How do I learn to think like a scientist? How do I form good hypotheses?

[ygolo]

This is perhaps a philosophical question, but I am trying to make the transition from engineer to scientist, and I am trying to relearn how to think and ask questions.

As an engineer, a lot of times, we can get away with making something that consistently worked without understanding it.

A common scenario where I worked earlier was that the chips we designed were such a limited run, that we could simply run a screen that tested functionality that we needed long enough despite having very low yield. We didn't have to publish a guarantee, and if a chip failed, we could replace it. The customers would not been around if we waited for a good understanding of what the issues were, but were perfectly happy with us replacing parts when they failed.

Here, just finding the maximum-likelihood Poisson distribution seemed good enough (and perhaps even overkill).

It seems to me that some of the best scientists are the ones good at forming hypothesis. But a lot of time I see people advocating that we should let the data speak for itself, and that we ought to make no hypotheses, and let the data tell us what is true.

However, this sort of statistical thinking seems to lack something when it comes to forming and testing the types of hypotheses important in the science.

I could be wrong.

But if I am, I was wondering: if we pretended we didn't know Newton's laws, using inferential statistics, would it be possible to infer F=m*a directly from the data (IOW, without an apriori hypothesis)? What experiments would we do? What would we measure, and what sort of statistical analysis would we do?

 

[oddjobmj]

I can't really comment on how to think like a scientist as, even though I have an idea, it's definitely not the insightful or 'deep' understanding that I believe you're looking for.

As far as how we could inferm something like F=m*a from data we would simply gather a bunch of data and included in this set would be F, m, and a data points. By chance, of course, the algorithm we design will find a strong correlation and hint to a scientists to check it out. The algorithm could tell us how strong the correlation is and account for bad data.

However, there are vast numbers of data-types to consider and each one would have to be compared to one another which means a fantastic number of required calaculations. Even if there are only 100 types of variables (data types) to consider that's 100 factorial or 10^157 comparisons. That's also assuming the data is ideal and there are no other situational changes for each data type.

The above might be a good start but I'm also not a CS major so there are very likely better ways to go about this type of problem. Also, I do believe we have narrowed down the possibilities and would be providing specific data in a 'smart' way instead of brute-forcing all possibilities.

 

[Lamarr]

you cannot learn to think like a scientist. Either you already do, or you don't.

great scientists don't have special abilities. They just questioned everything around them, sought to understand this universe, and had a passion for knowledge.

The difference between scientists and laymen is that laymen simply don't give a damn. Theyre not curious, and don't question.

 

[Lamarr]

even some professional researchers aren't really scientists, they're more like science students. They regurgitate what they learned from textbooks, and are satisfied with a basic understanding. Sometimes they give lectures without fully understanding what they're lecturing about.

A real scientist is never satisfied, no matter how much knowledge you throw at him.

 

[Devils]

Which part of Bloom's taxonomy do you work at?
http://www.nwlink.com/~donclark/hrd/bloom.html

Also try lateral thinking
http://www.mindtools.com/pages/article/newCT_00.htm

At school from first principles I worked out the angle of bonds in methane, worked out the maths for perspective transformations, designed software out of interest. Essentially you have it or don't.

 

[Andy Resnick]

<snip>
As an engineer, a lot of times, we can get away with making something that consistently worked without understanding it.
<snip>

You are exactly correct. The difference between Engineering and Science is that Science is more concerned with 'understanding' rather than 'consistently worked'. Good scientists and good engineers make excellent teams, I have been fortunate enough to be a part of a few.

 

[maxwellic]

My answer when I read the title was to become an engineer! Engineers need to ask enough questions to be able to "get away with making something that consistently worked without understanding it". :)

 

[JakeBrodskyPE]

One of my mentors used to say that Engineering often works by habit. You find something that works, and you change it as little as possible to meet the goals set before you. Only when it becomes clear that doing the same old stuff is no longer going to work should you consider a completely different approach.

