对该模型进行了检验——一项是多阶段的实地研究，另一项是实验研究，结果均支持核心命题。本文对信息系统研究的贡献体现在三个方面：第一，通过提出技术阴谋信念的概念，我们构建了一个以 IT 产物为中心的新研究视角，使学界能够探索以往被忽视的技术阴暗面。第二，我们确立了技术阴谋信念与阴谋思维之间的双向关系，表明技术阴谋信念的认可可能引发一种恶性循环，使个体越来越多地通过阴谋论来解读其所处环境。第三，我们初步揭示了哪些技术特征和发布方特征更容易使某项技术成为阴谋信念的对象。研究结果提醒技术开发者和政策制定者：他们的决策可能引发或缓解技术阴谋信念，而这对社会有着深远的长期影响。

Abstract:  Conspiracy theories, which allege that powerful groups hatch malicious plots, are increasingly recognized as a threat to contemporary society. Although research acknowledges the role of information technology (IT) in spreading such theories, the understanding of conspiracy beliefs related to technology, their formation, and their effects remains limited. Building on theoretical insights on conspiracy beliefs and information systems (IS) research dealing with the impact of individuals’ perceptions of technology, we theorize on technology conspiracy beliefs. We define technology conspiracy beliefs as an individual’s endorsement of an unverified narrative that purports that an organization issuing technology is using that technology to secretly pursue evil goals. We then develop the TECONMIND (technology-conspiracy-mindset) model, which suggests that an individual’s perceptions of a technology’s characteristics (performance expectancy, effort expectancy, and perceived risks) and its issuer’s characteristics (perceived malevolence and perceived power) can lead to the formation of technology conspiracy beliefs. Moreover, the model proposes a reciprocal relationship between technology conspiracy beliefs and a broader conspiracy mindset. Accordingly, we expect technology conspiracy beliefs to foster a conspiracy mindset, which in turn, promotes further technology conspiracy beliefs. We provide empirical evidence for the prevalence of technology conspiracy beliefs and then test this model in two studies—a multiwave field study and an experiment—that support our central propositions. Our study contributes to IS research in three ways. First, by developing a conceptual understanding of technology conspiracy beliefs, we introduce an IT artifact-centered concept that allows researchers to explore a previously overlooked dark side of technology. Second, we establish a reciprocal relationship between technology conspiracy beliefs and conspiracy mindsets, indicating that the endorsement of technology conspiracy beliefs can set in motion a vicious cycle in which individuals increasingly interpret their environment using conspiracy theories. Third, we provide an initial understanding of which perceived technology and issuer characteristics make technologies prone to become objects of conspiracy beliefs. Our findings should sensitize technology developers and policymakers as to how their decisions can instigate or mitigate technology conspiracy beliefs, which have significant long-term societal consequences.

摘要：远程工作的日益普及加速了监督系统的采用，这些系统用于追踪劳动者的行为，尤其是在在线劳务平台上。与既有文献主要聚焦于监督对生产率的影响不同，本研究从契约治理的视角探讨监督的作用。理论上，通过实现对劳动者进展的实时细致观察，监督系统的部署有助于提升契约控制与协调，从而降低雇主对本国劳动者的偏好（即“本土偏好”）。本文利用某领先在线劳务平台在计时项目中外生性引入监督系统的事件，采用双重差分模型评估监督系统在减少本土偏好方面的作用。研究发现，在监督系统引入后，对外国劳动者的偏见显著减弱并且在统计上不再显著，这凸显了监督系统在促进全球劳动者公平竞争中一个被忽视的作用。进一步分析表明，监督系统的引入使外国劳动者的雇佣比例提升了15%。此外，本土偏好的减弱在以下两类情境中更为明显：其一是高度常规化的项目（外部不确定性较低）；其二是雇主缺乏与外国劳动者的正面合作经验（内部不确定性较高）。另外，当劳动者评分较高（预期的道德风险较低）或与雇主处于相同时区（预期的协调成本较低）时，雇主也不再表现出更强的本土偏好。这些发现支持了监督系统通过提升契约控制与协调来缓解雇主本土偏好的有效性。研究结论对在线劳务平台的设计具有重要的管理启示。

Abstract: The increasing prevalence of remote work has accelerated the adoption of monitoring systems to keep track of worker behavior, especially on online labor platforms. In contrast to the existing literature that predominantly focuses on the effect of monitoring on productivity, this study investigates the impact of monitoring from the perspective of contractual governance. In principle, by enabling the detailed real-time observation of worker progress, the deployment of monitoring systems has the potential to improve contractual control and coordination, thereby reducing employers’ preferences for domestic workers (home bias). Leveraging the exogenous introduction of a monitoring system for time-based projects on a leading online labor platform, we employ a difference-in-differences model to estimate the impact of monitoring systems in reducing home bias. Our findings reveal that following the monitoring system’s introduction, the bias against foreign workers becomes substantially weaker and statistically insignificant, highlighting the overlooked role of monitoring systems in fostering a more level playing field for global workers. Our further analysis indicates that monitoring leads to a notable 15% increase in the hiring of foreign workers. Moreover, the decrease in home bias is more pronounced in high-routine projects or when employers lack prior positive experiences with foreign workers, two scenarios characterized by low external uncertainty and high internal uncertainty, respectively. Additionally, employers no longer exhibit a stronger home bias when workers have higher ratings, where the expected moral hazard risk is lower, nor when workers reside in the same time zone, where expected coordination costs are lower. These findings lend support to the effectiveness of monitoring systems in mitigating employers’ home bias through enhancing contractual control and coordination. Our findings provide important managerial implications for the design of online labor platforms.

