Ridderprint

Discussion 163 models would update following changes in the underlying relationships; or what acceptable levels of predictive performance metrics are for different HRM processes. Moreover, scholarly attention is required on the ethics surrounding machine learning and decision-support systems in general. Where are the ethical boundaries in exploration through experiments with HRM practices? What are the implications of (not) including conscious or random learning experiences in predictive systems? What do we consider an optimal tradeoff between exploration and exploitation in light of a long-term balanced perspective? HRM scholars should get on this bandwagon and make sure that the developments related to machine learning in HRM contexts follow a balanced approach. 7.5.2 Expatriate Management The field of expatriate management research may struggle with the inherent characteristics of expatriation. Expatriates are hard to come by and can differ quite considerably from one another in terms of their personal characteristics (e.g., host/home country, nationality, personality) and their assignment and organizational context (e.g., duration, type, hardship, mobility support). Fortunately, I see two avenues by which a people analytics approach may help expatriate management research overcome these challenges. First, technological advances may help to circumvent sample size issues by facilitating within-person research. For instance, experience-sampling methods (ESM) gather intensive repeated assessments with brief intervals and study durations (Beal, 2015). Such study designs have the potential to capture the expatriate experience on a much more granular level compared to repeated or cross-sectional surveys while being relatively simple to implement with the current technological developments. ESM may help to capture a wide range of expatriate experiences, close to the moment of occurrence, as they are experienced in real-life (Beal, 2015). Concrete examples of how ESM could improve our knowledge of the expatriation process lie in an investigation of how home and host country social networks change over time, how the perceived organizational support and psychological contract change during the expatriation cycle, or in uncovering the true shape of the adjustment curve. Alternatively, text mining methodology and sentiment analysis can generate rich data regardless of the sample size. For instance, scholars could mine interview data to explore what themes and concepts are considered to contribute to or impair expatriate experiences. Here, scholars could combine old and new interview data to compare what topics are mentioned in relation to international assignment and whether these have changed due to advances in communication technology or the rise of alternative assignments and dual-career families. On a comparative level, the textual content of the global mobility policies of different companies could be processed and analyzed to explore where differences exist and what these imply for assignment success. Moreover, social network analysis may be interesting in light of expatriation and sample size. Here, scholars could explore to what extent the social networks of assignees change during and after the assignment. Rich information could be gathered in relation to the home and host country networks, and both in the work and personal atmospheres.