I heard someone claim there have been more scientific discoveries outside of academia, in industry, the past 50 years than in academia. Is this true? Has anyone tried to quantify this?
This is definitely true in regard to applied sciences: there have been significantly more discoveries of things that make money in the institutions designed to make money than in the institutions designed to teach. The amount of industrial research in electronics and computer science and other branches of engineering is very impressive.
I don't think though that the magnitude of the industry-driven research can be accurately quantitatively measured because many discoveries remain unpublished proprietary information for years. Very often a trade secret would make one more money than a patent. For example, RSA cryptography was known to British Intelligence 20 year before it was published by RSA, and that fact was not publicly disclosed until 30 years after RSA publication. In academia it's "publish or perish"; in industry it's often "don't publish or perish".
That said, I think the situation is different in regard to fundamental discoveries. Things that would open humanity's eyes but won't pay the discoverers' bills are still researched primarily in academia.
This is not a philosophical question, but it has some philosophical offshoots, whatever answer is given. Most importantly -- how are discoveries made and what conditions are conducive? If you know that, the main question is whether academia provides such conditions. This is a controversial topic, and I liked an article last year about this in the context of Vinay Deolalikar's P≠NP proof.
Anyway, I don't even think it would be possible to quantify this in a strict way. After all, you would need to select some criteria for what a discovery is (was the lightbulb or search engine a scientific discovery?) and whether it was made in industry or academia (the line is usually quite blurry and becoming more so).
The phrase "more scientific discoveries" has three words in it that are vague and that can be made precise in many many different ways. Depending on how one defines these terms, the truth of this statement will vary.
The National Science Foundation regularly publishes statistics for various comparisons between academia and industry. Here are excerpts from two articles. The emphasis is from the original articles.
Business R&D Performance in the United States Increased in 2011, by Raymond M. Wolfe
During 2011, companies in manufacturing industries performed $201 billion (68%) of domestic R&D, defined as R&D performed in the 50 states and the District of Columbia (table 2). Most of the funding was from companies’ own funds (81%). Companies in nonmanufacturing industries performed $93 billion of domestic R&D (32% of total domestic R&D performance), 81% of which was paid for from companies’ own funds.
Science and Engineering Indicators 2012, National Science Board
The business sector continues to account for most of both U.S. R&D performance and R&D funding.
The business sector performed an estimated $282 billion of R&D in 2009, or 71% of the U.S. total, drawing on business, federal sources, and other sources of R&D support. The business sector itself provided an estimated $247 billion of funding for R&D in 2009, or 62% of the U.S. total; almost all of which supported R&D performed by business.
The levels of business R&D performance and funding were both higher in 2008 than in 2009 ($291 billion and $259 billion, respectively). Even with the decline in 2009, expanded business spending has accounted for most of the nation's R&D growth over the last 5 years.
The academic sector is the second-largest performer of U.S. R&D, accounting for an estimated $54 billion in 2009, or about 14% of the national total.
The federal government is the second-largest funder of U.S. R&D, providing an estimated $124 billion, or 31% of the U.S. total in 2009.
U.S. R&D is dominated by development activities, largely performed by the business sector. The business sector also performs the majority of applied research, but most basic research is conducted at universities and colleges and funded by the federal government.
In 2009, basic research was about 19% ($76 billion) of total U.S. R&D performance, applied research was about 18% ($71 billion), and development was about 63% ($253 billion).
Universities and colleges historically have been the main performers of U.S. basic research—and accounted for about 53% of all U.S. basic research in 2009. The federal government remains the primary source of basic research funding, accounting for about 53% of all such funding in 2009.
The business sector is the predominant performer of applied research, accounting for 58% of all U.S. applied research in 2009. Business is also the largest source of funding for applied research, providing 48% in 2009.
Development is by far the largest component of U.S. R&D. Funding for development comes primarily from the business sector, 78% in 2009; nearly all of the rest comes from the federal government.
A few years ago I was invited to participate in a workshop hosted by the EC on fundamental research in software engineering. I remember Wolfgang Emmerich from University College London talking about their Impact project, which had found evidence that most innovations in industry are traceable to academic research, but only after some time has passed. The report from the workshop with the details can be found here.