摘要：学习消费者的个性化隐私偏好对企业和政策制定者至关重要，它有助于建立信任与合规机制，并为有效的政策制定提供指导。现有方法主要依赖于私有信息，例如专有的用户行为数据、个体层面的社会经济与人口统计特征，或需要用户明确输入。这些方式可能具有侵入性和负担，甚至导致用户不满。当下，个体在公共领域中不断生成和分享大量与自身相关的信息，这些信息能够从多维度反映其行为、态度与偏好，从而为推断隐私偏好提供宝贵洞见。基于此，本文提出了一种全新的框架，利用一种普遍存在的公共数据来源——社交媒体发帖，来预测个性化隐私偏好。在心理学与隐私理论的支撑下，我们通过深度学习模型与自然语言处理算法，从社交媒体文本中提取并学习一系列理论驱动的心理社会特征，包括生活方式、风险偏好、人格特质、与隐私相关的经济偏好、语言风格等。有趣的是，我们发现来自公共数据的心理社会特征在预测力上优于私有信息。此外，本文进行了多种可解释性分析，以揭示模型性能的驱动因素。最后，我们通过实证展示了该模型的实际应用价值，证明该框架能够帮助平台和政策制定者预测隐私政策的后果。总体而言，本研究为增强消费者隐私控制与信任、优化平台数据管理、以及为政策制定者提供更优的数据隐私监管建议提供了重要的管理启示。

Abstract: Learning consumers’ personalized privacy preferences is crucial for firms and policymakers to establish trust and compliance and guide effective policymaking. Existing approaches rely mostly on private information such as proprietary user behavior data and individual-level demographic and socio-economic factors, or require explicit user input, which can be invasive and burdensome, potentially leading to user dissatisfaction. Nowadays, individuals generate and share vast amounts of information about themselves in the public domain, which can provide a valuable multifaceted view of their behaviors, attitudes, and preferences. This information thus has the potential to provide valuable insights into individuals’ privacy preferences. In this study, we propose a novel framework to predict personalized privacy preference by leveraging a ubiquitous source of public data—social media posts. Deeply rooted in psychological and privacy theories, we use deep learning model and natural language processing algorithms to learn theory-driven psychosocial traits such as lifestyle, risk preference, personality, privacy-related economic preferences, linguistic styles, and more from social media posts. Interestingly, we find that psychosocial traits from public data provide greater predictive power than private information. Furthermore, we conduct multiple interpretability analyses to understand what drives the model’s performance. Finally, we demonstrate the practical value of our model and show that our framework can assist platforms and policymakers in forecasting the consequences of privacy policies. Overall, our framework provides managerial implications for enhancing consumer privacy control and trust, optimizing platform data management, and informing policymakers about better data privacy regulations.

摘要：数字平台通常通过货币激励来驱动用户完成特定任务。现有研究主要考察了在公共平台上引入此类货币奖励对相关结果（如参与度与绩效）的影响。然而，对于取消奖励的效果（即是否仅仅是其引入作用的逆向过程），尤其是在企业平台上的影响，学界知之甚少。本文基于企业平台的准自然实验研究了取消货币激励的影响。与既有关于公共平台的研究一致，我们发现，引入基于数量的货币激励可以提升参与度（贡献数量），但对绩效（贡献质量）没有显著影响。然而，我们的核心实证结果显示：取消货币激励并非其引入作用的简单逆转。具体而言，相比于最初引入奖励时带来的参与度提升，取消奖励导致的参与度下降幅度更大，这表明取消奖励对参与度有净负面效应。此外，取消货币奖励还会引起绩效显著下降，这说明在绩效层面上，取消与引入激励的效果并非对称。进一步地，我们考察了不同自我动机类型与工作能力水平的个体对货币激励的差异化反应，揭示了激励“引入”与“取消”之间的不对称性。最后，本文讨论了企业平台与公共平台在货币激励影响上的异同。研究结果对企业信息系统及一般信息系统在平台战略设计方面具有重要实践启示。

Abstract: Digital platforms commonly use monetary incentives to motivate users to perform specific tasks. Extant studies have shown the effects of introducing such monetary rewards on the outcomes of interest (e.g., participation and performance) on public platforms. However, little is known about the impact of canceling rewards (i.e., whether it simply reverses the effect of their introduction), and particularly less attention is paid to corporate platforms. Our study examines the impact of canceling monetary incentives using quasi-natural experiments on a corporate platform. Similar to prior studies focusing on public platforms, we find that introducing quantity-based monetary incentives increases participation (contribution quantity), but has no significant effect on performance (contribution quality). Yet, in contrast, our main empirical analysis reveals that canceling monetary incentives is not simply the reverse process of their introduction. In particular, compared with the increase in participation when monetary rewards were initially introduced, the cancellation of these rewards leads to a sharper decrease in participation. This suggests that canceling rewards has a net negative impact on participation. In addition, canceling monetary rewards also causes a significant decline in performance, which indicates that the effects of canceling and introducing rewards on performance are not simply the opposite of each other. Furthermore, we examine the heterogeneous responses of individuals with different self-motivation types and working competency levels to monetary incentives, highlighting the “asymmetry” between canceling and introducing incentives. We also discuss the similarities and differences between corporate and public platforms regarding the impact of monetary incentives. Our results provide important practical implications for enterprise information systems and general information systems regarding their design of platform strategies.

摘要：近年来，一些用户生成内容（UGC）平台引入了购物功能，希望通过应用内交易的佣金分成获得额外收入。这类平台允许创作者制作可购物内容，即在内容中标记带有联盟链接的推荐产品，创作者从商家处获得佣金后，平台再从中抽取一定比例。这种新型商业模式被称为 平台主导的联盟营销。尽管其日益流行，但现有文献尚未系统分析这种商业模式对关键利益相关者（即 UGC 平台、内容创作者和内容消费者）的影响。本文采用博弈论模型来填补这一研究空白。研究结果表明，平台主导的联盟营销可以在 UGC 平台、参与联盟营销的创作者和内容消费者之间创造“双赢”局面。此外，即便是不参与联盟营销的创作者，有时也能从其他创作者的参与中间接受益。然而，出乎意料的是，UGC 平台采用该模式并不必然受益，因为可能出现流量收入下降的情况。这一发现为平台是否以及如何利用联盟营销提供了重要的管理启示。进一步的分析还表明，在竞争加剧的环境下，创作者反而可能通过减少内容产出而获益。更重要的是，创作者的最优内容生产决策会因多种因素（如创作者数量增加、内容可替代性增强）而显著不同，这些因素共同推动了竞争的加剧。综上，本研究为 UGC 平台的战略制定与创作者的内容生产优化提供了有价值的洞见。

Abstract: Recently, some user-generated content (UGC) platforms have introduced shopping features to generate additional commission-based revenue from in-app transactions. These platforms allow creators to produce shoppable content, where promoted products are tagged with affiliate links, and in return, they earn a percentage of the commission the creators receive from merchants (such a new business model is referred to as platform-led affiliate marketing). Despite its growing popularity, the impact of this business model on the key stakeholders (i.e., UGC platforms, content creators, and content consumers) has yet to be systematically analyzed in the existing literature. This paper aims to address this significant literature gap by employing a game-theoretic model. Our findings demonstrate that platform-led affiliate marketing can create a win-win situation for UGC platforms, creators participating in affiliate marketing, and content consumers. Furthermore, we find that even creators not participating in affiliate marketing can sometimes benefit indirectly from other creators’ participation. On the other hand, surprisingly, we find that UGC platforms may not necessarily benefit from adopting this emerging business model because of potential losses in traffic revenue. These results carry significant managerial implications for UGC platforms regarding whether and how to leverage affiliate marketing. Moreover, our findings suggest that creators may benefit from reducing their production efforts in the face of intensified competition. More importantly, optimal production decisions for creators can vary significantly, depending on specific factors (e.g., an increase in the number of creators or content substitutability) contributing to heightened competition. These findings offer valuable insights that can guide creators in refining their production strategies.

