Continuous-time modelling of irregularly spaced panel data using a cubic spline model |
| |
Authors: | Sy-Miin Chow,Guangjian Zhang&dagger |
| |
Affiliation: | University of North Carolina, CB#3270 Davie Hall, Chapel Hill, NC 27599-3270, USA; University of Notre Dame, 108 Haggar Hall, Notre Dame, IN 46556-5611, USA |
| |
Abstract: | Continuous-time modelling remains a somewhat 'idealized' representation tool. Even though conceptualizing a dynamic process as a continuous process has clear appeal from a theoretical standpoint, practical tools that allow researchers to effectively map an idealized continuous model onto a set of discrete-time observed data are still lacking observed data. Irregularly spaced longitudinal data frequently arise in empirical settings because of the prevalence of longitudinal studies with partially randomized measurement intervals and other related designs. We present a practical approach that capitalizes on a nonparametric spline interpolation approach to impute the gaps in irregularly spaced panel data. Simulated and empirical examples are provided to demonstrate the applicability of the proposed approach to studies of group-based dynamics using panel data. |
| |
Keywords: | stochastic differential equation state-space modelling Kalman filter smoothing exact discrete time panel data |
|
|