The metaphor is extremely helpful for deciding when and how to learn a new stat package, and to keep you going when the going gets rough.
- Whether you need to learn any particular package/language depends on your context.
Knowing Icelandic is a huge advantage if you need to communicate with people who only speak Icelandic. Otherwise, not so much.
Likewise, there are specialized packages that are a huge advantage to people working in very specific research areas. If you’re never going to work in that area, you don’t need it.
- Learning the second one is easier than the first.
Like languages, stat packages have an underlying logic and an underlying grammar. Once you’re used to how these work in one, you know what to look for in another. It’s always easier to learn the second one.
- It’s really helpful to get training that takes into account how people actually use the language/package.
Ever try to learn French from a French dictionary? Of course not. If you try to learn SAS from a manual, you’ll be frustrated.
Manuals and dictionaries are excellent reference books and can assist your learning, but they’re not designed as tutorials.
Get good training at the beginning, and you’ll be well set up for self-study later on.
- Each one has its strengths.
There are expressions and words that only exist in certain languages. There are supposedly dozens of words for snow in Inuit, but (and I’m only guessing here) probably not many words for crocodile. Some words are important in specific contexts.
If you become a tropical herpetologist, you are going to need a language that includes ‘crocodile.’
Likewise, every general package has certain areas it does very well, and others, not so much. Being able to easily whip off the statistical tests you need often is a huge advantage.
- It’s always helpful to learn more packages/languages.
Do you need to learn more than one language? Certainly many people get away with only speaking one their entire life.
But if you live in a place where the language you learned at home is not what most people use, you’re going to be limited. In that situation, learning another language is extremely important. (If you’re the only R user where everyone else uses Stata, it gets hard to ask colleagues for help or share results).
But even if your first language works for everyday life, having more skills opens you up to many more opportunities and levels of effectiveness.
So being able to use R when you need to is good, even if you and everyone around you, usually uses Stata.
- Understanding more that one language/package teaches you about how languages/packages work.
It’s hard to see the bigger context of how grammar works when you’ve only experienced one language. Likewise, you start to understand the import of defaults, options, and syntax rules when you see them done two different ways.
- You don’t need to be fluent in a package/language in order for it to be helpful, but you do need a basic understanding.
There is a turning point in your learning before which you’re just not going to be able to use the language/package with anything but frustration. You just won’t have enough vocabulary and basic ability to use the grammar without looking everything up.
But once you’re past that point, you can do a lot, even before you’re fluent. More importantly, you’re poised to learn even more very quickly.
- You won’t really learn a language or a package unless you practice it.
You know this. You’ve got to practice.
- Immersion can be overwhelming, but a great way to learn quickly.
If you keep going back to the language/package you know well, you won’t make as much progress with a new one.
- No one language or package is the best one that everyone should use.
It would sound ridiculous to say that one language is the best.
The same is true for stat packages. There are many great packages.
Yes, they each have great advantages, and the advantages of a particular package may be just the ones you need. Yes, it’s the best for you. But not necessarily for someone else working in a different area, with different needs.