Show me the numbers. Causality in multilevel HRM, employee attitude and organization performance relations

D. Madinabeitia (Speaker), Kroon, B. (Speaker), U. Elorza (Speaker), de Reuver, R. S. M. (Speaker)

Activity: Talk or presentation typesOral presentationProfessional


The mutual gains thesis holds that investing in sophisticated human resource management leads to both socially and financially beneficial outcomes. However, testing its evidence requires investigating multiple data levels (employee outcomes and organization performance in a large number of organizations) (Gelade & Ivery, 2003) in a preferably longitudinal research design (Wright, Gardner, Moynihan, & Allen, 2005). Few studies have both the multilevel and longitudinal data required to allow such analytics (Madinabeitia, 2016). The current paper builds on an ongoing big data project by Mondragon University aimed at helping organizations to use employee survey data to improve the social and financial performance of organizations. The current state of data allows analytics to test (quasi-) causality in multilevel mediating processes in the HRM and organization performance relationship. Relying on survey data of 9224 employees in 59 industrial organizations in Spain combined with longitudinal data on past (-1/-3 year before the employee data collection) and future (+1 year after) financial indicators of organization productivity, we performed quasi-longitudinal multilevel structural equation modeling tests (MSEM) on the mediating role of employee attitudes (satisfaction and commitment) in the relationship between high-involvement work systems (HIWS) and organizational productivity. Findings indicate that the positive relationship between HIWS and +1Y organization performance is fully mediated by employee positive attitudes, and that although previous organization performance explains a large amount of +1Y organization performance, employee attitudes incrementally explain variance of +1Y organization performance. In contrast to previous research (Shin & Konrad, 2014), the relationship between previous organization performance and HIWS was insignificant. The main limitation of the current data is that the employee data are still cross-sectional. Future rounds of data collection will enable more sophisticated causal modelling. The key theoretical implication is the support for the mutual gains thesis. The project also highlights the challenges to obtain true longitudinal multilevel data on human resource management. Projects like this one at Mondragon stress the importance of project management support structures and funding.
Period9 Nov 2017
Event title10th International Conference of the Dutch HRM Network
Event typeConference
LocationNijmegen, Netherlands
Degree of RecognitionInternational