摘要：移动应用是证券经纪行业中最重要、最广泛使用的金融科技（fintech）创新之一。令人意外的是，尽管其经济意义和理论价值日益增加，但关于移动应用如何影响个人投资者金融决策与绩效的研究仍然有限。本文利用中国一家大型证券公司的专有纵向数据（2012年12月至2015年11月），在准自然实验的框架下探讨移动应用对投资者交易行为与投资组合绩效的影响。我们利用移动应用的引入作为识别事件，在20,665名投资者的样本中考察移动应用采纳的效果。采用广义合成控制法的结果表明：若仅以二元指标衡量移动应用使用，整体上看，采纳移动应用对投资者的投资组合绩效没有显著影响。机制分析显示，移动应用的使用显著降低了时间约束（作为交易摩擦的代理变量），但同时也略微提升了追涨偏好，表现为短视决策倾向。由于时间约束的降低有助于改善绩效，而追涨行为可能损害绩效，两者效应相互抵消，从而解释了移动应用采纳对整体投资组合绩效“无显著影响”的结果。进一步分析用户采纳后的行为，发现移动应用的使用强度与投资组合绩效呈现倒U型关系。该结果在不同样本下、排除市场高波动期，以及通过多种方法（倾向得分匹配、动态匹配、堆叠双重差分、工具变量法等）检验后均保持稳健。本文的研究为学术研究与实践提供了重要启示。

Abstract: Mobile apps are among the most important and widely used financial technology (fintech) innovations in the brokerage industry. Surprisingly, despite their increasing economic importance and theoretical significance, few studies examine the effects of mobile app use on individual investors’ financial decisions and performance. This study seeks to understand how mobile apps influence investors’ trading behaviors and portfolio performance by using a proprietary longitudinal data set from December 2012 to November 2015 from a large securities company in China with a quasi-experimental setting to answer our research questions. We leverage the introduction of an app to identify the effect of mobile app adoption by using a sample of 20,665 investors. We use the generalized synthetic control method and find that mobile app adoption does not affect investors’ portfolio performance when one examines aggregate impacts using a binary indicator of mobile app use. Our analyses of the mechanisms indicate that adopting mobile apps results in a noticeable decrease in time constraints, a proxy for transaction friction, and a modest increase intrend-chasing bias, reflecting tendencies toward myopic decision making. Because the reduction in time constraints can benefit investors’ performance, the increase in trend-chasing can be detrimental to investors’ performance; our findings explain why mobile app adoption has no overall effect on portfolio performance. Further analyses of adopters’ postadoption behaviors provide interesting insights and show that the mobile app usage intensity has an inverted U–shaped relationship with portfolio performance. The results are robust to using different samples or excluding high market volatility periods and by using a variety of methods, such as propensity score matching, dynamic matching, stacked difference in differences, or an instrumental variable approach. We discuss the implications for research and practice.

摘要：“打卡返现”（Clocking-in cash-back, CIC）作为一种新兴的在线学习游戏化商业模式，近年来引起了广泛关注。该模式允许用户在规定时间窗口内连续完成特定任务，即“打卡”（如每日作业、在社交媒体分享学习进度等），从而获得课程费用的全额返还。通过这一机制，平台能够有效监控用户的学习投入，并依据打卡完成情况将用户划分为“坚持者”或“放弃者”。尽管日益流行，但现有文献尚未对这一新模式进行系统性分析。本文填补了这一空白，探讨了在线学习企业应如何设定课程的最优时间窗口，以及该时间窗口如何受到情境因素的影响。研究识别了延长时间窗口对用户放弃时间的两种相反作用：一是“心理负效应增强作用”（负面影响），二是“努力成本降低作用”（正面影响）。研究结果表明，当放弃者的正面口碑效应增强时，企业在某些情况下反而应缩短时间窗口，因为心理负效应的增加超过了努力成本的下降。此外，研究发现，在高能力用户比例上升的情况下，延长时间窗口并不总是有利。进一步结果显示，当边际内容生产成本上升时，企业不应通过降低单个任务的难度来应对，反而可能通过缩短时间窗口、提高任务难度来获得更大优势。本研究为在线学习企业优化CIC机制的设计提供了有价值的管理启示。

Abstract: Clocking-in cash-back (CIC), an emerging gamified business model in online learning, has recently garnered significant attention. CIC allows users to secure a full refund of the course fee through consecutive completion of specific tasks within a required time window. These tasks, known as clocking in, encompass activities such as daily assignments and sharing progress updates on social media. By employing this gamification system, the firm effectively monitors user efforts, categorizing them as winners or quitters based on clocking-in completion. Despite its growing popularity, this new business model has yet to be systematically analyzed in the literature. This paper fills this critical gap by examining how an online learning firm should set the optimal time window for its course and how the time window is affected by context-specific factors. We identify two opposing effects associated with extending the time window on users’ quitting time: the psychological-disutility increasing effect (negative) and the effort-cost decreasing effect (positive). Our results indicate that as quitters’ positive word-of-mouth (WOM) effects increase, there are cases in which the firm should opt for shortening the time window, primarily because of the psychological-disutility increasing effect outweighing the effort-cost decreasing effect. Furthermore, we find that it is not always beneficial for the firm to extend the time window when there is an increased presence of high-ability users. Additionally, we find that as the marginal content creation cost rises, instead of reducing the difficulty level of each task, the firm may find it more advantageous to raise the difficulty level by shortening the time window. Our findings provide valuable insights that online learning firms can utilize to enhance their design of the CIC mechanism.

