[mattyb]

From Wired : The End of Theory: The Data Deluge Makes the Scientific Method Obsolete

Really interesting.

[dcmacdaddy]

Originally Posted by Wired News article

Correlation supersedes causation, and science can advance even without coherent models, unified theories, or really any mechanistic explanation at all.

I have a hard time accepting the simplifying nature of this summary statement from the article. What does it mean when it says "correlation supersedes causation"? By the overall tone of the article it seems to me that the article is saying that the value/usefulness of data is to be found only in its indicative role and not in its predictive or explicatory role. (e.g.: a large enough data set is indicative of a trend or set of behaviors but the data set does not appear "coherent" to provide an explanation of why the trend or set of behaviors is occurring.)

For me, noting that something is happening is not nearly as useful as understanding "why" something is happening. This article seems to suggest that the "why" question is losing relevance . . . and that bothers me.

[Dakar the Fourth]

Originally Posted by dcmacdaddy 

This article seems to suggest that the "why" question is losing relevance . . . and that bothers me.

That may be true within the general masses, but for any kind of person who is still capable of critical thinking, that question does not lose any relevance.

[mattyb]

I'm not a scientist, just a simple IT worker, but maybe this (the "why" question losing relevance) is a side effect of the 'time is money' world we live in at the moment.

For me, the results of J. Craig Venter show that the 'why' type of investigations are still required, but what is interesting is that his gene sequencing analyses discovered previously unknown bacteria. Basically mining the data, he found things that he wasn't even looking for.

Use this info, dig deeper and the scientists/researchers can get into the 'why', while the people operating in the 'real' world can put this information to use.

Just MHO though.

[mattyb]

The Ars Technica opinion.

[OreoCookie]

Nonsensical article.
He claims that the quest for correlations will replace the search for causal relationships. Most of the time, correlations in data point to new phenomena or mechanisms. Just because we have better tools than ever to look for correlations, doesn't mean we replace the search for causes.

Furthermore, you have to know what to look for in your data: without some idea what is going on, you will likely see nothing. I don't deny that many interesting scientific discoveries were made accidentally, but that happens at the beginning of new discoveries. When you do research in a mature field of science, you have a good idea (a mixture of intuition, experience, knowledge) what you will likely have to look for. So, in essence, I doubt you will be able to even find the important correlations when you don't have an idea of the mechanism (= `causes'). His examples (e. g. sequencing the genomes of various species) does not support his point at all. These large collections of data are the basis for further analysis, it provides scientists with the means rather than the ends.

Even if some scientists `only' gather high-quality data, e. g. measuring fundamental constants with higher and higher precision, then this is also not as easy as the author makes it sound. He just sees the end result (the numbers and graphs, so to speak), but not the effort that went into it. Sequencing the human genome is simple in the sense that you can explain it rather easily to even laymen. But to actually do it was a concerted effort of many talented individuals and experts in many, many fields.

The fact that our knowledge is exploding exponentially gives lots of challenges to scientists and society alike. Very often, only a handful of people understand all particular details of a given theory or technique. Islands can form independently where people do `the same thing twice'. The world just is a lot, lot more complicated than during Newton's lifetime.

[Tiresias]

Originally Posted by mattyb 

From Wired : The End of Theory: The Data Deluge Makes the Scientific Method Obsolete

Really interesting.

The irony is the whole concept of Data Deluge is itself a new model of the scientific method.

[Cipher13]

Poor Wired.

It's a good article. It just underscores the public's search for easy answers. The public and media fall for these non-sensical correlation studies all the time. Especially when they pertain to small children. Panic over BPA in bottles, autism after vaccinations, and others have convinced parents to do stupid things without any supporting evidence.

[Kerrigan]

Reading that article is like eating the re-heated leftovers from last night's expensive feast with Lyotard and Baudrillard.

That is to say, the author displays a lot of chutzpah in boldly declaring the end of scientific theory, when postmodernist thinkers have been elucidating a similar but much more profound discourse for decades. But despite decades of postmodernist writing, the new electronic world that we live in has not destroyed the old axioms of analytical philosophy, and the profusion of data has not lead to the formation of an unintelligible hyperreality.

Yes, there will always be Humean doubts about the nature of human reason and the existence of such things as causation, but the scale of data available to mankind does not somehow confirm those doubts. There will always be convincing arguments for both skepticism and metaphysicism.

[analogika]

Silly premise.

Data is not knowledge.

Data has no meaning in and of itself.

Simply amassing huge amounts of data means absolutely nothing until *somebody* imbues it with sense by applying structure, by asking questions and proposing answers and USING that data for research.

Accidental discoveries are just that.

[Tiresias]

"Where is the wisdom we have lost in knowledge? Where is the knowledge we have lost in information?"
—T. S. Eliot

And where is the information we have lost in data?

[lpkmckenna]

Originally Posted by Kerrigan 

That is to say, the author displays a lot of chutzpah in boldly declaring the end of scientific theory

Whoa, that's a highly erudite way of saying he's an idiot.

[mattyb]

Originally Posted by Tiresias 

And where is the information we have lost in data?

VERY good question !!