Having decided to become a researcher after long career in industry, should I work independently or start a PhD?

Question

I'v been doing software development work for almost a decade. I really love doing research work but I am always doing programming work in companies. I could join a PhD program 15 years ago after my MS. degree in US. But I did not continue at that time for reasons (want more money, find a job for green card, etc.).

Recently I decided to follow my true heart on sciences, instead of being a "seasoned company guy" all the way to my retirement. The machine learning/deep learning/big-data are thriving in these a few years. I cannot sit here just watching this big revolution in computer science.

Money is always an issue but will not be a big issue for me. I have accumulated quite a bit of money in the past 10 years hard working. I can be mortgage free in a few years even in this most expensive city (Vancouver, Canada). My wife has a good job too. She “granted” me 2 years to do a starter’s non-profitable work on my own. Hear is the plan: Doing part-time work to earn half salary to support the living cost and the remaining mortgage. The rest of my band-wise will be dedicated to my research work. The question is, how to start?

Option 1: be an independent researcher and developer. Spend the next year to do online study including all the necessary math/statistics skills. Become an expert on certain field. And starting to publish papers, contribute to open sources or have some software inventions/products.

Option 2: apply a PhD program and spend the next 5+ years with professors to build up the foundation to become a truly powerful researcher. I am 42 now. Do I have that much time to spend?

Option 3 is to dedicate the first year to invent and develop a software product (iPhone App, Cloud service, etc.) and try to make money on it. And use the earning as the funding to support myself on the research work. In the mean while, try to get some investment and government funding.

None of the above will be an easy way. It is not a common career path, especially for a forty years old guy with a job/salary, wife and a kid...etc.. This will be a venture. I will spend years being dedicated /obsessed ( hopefully not half-crazy) in a house at the foothill of Canada north mountain, learning/reading/drawing/typing.....

Which way to go...?

Answer 1

A PhD is training for research. There's a whole lot of things that happen along the way, but the point of the process is that you gain skills in conducting research. In addition, you gain the 'worldview of a researcher': how to think critically, assimilate information, analyse data, collaborate and come up with new ideas.

It is, of course, possible to do all these things without doing a PhD. But the advantage of the PhD is it gives you opportunities that you might not otherwise have: access to resources, specific programmes (eg internships or field studies), people to collaborate with, exposure to new ideas. Those opportunities can be open to PhD students in a way that they aren't to random members of the public.

A big part of research is interacting with other people, not just working on your own single path. PhDs force much more interaction than you might do on your own (though it varies by subject).

The other side of 'training for research' is that, once you've come out of the process, you have simple proof that you have been trained. That's what the letters 'PhD' are code for. While the experience can be very different at different institutions, a PhD means the holder has untaken a major research effort and it has been validated by their peers, and learned some of the skills of research along the way.

That's the reason that many academic jobs expect you to have a PhD: they don't have time for a free-form assessment of your research skills - the PhD is proof that somebody else has already done that assessment.

By doing research without a PhD, you force people you interact with to do more work to assess your skills. You can do that by building up a good body of peer reviewed research, which itself is unlikely to care whether you have a PhD (especially if the review is blind). But in superficial contexts (eg screening for a job, a grant application or an invitation to some event) you're likely to find yourself at a disadvantage.

One advantage with working inside academia is you have more collaborators, but also as you become more senior you can build a research team: when students work for you, you (collectively) can achieve more. That's difficult to do when you're working independently.

Finally, we can all come up with examples of brilliant people who started working on their subject without a PhD. I suspect that's much less common than it was in the twentieth century, for two reasons:

1. Supply and demand: there were many fewer researchers, and the process was much more informal. Now every tenure job gets hundreds of applicants. Why would you pick one missing critical training?
2. War and social upset: when you have to work on radar/antibiotics/aeronautics/nuclear physics/intelligence/cryptography at short notice for the war effort, all employment norms go out the window. When those people came back to civilian life, they already had skills they could use in research. Or, alternatively, they proved their worth in opportunities available because everyone who would normally take them was away fighting.

(Side note: I come from the UK where PhDs are 3 years on paper and often more like 4 in practice. The North American process seems like a much more drawn out affair, and the time tradeoffs may be different)