regarding the software, the idea is that choose some, and use it. But in practice, you must consider some other elements. Such as the fact that my supervisor, my supervisor's friends, my supervisors' previous phd students, they choose, GUASS, so for me it is always optimal to use GUASS.

Take an example, I got a set of GAUSS source code of kernel density of nonlinear grids which is part of my current topic. So It depends on your own situation.

Another critical difference between using packaged software and eay-to-program one is that sometimes you may think that you have understood the econometric model and put the data into Eviews and click to get the results.

In many cases, it turns out that it is not. For example, In Eviews with EGARCH model, Eviews takes a shor-cut in hanling the expectation of the absolute z(t). If your intereste lies in the alpha and beta's, their size, their sign and trying to give some economic intuition, then it is OK. On the other hand if you are going to simulate the model after parameter estimation, this subtle point becomes noticeable.

I still insist that for Chinese master student (of course good student and good univ) in 计量经济学, it pays off to make their own code to implement the econometric model since they have ample time of study.

漠中麾 wrote: ..东财数量经济的，.., 我们用的软件是EVIEWS.

It depends on your major and filed.

At my current deparment(Economic Statistics), none of the prof. s and PhD students use eviews and the like. Some people may need to learn only because they need to teach the undergraduate student.

(sorry for 漠中麾, i am not blaming anyone, just trying to provide some info.)

多谢zhaoswenden赐教。但还是要澄清一个事实：用哪种软件只是个人偏好以及软件的实适用性。一个应用经济学的研究者往往不止用一种计量软件。

如果精通SAS，我想就没必要去研究其他软件了吧`