ABSTRACT
In some sociological applications, substantive theory or previous empirical research suggests a specific shape of time-dependence for the transition rate. When this time path can be modeled in terms of a tractable waiting time distribution, the likelihood can be calculated by using its density f(t) and survivor function G(t), as has been shown for the exponential model in section 4.1. An obvious estimation strategy is then to use any computer program or software package (e.g. BMDP, GLIM, GAUSS, LIMDEP, SAS, etc.) that offers function optimization routines and to maximize the likelihood function with respect to the unknown parameters. 1 This strategy is very flexible because researchers are able to build their own parametric models, but it requires mathematical and programming skills. Our experience however shows that applied social scientists normally do not want to be bothered with writing programs and maximizing likelihood functions. TDA therefore offers the opportunity to estimate widely used parametric models with a single command. The parametric models with time-dependence can be requested with the TDA command https://www.w3.org/1998/Math/MathML"> rate ( p a r a m e t e r s ) = m o d e l _ n u m b e r https://s3-euw1-ap-pe-df-pch-content-public-p.s3.eu-west-1.amazonaws.com/9781410603821/c161f26d-9afd-4c9a-b556-d0a6b36654d2/content/Eqn150_B.tif" xmlns:xlink="https://www.w3.org/1999/xlink"/> where model.number is one of the possible types shown in Box 7.1. For most models all parameters that can be specified in the rate command are optional.