摘要：在情境感知推荐系统（CARS）中，情境化场景（例如“周五与配偶在餐厅共进晚餐”）已经成为一种重要的上下文表示机制。以往研究表明，将情境信息嵌入欧几里得空间的潜在表示方法，能够提升推荐效果。然而，这类传统方法在建构层级化情境信息的嵌入时面临重大挑战，同时在生成有助于管理者理解与使用的可解释表示上也存在不足。为解决这些问题，本文提出了 HyperCARS方法，即在潜在的双曲空间中建模层级化的情境。HyperCARS将双曲嵌入与层级聚类相结合来构建情境，从而实现上下文建模组件与推荐算法的松耦合，使其能够灵活适配多种已有的推荐算法。实证结果表明，与欧几里得嵌入相比，双曲嵌入更好地捕捉了情境的层级特性，并在多个层级上生成了更加清晰、区分度更高的情境划分。同时，基于双曲嵌入的情境能够在标准推荐指标上实现更优的推荐效果，并提升生成层级情境的可解释性。此外，由于双曲嵌入在CARS之外的诸多应用中也具有潜力，本文进一步提出了一个潜在嵌入表示的框架，对以往嵌入相关研究进行了系统分类，并为双曲嵌入在信息系统应用中的未来研究路径提供了方向。

Abstract: Contextual situations, such as having dinner at a restaurant on Friday with the spouse, became a useful mechanism to represent context in context-aware recommender systems (CARS). Prior research has shown important advantages of using latent embedding representation approaches to model contextual information in the Euclidean space leading to better recommendations. However, these traditional approaches have major challenges with construction of proper embeddings of hierarchical structures of contextual information, as well as with interpretations of the obtained representations that would be useful for managers. To address these problems, we propose the HyperCARS method that models hierarchical contextual situations in the latent hyperbolic space. HyperCARS combines hyperbolic embeddings with hierarchical clustering to construct contextual situations, which allows to loosely couple the contextual modeling component with recommendation algorithms and therefore provides flexibility to use a broad range of previously developed recommendation algorithms. We demonstrate empirically that the proposed hyperbolic embedding approach better captures the hierarchical nature of context than its Euclidean counterpart and produces hierarchical contextual situations that are more distinct and better separated at multiple hierarchical levels. We also demonstrate that hyperbolic contextual situations lead to better context-aware recommendations in terms of standard recommendation metrics and to better interpretability of the resulting hierarchical contextual situations. Because hyperbolic embeddings can also be used in many other applications besides CARS, in this paper, we propose the latent embeddings representation framework that systematically classifies prior work on embeddings and identifies novel research streams for hyperbolic embeddings across information systems applications.

摘要：现有关于阻碍型技术压力源（hindrance technostressors）的研究模型众多，且各自包含不同的压力源集合。这使得研究者和实践者在选择某一模型或尝试整合多个模型时面临困难，因为缺乏关于其相对优势与适用性的指导。同时，现有模型并未涵盖阻碍型技术压力的完整概念广度，通常也缺乏对压力源导致的负面心理反应或结果的足够解释力。本文综合了零散的阻碍型技术压力研究成果，提出了一个统一的数字阻碍压力源层级模型，并构建了一个简洁而全面的测量模型，具备较高的预测力。研究基于技术压力理论与职业压力理论，采用以量化研究为主的混合方法。实证部分包括一项定性预研究以及多轮调查，总计超过 5,800 名参与者。数据支持了新模型的建构、验证与基准检验。最后，本文讨论了该模型在研究与实践中的相对优势，并为模型选择提供了指导。

Abstract: There are many models with various sets of hindrance technostressors. Researchers and practitioners face the challenge of selecting a model or mixing several models without guidance on their relative advantages and suitability for contemporary digital work. None of the existing models captures the full conceptual breadth of hindrance technostress, and the existing models typically offer suboptimal power to explain the negative psychological responses or outcomes of technostressors. We synthesize the fragmented works on hindrance technostressors and propose a unified hierarchical model of digital hindrance stressors. We provide an extensive and parsimonious measurement model with high predictive power. This work builds on technostress and occupational stress theory using a quantitative-dominant mixed-methods study. The empirical part of the study includes a qualitative prestudy and multiple surveys with more than 5,800 participants. The data support the modeling, validation, and benchmarking of the new models we introduce. We discuss the relative advantages of the models for research and practice and guide their selection.

摘要：云服务运行在一个具有动态特征的共享安全环境中：用户将固定成本转化为随时间变化的可变成本，而云服务提供商（CSPs）与用户共同为整体安全做出贡献。本文通过动态博弈研究共享安全的特性，其中用户的安全投入与云使用水平会影响其云服务商的脆弱性。同时，CSP的安全投入不仅考虑自身用户，还受到与其他CSP之间竞争的影响。Markov完美均衡揭示了云安全的长期动态演变路径。研究特别识别出一种新的时间路径上的战略互补性：云使用水平与CSP的安全状态之间相互影响。这表明，云安全是一种不同寻常的“不纯公共品”形式，其中个体对CSP安全的投入并非只为自己带来收益，而是为他人提供了选择性激励（私人收益）。由于这会推动云使用增加，CSP的脆弱性也随时间上升。与此同时，CSP在安全上的竞争可能既带来用户福利改善，也可能导致“锁定效应”。

Abstract: Cloud services exist under a shared security environment with a dynamic nature; users trade fixed costs for variable costs over time, and both cloud services providers (CSPs) and users contribute to overall security. We investigate the nature of shared security in a dynamic game where users’ security contributions and cloud usage figure into their CSP’s vulnerability. Furthermore, CSPs’ own security contribution takes into account both their users as well as competition with other CSPs. The Markov perfect equilibrium reveals the long-term time patterns of security of the cloud. In particular, we identify a novel form of time-path strategic complementary between usage and a CSP’s Markov state of security. This implies that cloud security is an unusual form of impure public good, whereby individual contributions bolstering a CSP’s security endow a selective incentive (private benefit) on others rather than on the contributor alone. Because this increases usage, CSP vulnerability increases over time. At the same time, CSP competition on security may lead to both welfare improvements for users and lock-in.

摘要：不确定性，尤其是分布不确定性（亦称“模糊性”），在学术研究与实践应用中都具有重要意义。然而，现有研究大多集中于结果不确定性（或风险）的应对，而对分布不确定性的关注相对不足。本文聚焦于这一关键但常被忽视的研究领域，指出在数据分析模型的开发与应用中亟需直接纳入对模糊性的考量，并推动建立和广泛应用明确定义的模糊性度量方法。

我们提出了一种新的量化模糊性度量方法，它能够比传统方法更准确地刻画分布不确定性。本文展示了该度量的性质与优势，强调其在提升实证模型、生成更可靠的参数估计以及改善决策过程中的价值。以金融市场的决策为例，我们进一步验证了该模糊性度量的应用价值。

本文有助于深化对不确定性的理解，并对研究方法和实践中的风险管理提供了重要启示。

Abstract: Uncertainty, particularly distribution uncertainty (a.k.a. ambiguity), holds significant relevance in both academic research and practical applications. Much of the existing research, however, has concentrated primarily on addressing outcome uncertainty (or risk), frequently neglecting the aspect of distribution uncertainty. This research delves into distribution uncertainty, a critical yet often overlooked aspect of empirical research. We argue that there is a pressing need to integrate considerations of ambiguity directly into the development and implementation of data analytics models, calling for the promotion and wider use of a well-defined measure of ambiguity. We introduce a quantitative measure of ambiguity that surpasses conventional approaches by precisely capturing distribution uncertainty. We illustrate the properties and advantages of this measure, highlighting its ability to enhance empirical models, yield more reliable parameter estimates, and contribute to the decision-making process. Using decision making in the financial market as an example, we demonstrate the value of this ambiguity measure. This paper promotes a more nuanced understanding of uncertainty and offers implications for both research methodologies and practical risk management.

