Stata

Commentaires: There are some things that Stata does extremely well - the statistics and econometrics in particular. And it does these things reliably and fast, with a good support network. If you're willing to pay for it, performance is even better - using up to 64 cores. But for more general purpose statistical work, R is not only free, but better (even though it is more buggy).

Avantages: This is statistical programming software that is excellent for canned econometric procedures. Reliable and fast, with a huge library of procedures built in, and an even bigger user-contributed library. Online support through the statalist forum is also amazing. It also has a matrix programming language (Mata). The graphics capabilities are also easy to use and very high quality. Has a lot of related features like heatmaps, producing production-quality regression tables, etc.

Inconvénients: This software is expensive. You're paying for something that is reliable, tried and true, with a great support network. But there are other free options, such as R. Furthermore, while this is excellent for canned statistical and econometric procedures, the programming language itself is not anywhere near as developed as something like R. For example, work with text etc is out of the picture. It is quite un-intuitive for someone coming with a programming background, and has weird quirks - such as only being able to keep on dataset in memory at a time. These issues mean that achieving what could be achieved in 3 lines of code in another language, might take 100s in Stata.

Commentaires: I would say that having worked with stata using different datasets have improved my econometrics understanding. It is a great choice for research.

Avantages: Stata is a statistical analysis software which is fairly easy to use. I believe It is the 2nd best software (after R) for learning econometrics. It has both coding and drop down interfaces. Easy to learn and there are ample of notes on internet which can be used in the learning process. Fast and powerful enough to run econometric models (even runs tobit model, which R can not do. R returns infinite SE many times). Since it is a paid software, the results it provides are reliable and that is why it is widely used by econometricians for their research. Since most of the economists use stata, it has a network externality, i.e. people can understand the research methodology by going over stata codes.

Inconvénients: The reason I think it is the 2nd best is that it is not cheap. Also, it offers different licenses. The basic license costs 200 usd (approx). But it can work with only 2048 variables. For mor variables, you have to buy stata SE, which can work with 32,767 variables. Since most of the advanced public datasets has more than 2048 variables, it is difficult to manage with IC. So, they basically force you to buy SE. They should increase the number of variables to at least 10000 for the IC version.