Science is probably a lot like that too. Using Occams Razor (Use the answer with the fewest assumptions), they plod on until experimental evidence can no longer be explained by the theory.

Both have invested enormous resources getting to where they are today. We do not casually discard those lines of design or theory until it is absolutely essential. Why? Because we want predictable results. Radically different theories, and radically different designs may have unusual and perhaps unknown or untested limits. They are not known to the degree that the older theories and designs were. The goal is to develop a body of design or theory that is known, accurate, and predictable.

Yes these are very conservative (in a classic sense) fields of study. While journalists love to discuss radical new ideas, such as string theory, the reality is that until there are explicit and exclusive experimental results that justify this theory over others, it's just a lot of interesting concepts or hypotheses.

In my humble opinion, that's how Real scientists and engineers think.

 

[chiro]

Science is all about decentralized knowledge discovery, and the paradigm of decentralization is critical to understanding the scientific method.

If a scientist hears a claim, they will test it themselves or they will read widely, do their own thinking, and come to some conclusion.

There is a lot of disagreement, competition for data, theories and result but the important thing to remember is that it is decentralized.

The other important thing to remember is that it is done under uncertainty.

The uncertainty comes from the fact that the information is incomplete and that you should never rely only on a single source, so read widely and get a lot of different viewpoints even if some of them seem strange and out there: diversity is a good thing.

Finally become adept at mathematics particularly statistics.

Mathematics is the only language that every kind of culture can agree on: it transcends cultural, social, and any other bias. People use mathematics widely because everyone can agree on it: it is impersonal for a reason and science uses it because it has the same kind of decentralization properties that science does.

Like science, mathematics allows anyone to check something for themselves. It allows people to talk on a language that has a level playing field which means it can be challenged and critiqued across the board. Once the barrier of learning the initial language, terminology and accepted structures have been passed, everyone can join in and everyone is able to look at things independently. It also doesn't care if you are right or wrong or what your name is.

Statistics and probability are the mathematical languages of uncertainty and this is why scientists use them.

The final piece of advice is that if you use mathematics to draw your own conclusions or if you read other peoples conclusions based on mathematics, then learn what all these assumptions and results mean. If you don't know what something means then either come back to it later when you do, or go to something else that is easier to digest.

Don't blindly just believe anything: always think about what is going on before you make any conclusion. If there ever was an attribute that a great scientist ever needed it was an ability to think. Also this applies to everyone no matter who is saying it: you will have to draw the line when you choose what to accept and reject and that's everyones right as a human being, but always keep in the back of your mind what these reasons are.

If you are willing to suspend your disbelief even slightly when listening to something else that is even partially contrasting to your own views or experience, then you may find that you can learn something new quite frequently.

 

[scout6686]

Very interesting and important thread in many ways. Much of "lower" sciences course (eg High School) miss the philosophy of science and theory behind it. Facts are not treated as questionable but rather is largely rote, I think, this gives the wrong impression of what science is about. Many people are turned off by this approach early on, however maybe it is like a Spanish teacher teaching Spanish when they are hardly bi-lingual (or not native). Anyways, I digress.

 

[JakeBrodskyPE]

Very interesting and important thread in many ways. Much of "lower" sciences course (eg High School) miss the philosophy of science and theory behind it. Facts are not treated as questionable but rather is largely rote, I think, this gives the wrong impression of what science is about. Many people are turned off by this approach early on, however maybe it is like a Spanish teacher teaching Spanish when they are hardly bi-lingual (or not native). Anyways, I digress.

Much of this is like learning music theory. You can make music without it, but few will take you seriously unless you can read and write music; and describe timing, scales, and chords.

To participate in science, you have to at least understand what is "known." This includes notation, experimental evidence, and the like. Questioning the facts and relationships before the whole framework is understood is poor practice at best. That's why so many scientists have to drink from the firehose of knowledge and experience just to get to the edge of what is known.