摘要：商业模式的选择对数字内容平台的成功具有深远影响，既会影响消费者，也会影响创作者。在连载数字内容领域，如何建立符合创作者利益的激励机制是一大挑战。网络小说平台允许作者在线创作并发布作品，传统上多采用按章节付费模式，读者需付费才能获取内容。然而，近年来出现了一种广告赞助模式，通过免费内容吸引读者。尽管该模式拓展了读者的可及性，但其对作者创作的影响仍不明确。本文利用某领先网络小说平台的政策变动作为自然实验，研究了广告赞助模式对作者创作行为的影响。结果显示：在广告赞助模式下，作者整体创作产出有所下降。为揭示其内在机制，我们聚焦读者参与度并进行了中介效应分析。结果表明，广告赞助模式降低了读者的参与度，进而减少了作者的创作投入。进一步的读者层面分析发现：消费广告赞助内容的读者，与作者的互动显著减少，这为主结论提供了更多支持。研究表明，相比付费内容，免费广告赞助内容使读者感知到的“沉没成本”降低。从心理学角度看，付费内容的沉没成本会激励读者持续参与并与作者互动，而广告赞助阅读则消除了这一机制，从而导致参与度下降。本研究对数字内容平台具有重要启示，强调平台管理者在实施广告赞助模式时必须谨慎权衡其对创作者和消费者的双重影响。

Abstract: The selection of a business model significantly impacts the success of digital content platforms, influencing both consumers and creators. Within the field of serialized digital content, establishing incentives that align with creators’ interests poses a notable challenge. Web novel platforms, which allow writers to produce and market their work online, have traditionally adopted a pay-per-view model, where readers pay for access to content. However, a recent trend has seen the introduction of an ad-sponsored model, aimed at attracting readers with free content. Although the ad-sponsored model expands reader accessibility, its effects on writers remain uncertain. Our study investigates the impacts of the ad-sponsored model on writers’ content creation efforts, leveraging a policy change within a leading web novel platform as a natural experiment. Our results indicate a general decrease in writers’ productivity under the ad-sponsored model. To uncover the underlying mechanism, we focus on the role of reader engagement and conduct a mediation analysis. Our results demonstrate that the ad-sponsored model is associated with less reader engagement, which leads to a subsequent decrease in writers’ content creation efforts. Further, we conduct a reader-level analysis and find a significant decline in engagement with writers among those who consume ad-sponsored content, providing additional evidence supporting our main findings. These results substantiate our hypothesis that diminished reader engagement stems from a perception of reduced “sunk costs” associated with free, ad-sponsored content compared with paid books. From a psychological standpoint, the sunk costs inherent in paid content motivate continued reader engagement and interaction with writers, whereas ad-sponsored reading removes such costs, thereby decreasing reader engagement. This study holds profound implications for digital content platforms, underscoring the necessity for platform managers to carefully evaluate the impacts of implementing the ad-sponsored model on both creators and consumers of serialized digital content.

摘要：尽管在线学习平台已广泛普及，但其在维持学习者动机方面仍面临挑战，表现为用户参与度低、学习绩效差。为解决这一问题，本文探讨平台如何通过营造有效学习环境并诱发不同的情境成就目标来提升学习动机。研究基于成就目标理论，在一门拥有来自171个国家、超过2000名学习者的大规模在线课程中，设计并实施了随机实地实验。通过多种计量经济学分析，本文估计了三类情境成就目标干预的效果：学习目标、成绩-证明目标（performance-prove）与成绩-回避目标（performance-avoidance）。与传统线下教育文献中“学习目标最优、成绩目标次之”的结论不同，本文发现：成绩-证明目标在提升在线学习参与度与绩效方面最为有效。这一结果源于线上与线下学习环境结构及学习者心理需求的差异。进一步的机制分析表明，不同情境成就目标通过激发不同的成就情绪，从而维持学习过程中的参与度。此外，研究发现目标效果依赖于学习者特征：先前成绩较高者更受益于成绩-证明目标；中等成绩者从成绩-回避目标中获益；成绩较低者则更受学习目标正向影响；社交隔离的学习者对成绩目标反应更佳。通过第二项实地实验，研究进一步探讨了目标的最佳组合，发现结合“学习目标”与“成绩-证明目标”能带来最高学习成效。本文对学术研究与实践均有重要意义，为学者与平台提供了关于如何有效利用情境成就目标及相关情绪机制来改善在线学习结果的启示。

Abstract: Despite their prevalence, online learning platforms have difficulty sustaining user motivation, resulting in low engagement and poor performance. Addressing this challenge, we study how these platforms can boost learner motivation by fostering an effective learning environment and inducing different situational goals. We conducted a randomized field experiment with behavioral interventions inspired by the achievement goal theory in a massive open online course with more than 2,000 learners from 171 countries. Using various econometric analyses, we estimate the effects of interventions based on three situational achievement goals: learning, performance-prove, and performance-avoidance. In contrast to the traditional (offline) education literature, which finds learning goals to be the most effective and performance goals to be inferior, we demonstrate that performance-prove goals are the most effective in enhancing online engagement and performance. We trace the roots of this finding to the differences in the structure of online and offline environments and the psychological needs of online learners. We uncover the underlying mechanism by empirically examining how situational achievement goals stimulate different achievement emotions necessary to maintain engagement in the learning process. We further find that the effectiveness of each goal depends on learners’ characteristics: prior performance and social activity. Learners with stronger prior performance benefit more from the performance-prove goal, whereas those with moderate performance levels gain from the performance-avoidance goal, and those with lower prior performance are positively influenced by the learning goal. Socially isolated learners respond best to performance goals. Through a second field experiment, we explore the optimal combination of situational goals. We find that combining learning and performance-prove goals leads to the highest learning outcomes. Our study offers theoretical contributions and practical implications for scholars and platform providers on how to effectively leverage situational achievement goals and related achievement emotions for improving online user outcomes.

