I’ve been active in learning SAS for about a week. During this time, I found out something about SAS that really make it on top of statistical software for a long time. I want to talk about my feeling about it here, as a reminder.
First of all, I want to ask a question: what softwares would you think of when you heard other people talk about statistics? Answers may vary: R, Python, SAS, STATA, SPSS……. To be honest, I have never touched on one of them and that is SAS. Due to my major in undergraduate study, I used STATA quite a lot and even my whole paper counted on it. Additionally, I have tried to broaden my vision about statistical software, so I have installed and played with R Studio and SPSS before. Except for the various capabilities of R, I am not stunned by any kind of statistical softwares. At least that’s my initial impression. I love python very much, but I have not much experience in the statistics-related packages such as Numpy.
Until I am forced to learn SAS can I find out the most interesting point in SAS — it separates the data input procedure and data analysis procedure with DATA and PROC command! I appreciate this setting very much because when the command file is becoming chunkier, this setting would help you recognise what you want to navigate to more efficiently. At first, it has some learning curve but once you get the hand of that, you will treasure the clarity it gives you.
When I wrote my undergraduate paper using STATA, I thought of it as a python-based statistical software because both of them have similar coding logic. It is simple but is still not easy for a non-programmer to understand at once. As for R, it simplifies a lot of the coding command and makes it easier for statistician to code, but it has way harder learning process than SAS does because it is more versatile and contains a lot of unnecessary packages for researcher. It’s true that people have different preference on their tastes, so that’s also why there are many similar but different tools in one area.
Frankly speaking, the coding structure of SAS is not as efficient as that of R, for example, I have to type “RUN;” to make it run each time. What I like about is the designing of its logic that separates the data input and analysing process. Just think about it, what if Python could open every type of data files using “DATA” command, it would be of great convenience!