In recent years, enterprises have increasingly recognized the pivotal role of external users in driving product innovation. Open innovation platforms (OIPs), which facilitate interactions between companies and external innovators, have emerged as critical conduits in this regard. However, OIP managers face the challenge of motivating innovators to sustain their contributions. While some OIPs have implemented material incentives, the impact of such rewards on users’ ongoing innovation efforts remains uncertain. This study utilized a large-scale dataset from an OIP to examine how performance-contingent material rewards influence the subsequent behaviors of online innovators. Employing a quasi-experimental design involving propensity score matching (PSM) and difference-in-differences (DID) analysis, we found that receiving performance-contingent material rewards led to a decrease in the quantity of subsequent ideas generated by innovators. However, these rewarded innovators produced ideas of higher quality. Interestingly, the novelty of ideas submitted by innovators declined following their receiving of rewards. Moreover, newly enlisted innovators exhibited a more positive response to these incentives. Our findings provide valuable insights for platform managers seeking to optimize incentive mechanisms. We suggest adopting diversified incentive approaches and refining incentive strategies to effectively motivate continuous innovation among users on OIPs.
22 February 2008 in Communications in Statistics - Theory and Methods
In this article, we extend a class of semi-parametric density estimators to time-series context. The asymptotic theory and simulation study are discussed. Theoretical results and numerical comparison show that in the time-series case, the estimators in this class are better than, or at least competitive with, the traditional kernel density estimator in a broad class of densities.
12 October 2009 in Communications in Statistics - Theory and Methods
In this article, we extend a semiparametric regression estimator with multiplicative adjustment to time series context. The asymptotic theory and results from a simulation study are discussed. Theoretical results and numerical comparison show that, in the time series case, the semiparametric estimator is better than the traditional local polynomial estimator in a wide neighbourhood around the true regression function.