摘要：虚拟现实（VR）已成为房地产平台的一项变革性补充，正在革新房产信息的呈现方式。与传统的展示技术（如图片和视频）相比，VR为消费者构建了一个交互式的三维（3D）环境，以获取房产信息。以往文献主要从行为学角度考察VR的效果（如生动性和交互性），对象多为图书、服装等低涉入产品的销售。房地产则是典型的高涉入产品，在产品价值、重要性、风险以及中介参与度方面具有显著差异，因此需要全面的产品信息。我们利用某领先房地产平台的大规模数据集，从信息视角研究了VR对房产市场结果（包括上市时间和成交价格）的影响。研究发现，VR是一种“效率提升器”，能够加快房产的成交时间，而不是像“市场价值提升器”那样显著提高成交价格。对于产品涉入度更高、产品质量更好以及中介服务质量较低的房产，VR的加速作用更为明显。这表明VR在以下方面的能力：增强信息丰富性，以满足消费者对高涉入产品的信息需求；建立信息可信度，从而帮助识别产品质量；在中介服务质量不足时，作为替代性信息来源。此外，我们区分了VR实景看房（VR Tour）的直接效应和VR标识（VR Badge）的间接效应，揭示了 VR 在消费者房产评估中的有效性。本文对非沉浸式VR的相关文献作出贡献，并提供了管理启示，尤其是在房产销售、高涉入产品营销以及电子商务中的产品展示技术方面。

Abstract: Virtual reality (VR) has emerged as a transformative addition to real estate platforms, revolutionizing the presentation of property details. In contrast to traditional presentation technologies (e.g., pictures and videos), VR constructs an interactive three-dimensional (3D) environment for consumers to obtain property information. Prior literature mainly examined VR’s effects (e.g., vividness and interactivity) on selling low-involvement products such as books and apparel from a behavioral perspective. Real estate properties are typical high-involvement products, differing significantly in product value, importance, risk, and agent participation, necessitating comprehensive product information. We use a large-scale data set from a leading real estate platform to investigate VR’s influences on properties’ market outcomes (including the time-on-market and selling price) from an informative perspective. We find that VR is an “efficiency enhancer,” accelerating properties’ time-on-market rather than increasing properties’ selling price as a “market value influencer.” VR shows a greater acceleration effect for properties with higher product involvement, higher product quality, and lower agent service quality. This demonstrates VR’s ability in enhancing information richness to satisfy consumers’ information needs for a higher level of product involvement; establishing information credibility in discerning product quality; and serving as an alternative information source in lieu of low-quality agent services. We also disentangle the direct effect of VR Tour from the indirect effect of VR badge on consumers’ decision making to reveal VR’s effectiveness in product evaluation. This work contributes to the literature on nonimmersive VR and offers managerial implications, particularly in the context of property sales, high-involvement product marketing, and product presentation technologies in e-commerce.

摘要：为缓解漏洞被恶意利用带来的威胁，越来越多不同行业的组织开始将漏洞奖励计划（Bug Bounty Programs, BBPs）纳入漏洞管理周期。漏洞奖励计划能够吸引外部安全研究人员，帮助发现组织信息技术系统中的漏洞，但同时也会在漏洞被发现后增加风险。为应对这种权衡，组织需要理解如何设计最优奖励金额，并根据若干关键因素评估漏洞奖励计划的总成本。业界关心的问题包括：(i) 组织特征（如安全状况与修补复杂度）；(ii) 安全研究人员的因素（如研究人员之间的异质性及其数量）；以及 (iii) 计划所处的法律框架等其他因素如何影响奖励金额和总成本。然而，由于缺乏形式化分析，本文借助博弈论模型来探讨相关问题，并提供了一系列有益的结论与管理启示。首先，虽然组织的修补复杂度与奖励金额之间存在替代关系，但安全状况与奖励之间的关系却未必是替代或互补的。其次，更多或更有能力的安全研究人员，并不必然意味着奖励金额的增加或总成本的降低。此外，尽管业界普遍认为对安全研究人员提供更高的法律保护会增加漏洞奖励计划的成本，但研究发现，无论奖励金额还是总成本都不必然增加或减少。这一复杂结果取决于两类成本：漏洞本身带来的固有成本，以及因奖励计划可能泄露而产生的相关成本。本研究为安全专业人士、组织和政策制定者在设计具成本效益的漏洞奖励计划时提供了新的见解与参考。

Abstract: To mitigate the threats from malicious exploitation of vulnerabilities, an increasing number of organizations across different industries have started incorporating bug bounty programs (BBPs) in their vulnerability management cycles. Whereas a BBP attracts external security researchers to facilitate the discovery of vulnerabilities in organizations’ information technology systems, it also increases the risks after the vulnerabilities are discovered. To deal with the trade-offs, organizations need to understand how to design an optimal bounty and evaluate the total cost of a BBP depending on several key factors. The industry is motivated to understand how the bounty and total costs are impacted by (i) the characteristics of the organization (e.g., security posture and patching complexity), (ii) security researchers (e.g., the heterogeneity among security researchers and their number), and (iii) other factors such as the legal framework surrounding the BBP. However, because there is a lack of formal analyses regarding these issues, we use game-theoretical models to shed light on relevant questions and provide several useful results and managerial insights. First, although an organization’s patching complexity and the bounty act as substitutes, the relationship between security posture and the bounty is not necessarily substitutive or complementary. Furthermore, having a larger number of or more capable security researchers does not necessarily imply an increased bounty or lower total costs. Moreover, whereas the prevalent business belief is that an increased level of legal protection offered to the security researchers increases the cost of the BBP, we find that neither the cost of the BBP nor the offered bounty necessarily increases or decreases. This nuanced finding depends on different types of costs incurred because of the inherent vulnerability itself and costs related to possible leaks out of the BBP. Our study provides insights to security professionals, organizations, and policymakers in designing cost-effective BBPs.

