Sharing Results is the Engine of Scientific Progress

by David Wojick in Products and Content | Comments [8] Tags: knowledge mission research-results

OSTI's mission is to help scientists share their results, but what role do results play in science? Here we present a simple model of one of the most basic uses of results, namely as the engine of scientific progress. Research results are more than just accumulated knowledge. Research results make possible new questions, which in turn lead to even more knowledge. The resulting pattern of exponential growth in knowledge is called an issue tree. It shows how individual results can have a value far beyond themselves, because they are shared and lead to research by others.

The reader is referred to the Sharing Results Issue Tree. This is an abstract example of a fundamental pattern that occurs throughout science. It begins with Result 1, which is an important finding by a researcher named Smith. Given this result there are three important new questions that can be formulated -- Questions A, B & C. It is important to realize that these questions could not have been asked until Result 1 occurred. Result 1 does much more than simply add to our knowledge, it raises important new questions.

Each of the three questions now becomes the object of new research. It is important to realize that in many cases this new research will be undertaken by researchers other than the one who got Result 1. This could not happen unless these new researchers know about Result 1, which requires sharing of results in some way or other. Thus sharing is essential for scientific progress.

The new questions that grow out of Result 1 yield Results 2 through 9. These new results are obtained mostly by researchers other than Smith, such as Brown, Gupta, Kim, etc. This is a large increase in knowledge, which is only made possible by the sharing of Result 1. Thus Result 1's value extends far beyond its contribution to knowledge. This is the essential point:

The value of a scientific result often lies, not in the knowledge it adds, but in the knowledge it leads to when it is shared.

Moreover, the process does not stop there. Note how one of these new results, Result 5, generates four new questions. One of these in turn yields two new results, Results 10 & 11. These too are a product of the sharing of Result 1. In the real world this sharing process often leads to whole new subfields and specialties. A single result, such as the discovery of the electron, can yield millions of important results by other researchers.

This pattern of progressive, spreading questions and answers is called an issue tree. The issue tree diagram provides a visual map of the progress of a little bit of science. But the underlying pattern is there in science itself, whether we map it or not. It is one of the fundamental patterns of scientific thinking.

Progress is not just the cumulative product of individual efforts, it requires sharing for its very being. We take this sharing for granted but it is by no means assured, and it is far from being efficient. The Internet promises to greatly improve the process of sharing scientific results, which should speed up progress. But this promise is still largely unmet. This is the challenge that OSTI is working on, how to speed up scientific progress by making sharing efficient.

David Wojick

Comments:

Readers may further note that the particular issue tree in the figure illustrates another important effect of sharing results. If the same researcher Smith is responsible for Results 1, 8, and 11, we see that while Smith's Result 1 raised Question C, which Smith answered directly (though only in part) by getting Result 8, Smith's Result 11 was obtained because of a question raised by Chu's Result 5, which Chu got in answer to another question raised by Smith's Result 1. Smith might have obtained Result 8 without sharing Result 1, but might not have obtained Result 11 at all--and no one might have obtained the other seven results of answering questions raised by Results 1 and 5.

What would the other 8 researchers have done otherwise? At least some of them might have been duplicating Smith's initial work, since they wouldn't have known Smith had already done it. No sharing of results -> duplication of effort along the same small frontier, resulting in slower accumulation of results. Sharing of results -> less duplication of effort along a wider, more rapidly expanding frontier of more results.

Posted by William Watson on June 19, 2009 at 10:02 AM EDT #

Hi David, this fundamental concept should be explored for 'idea development' in the same way that an immature idea would have 'value-added' responses from members of the tree. At the end, the idea is either parked (for more information) or is matures to be used as a research product or feature improvment in research product. Ham

Posted by Hamilton Hunter on June 22, 2009 at 02:52 PM EDT #

I agree with William that reducing redundancy is a potential benefit of OSTI's work. What is called "simultaneous discovery" is a well known problem. And as Hamilton suggests, a Web-based process for idea development may well be worthwhile. The basic point is that knowledge can grow exponentially, explosively, in a chain reaction, but only if knowledge is shared. Today it is very difficult for a researcher to know about closely related work going on elsewhere. The Web does not yet work for science.

Posted by David Wojick on June 23, 2009 at 05:23 PM EDT #

Nice work, David. But, although reducing redundancy is a valuable outcome, for true progress in science results must also still be repeatable. That is, an advancement in science should be verifiable as fact. We both know if situations in which the scientific outcome was not replicated for a long time only to find that the result was biased or compromised by experimental design or other factors. The reduction in redundancy may have its greatests value in allowing our limited resources for scientific investigation to be more accurately targeted to either validation of claims or original research.

Posted by Gene L. on July 07, 2009 at 04:10 PM EDT #

If I may add an aside and a caveat to this admirable goal at OSTI and the mission of improving efficiency in the sharing of results. And it comes, perhaps, from a more philosophical standpoint, though at its basis is the fact of being human. The old addage that knowledge is power, (unfortunately/fortunately?) still pervades all aspects of life. Almost everything invented is patented, companies closely guard their industrial secrets and knowledge-base and researchers are often close-handed (whether voluntarily or under duresse) when they are onto some breakthrough which is regarded by the employer as intellectual asset. It is human nature to want to reap the rewards of the end result. And all of the rewards! This can go down to the individual level. In order to implement such an efficient system once it becomes available, and that seems to me to be only a matter of time the way the web and a.i. are growing exponentially, the greatest hurdle that will be bestowed upon the enterprise will be the willingness of the majority to "share results" in the first place. Of course, anyone that realises that the end result would be more beneficial would welcome the new tools and systems but how many of us are there left? Cat Jarritt www.pubsofa.com

Posted by Cat Jarritt on September 17, 2009 at 05:25 PM EDT #

When David says, "These too are a product of the sharing of Result 1. In the real world this sharing process often leads to whole new subfields and specialties." This could mean a whole new data set, correct?

Posted by problem children on October 09, 2009 at 10:51 AM EDT #

Jarritt points to the well known issue that there may be a conflict with the goal of sharing and the protection of intellectual property. This is especially true in technology, but much less so in science, where publication and citation is often the measure of success. The economics community is doing a lot of research on this issue, but I take no position on it. There is enough sharing already to justify putting more effort into making it efficient.

Posted by David Wojick on October 10, 2009 at 07:36 AM EDT #

Point well made and well taken, David. Perhaps I am too much the cynic. Will be interesting to follow the findings of the research. I'd suggest that there may be a "creep rate" pervading from technology through to science with this issue that may be increasing to the point of structural failure. Perhaps its the (again, cynical?)feeling that I get with publication, peer review and citation in many cases seeming to have more of a deep-rooted basis in one-upmanship rather than sharing. Its the same feeling that I get when I see the sponsorship of science controlled more and more by "the big few". But it is sharing nonetheless and and there is enough of it, as you state; your goals of improving efficiency of sharing are admirable, interesting to follow and seem to be breaking new ground and I wish you well with them. Cat Jarritt

Posted by Cat Jarritt on October 21, 2009 at 07:16 PM EDT #