Continuous time modeling of panel data: SEM versus filter techniques

• Published in: Statistica Neerlandica Vol. 62; no. 1; pp. 4 - 28
• Main Authors: Oud, Johan H. L., Singer, Hermann
• Format: Journal Article
• Language: English
• Published: Oxford, UK Blackwell Publishing Ltd 01.02.2008; Netherlands Society for Statistics and Operations Research
• Series: Statistica Neerlandica
• Subjects: Comparative studies; damped linear oscillator; exact discrete model; Kalman filter; Mathematical models; Maximum likelihood method; Monte Carlo simulation; random subject effects; state-space model; structural equation modeling
• ISSN: 0039-0402; 1467-9574
• DOI: 10.1111/j.1467-9574.2007.00376.x

Although convincing arguments have been put forward for continuous‐time modeling, its use in panel research is rare. In one approach, classical N = 1 state‐space modeling procedures are adapted for panel analysis to estimate the exact discrete model (EDM) by means of filter techniques. Based on earlier less satisfactory indirect methods, a more recent approach uses structural equation modeling (SEM) to get the maximum likelihood estimate of the EDM by the direct method. After an introduction into continuous‐time state‐space modeling for panel data and the EDM, a thorough comparison is made between the two distinct approaches with quite different histories by means of Monte Carlo simulation studies. The model used in the simulation studies is the damped linear oscillator with and without random subject effects.