摘要：在在线竞争环境中，关于竞争对手专业性的资讯会如何影响核心用户的参与度？在线竞争环境与以往研究的其他在线情境（如电商、社交媒体平台）存在根本区别，而这种情境差异至关重要，因为它决定了他人信息与核心用户参与度之间的关系。这一问题也与学界对地位与过往表现作为专业性信号的相对效应和绝对效应存在的理论模糊性相关。本文基于一项实地实验，在一款多轮移动双人对战游戏的界面中，随机向玩家展示竞争对手的地位与过往表现信号，并考察其对核心用户参与度的影响。我们将用户在每轮结束后是否继续游戏作为参与度的衡量指标。基线结果显示：当竞争对手的当前表现更高时，用户继续游戏的可能性降低。然而，当竞争对手具有较高地位或中等水平的过往表现时，这一“当前表现效应”会显著增强。进一步分析发现，这些或然效应部分取决于核心玩家的竞争动机。本研究为提升在线竞争环境下的用户参与提供了重要启示，同时也深化了我们对地位与过往表现信息如何影响在线参与度的理解。

Abstract: How does information on a rival’s expertise influence the focal user’s engagement in online competitive settings? Online competitive settings differ fundamentally from other online contexts, such as e-commerce and social media platforms, explored in the prior work. This contextual distinction, we argue, is important as it determines the relationship between information about others and the focal user’s engagement. The relevance of this question relates to broader theoretical ambiguities concerning the effects of status and past performance as relative and absolute signals of expertise. Our findings are based on a field experiment in which we modified the multiround mobile two-player gaming app interface, randomly exposing players to rival’s status and past performance signals. We measure the focal user’s engagement regarding their decision to continue the game after each round. As a baseline, we find that the focal user is less likely to continue the game if the rival exhibits higher current performance. However, the rival’s status and past performance signals create strong contingencies wherein the principle effect of current performance is stronger if the rival has a high status or moderate past performance. Further, these contingent effects are partly predicated on the focal player’s motivation to compete. These findings offer several important implications for driving user engagement in online competitive settings and meaningfully advance our current understanding of the effects of status and past performance information on online engagement.

摘要：随着社交媒体和在线社区的参与量不断增长，内容审核已成为平台治理的重要组成部分。迄今为止，志愿者（人工）审核员一直是内容审核的核心力量。然而，基于志愿者的审核模式在实现规模化、有效性与可持续性方面面临挑战，因此，许多在线平台近年来开始引入基于算法的内容审核工具（即“机器人”）。当机器人被纳入平台治理后，志愿者审核员会如何在社区“监管”和“培育”工作上作出反应，仍不清楚。为了解这些日益普及的机器人审核员的影响，本文基于 Reddit 上 156 个社区（子版块）的数据开展了实证研究。通过一系列计量经济学分析，研究发现：机器人能够促进志愿者审核员的工作，激励他们处理更多帖子，这一效应在大型社区中尤为显著。具体而言，志愿者审核员在引入机器人后增加了20.9%的“社区监管”活动，尤其集中在带有主观性的规则上。同时，在规模较大的社区中，志愿者还会投入更多精力，提供更详细的解释与建议。值得注意的是，活动量的增加主要源于在主观监管扩展的同时，对“社区培育”投入需求的上升。此外，机器人内容审核的引入还提升了志愿者审核员的留存率。总体而言，本研究表明，将基于算法的内容审核纳入平台治理，有助于维持数字社区的可持续发展。

Abstract: With increasing volumes of participation in social media and online communities, content moderation has become an integral component of platform governance. Volunteer (human) moderators have thus far been the essential workforce for content moderation. Because volunteer-based content moderation faces challenges in achieving scalable, desirable, and sustainable moderation, many online platforms have recently started to adopt algorithm-based content moderation tools (bots). When bots are introduced into platform governance, it is unclear how volunteer moderators react in terms of their community-policing and -nurturing efforts. To understand the impacts of these increasingly popular bot moderators, we conduct an empirical study with data collected from 156 communities (subreddits) on Reddit. Based on a series of econometric analyses, we find that bots augment volunteer moderators by stimulating them to moderate a larger quantity of posts, and such effects are pronounced in larger communities. Specifically, volunteer moderators perform 20.9% more community policing, particularly over subjective rules. Moreover, in communities with larger sizes, volunteers also exert increased efforts in offering more explanations and suggestions after their community adopted bots. Notably, increases in activities are primarily driven by the increased need for nurturing efforts to accompany growth in subjective policing. Moreover, introducing bots to content moderation also improves the retention of volunteer moderators. Overall, we show that introducing algorithm-based content moderation into platform governance is beneficial for sustaining digital communities.

摘要：越来越多的公司开始部署高度逼真的数字人类代理（Digital Human Agents, DHAs），由日益逼真的人工智能（AI）控制，用于在线客户服务，这些任务往往由聊天机器人承担。本文通过四项主要实验，结合定量问卷、定性访谈、直接观察与神经生理测量的混合研究方法，考察了用户在使用DHA时的感知（可信度、亲和力以及合作意愿）与行为。实验使用了四种不同的DHA，其中包括两款商业化产品（结果显示尚不成熟），以及两款未来导向的版本（在实验中，参与者被成功引导相信由人类操控的数字人类实际上由AI控制）。第一项研究比较了用户在阅读描述后对DHA、聊天机器人和人工客服的感知，发现DHA与聊天机器人差别不大。第二项研究比较了用户在实际使用一款商用DHA与聊天机器人后的感知，大多数参与者报告称DHA的当前实现存在问题，要么显得“怪异”或“机械化”，要么难以顺畅交流。第三和第四项研究采用“奥兹巫师”（Wizard of Oz）设计：让用户相信DHA由AI控制，但实际上由人类操控。结果显示，参与者仍然更偏好通过视频会议与真人客服互动；但在控制视觉逼真度后，我们未发现DHA与真人在简单客服任务中的差异。研究发现，当前存在交流问题的DHA在亲和力上优于聊天机器人，但其他方面差异不大。而当DHA的外观表现和交流能力接近人类水平时，其与真人客服之间的差异消失。此外，研究还对“算法规避”（algorithm aversion）文献作出贡献，表明DHA的拟人化界面可能缓解用户的算法规避心理。

Abstract: Companies are beginning to deploy highly realistic-looking digital human agents (DHAs) controlled by increasingly realistic artificial intelligence (AI) for online customer service tasks often performed by chatbots. We conducted four major experiments to examine users’ perceptions (trustworthiness, affinity, and willingness to work with) and behaviors while using DHA via a mixed-method approach with data from quantitative surveys, qualitative interviews, direct observations, and neurophysiological measurements. Four different DHAs were used in our experiments, which included commercial products from two different vendors (which proved to be immature) and two future-focused ones (where participants were successfully led to believe that the human-controlled digital human was controlled by AI). The first study compared user perceptions of a DHA, a chatbot, and a human agent from a written description and found few differences between the DHA and the chatbot. The second study compared perceptions after using a commercially available DHA and a chatbot. Most participants reported problems using a current production implementation of DHA, either finding it uncanny or robotic or having trouble conversing with it. The third and fourth studies used a plausible future-focused “Wizard of Oz” design by informing users that the DHA was controlled by AI when it was actually controlled by a human. Participants still preferred a human agent using video conferencing to the DHA, but after controlling for visual fidelity, we did not find evidence of differences between the human and the DHA. Current DHAs that have communication problems trigger greater affinity than chatbots but are otherwise similar to them. When the DHAs’ representation and communication ability match human ability, we failed to find differences between DHAs and human agents for simple customer service tasks. Our results also add to research on algorithm aversion and suggest that the anthropomorphic computer interfaces of DHA might alleviate algorithm aversion.

摘要：随着信息技术（IT）在商业环境中的重要性不断提升，IT 知识在劳动力市场需求侧（行业与企业层面）的关键作用已被充分论证。然而，关于 IT 劳动力供给侧的研究却相对不足，这使得 IT 专业人士如何战略性地培养自身知识结构以实现可持续且高回报的职业路径，成为一个重要挑战。本文通过实证研究填补了这一空白，探讨了不同类型的 IT 知识结构如何长期影响 IT 从业者的薪酬与职业安全。我们提出并量化了两个新的指标来刻画 IT 从业者的知识结构：知识通才度（knowledge omnificence），衡量个体知识结构的广度；知识差异化（knowledge differentiation），衡量个体知识集合与同侪之间的差异程度。基于 2000–2016 年 IT 从业者的职业数据分析，结果显示：平均而言，较高水平的知识通才度或差异化均能带来积极的经济回报。但这种正向关系并非单调：最佳策略是在适度水平上同时具备通才度和差异化。进一步研究发现，不同目标对应不同知识路径：若追求更高薪酬或职位，应着重提升知识通才度；若重视职业安全，则应侧重知识差异化。这一结果与动态能力框架及无边界职业理论的综合逻辑相一致。此外，研究还揭示了知识结构在缩小性别差距方面的积极作用：女性从业者因知识通才度每提高一个单位，薪酬平均提升14.74%；知识差异化每提高一个单位，薪酬平均提升1.38%。本研究对 IT 从业者、企业和政策制定者具有重要的管理启示，强调了战略性管理 IT 知识结构的重要性，帮助 IT 专业人士在动态且竞争激烈的劳动力市场中实现可持续发展。

Abstract: Amid the growing importance of information technology (IT) in the business landscape, the pivotal role of IT knowledge on the demand side of the labor market, at both industry and firm levels, is well documented. However, the important labor supply side concerning IT workers has remained largely unknown. This raises challenges about how IT professionals should strategically cultivate their IT knowledge structures toward a sustainable, well-compensated career path. This paper bridges this gap by examining how different types of IT knowledge structures of IT workers impact their salaries and job security over time. We theorize, define, and operationalize two new metrics to characterize the knowledge structures of an IT worker. Knowledge omnificence measures the breadth of an IT worker’s own knowledge structure, whereas knowledge differentiation assesses the extent of difference between one’s knowledge set and those of their peers. By analyzing extensive career data of IT workers from 2000 to 2016, we demonstrated that, on average, a high level of IT knowledge differentiation or omnificence yields positive economic returns for IT workers. However, there is an intriguing twist: such a positive relationship is not monotonic. The most advantageous strategy is to acquire IT knowledge at moderate levels of knowledge omnificence and differentiation. Further, our results revealed another new twist: to increase salary potential or pursue a better position, one should aim for knowledge omnificence, whereas those valuing job security should aim for knowledge differentiation. This aligns with our theoretical rationale that utilizes a structured framework, integrating the dynamic capability framework and the boundaryless career theory. Besides, we found that both knowledge omnificence and differentiation reduced gender disparity in the labor market. In particular, females benefited more, with a 14.74% (or 1.38%) increase in salary, from having a one-unit increase in knowledge omnificence (or differentiation). This study holds critical managerial implications for IT workers, firms, and policymakers. It emphasizes the importance of strategic management of IT knowledge structure in enabling IT workers to thrive in the dynamic and competitive IT job market.

摘要：语音购物助手（Voice AI），如亚马逊Alexa与阿里巴巴天猫精灵，作为一种新的线上购物渠道，正在全球范围内迅速普及。本文利用全球最大电商平台阿里巴巴的大规模消费者购买记录档案数据，实证研究了消费者采纳天猫精灵后对其消费行为的影响。结果显示：在采纳语音AI后的前四个月内，平均消费者在电商平台的周消费额增加了16.6%。进一步分析了重复购买、产品可替代性与熟悉度等产品特征如何调节天猫精灵采纳的效果，结果支持“语音 AI通过降低信息获取成本发挥作用”的机制。长期来看，天猫精灵的积极作用在重复购买上仍然显著，但效果会随时间衰减。此外，分析发现，语音渠道平均对PC渠道消费具有正向溢出效应，但对移动端消费没有显著影响。这种渠道间的动态效应依赖于具体的购物情境。本研究表明，具备购物功能的语音AI设备能够促进其所依附电商平台的增长。作为首个实证检验语音AI采纳对电商消费影响的研究，本文为电商平台与零售商在利用语音购物助手方面提供了重要启示。

Abstract: Voice-activated shopping assistants (voice AI), such as Amazon Alexa and Alibaba Tmall Genie, have been gaining popularity worldwide as a new channel for online shopping. In this paper, we analyze large-scale archival data of consumer-level purchase records from Alibaba, the world’s largest e-commerce platform, to empirically investigate how consumers’ adoption of Tmall Genie affects their consumption. The results show that the average consumer’s weekly spending the e-commerce platform increased by 16.6% within the first four months after adopting voice AI. Additionally, we explore specific product features that moderate the effect of Genie adoption by examining repeat purchases, product substitutability, and familiarity, supporting a mechanism that involves reducing information acquisition costs. The positive effects of Genie adoption remain significant on repeat purchases in the long term, although they attenuate over time. Furthermore, our analyses reveal that on average, the voice channel has a positive spillover effect on spending on the PC channel but no significant effect on the mobile channel. The channel dynamics are contingent on specific shopping contexts. Our results demonstrate that voice AI devices with shopping capabilities can enhance the growth of the affiliated e-commerce platform. As the first study to empirically examine the impact of voice AI adoption on e-commerce consumption, our paper provides valuable implications for e-commerce platforms and retailers leveraging voice-activated shopping.

来源：浙江工商大学数字创新与全球价值链升级研究中心

本篇文章编辑：林臻杰 中国人民大学信息学院 

声明：本文版权归原作者及原出处所有，内容为作者观点，并不代表本公众号赞同其观点及对其真实性负责。如涉及版权等问题，请及时与我们联系，我们立即更正或删除相关内容。本公众号拥有对此声明的最终解释权。