U.S. Scientists top research fraud list? Concerned? Probably not.

I happened upon this story while reading Politics Daily’s[1] coverage of a Journal of Medical Ethics article about a study of retraction incidence for research papers. The article was published in November 2010. 

The study found that leading causes of invalid research were:

• retraction due to discovery of lab error after article submission to peer-reviewed journals
• inability to reproduce results

I see that as honest behavior. Which would be easier, trying conceal or deny a mistake, or admitting error? The latter couldn’t be easy.

Braver Path Dramatization: Researcher requests article retraction

Dear ACM or IEEE,

I am the author of that research article you featured in last month’s issue. You know, the paper that was covered by most of the scientific press and popular media because my findings had such wide-ranging implications?

Well, I just found a major error in my work as I was re-reading it today. None of the peer-reviewers caught it, nor did I, until now. Please issue a retraction in my name. I’ll return that $50,000 of prize money you awarded to me. And I’ll tell the research group at [ pick any of { IBM, Princeton Advance Studies, Google Labs, NIH, CDC, Stanford University, mongoDB, Betaworks, NVIDIA} ] who offered me that great new job based on my research, that I was wrong and understand if they rescind their offer of employment and funding….

Actually, I wish the article hadn’t used the word fraud at all, as it a study of retractions, only a small number of which were due to fraud. There were certainly some cases of outright, very predatory fraud, clearly motivated by greed. The article mentions that. But that was a small part of the total number of retracted papers. In fact, when considered in the context of relative and not absolute counts, the key finding was that the retraction rate in the U.S. was 1.64%, during a ten-year interval. This far surpasses quality standards for rate of failure in nearly every other industry.

The most troubling concerns were plagiarism and deliberate falsification. Cases of both were presented in the article. Source data was drawn from on-line medical research repository PubMed from 2000 – 2009.

The article covered some other trends. Fewer American and Japanese scientists are publishing as a percentage of the total number of publications than in the past. Other countries are now entering the ring. This doesn’t mean that the U.S.A. and Japan are in technological or academic decline! It means that researchers from other nations are gaining better access to education and research funding. That helps everyone.

Also, within the United States, research breakthroughs are becoming far less concentrated in the traditional bastions of Harvard, Stanford and University of Chicago. Duke, University of Kansas, University of Iowa, University of Southern Florida and other public and private institutions are coming their own, achieving prominence like never before.

1. Politics Daily is owned by America Online News.  AOL continues to produce quality content and services, despite the brand’s unfortunate lack of prestige and status.  AOL is much more than an outdated and unpleasant internet service provider, although that is my first thought when I see the triangular AOL logo.

Industrial Production and Capacity Utilization

The Philadelphia Branch of the Federal Reserve Bank released its September 2010 Industrial Production and Capacity Utilization report this morning. It is an interesting economic indicator, due to its frequency (monthly), timeliness (within a fortnight of the prior month-end), and long history of well-documented tracking, readily available from 1972 through the present.

The salient number, based on my review and recall from past days of buy-side proprietary trading of fixed-income securities, is the capacity utilization percentage. This is why:

• Rising industrial production levels, in absolute terms, were driven by technological progress in manufacturing methods, more than increased economic activity.
• The same can be said for capacity, although the Fed does provide helpful indexing as a percentage of the historical peak level in 2007.
• Utilization is the percentage of actual capacity used for production, which indicates to me the level of demand  required for actual goods (not services!) and thus the demand for U.S.-manufactured industrial products, as a percentage of the maximum possible supply that could be produced.

The preliminary annualized utilization for September 2010 is 74.7% of capacity. This is well below the yearly average of 80.6% from 1972 through 2009. However, utilization of capacity has increased from the 2008-2009 low of 68.2%, as well as the September 2009 value of 70.5%. While I feel some concern about the economic situation in the immediate months ahead, it is encouraging to note the breakdown by process stage:

• For crude production, the operating rate increased 0.7% points to 86.9%, almost half a point higher than the 1972 to 2009 average
• For primary and semi-finished stages, utilization declined 0.6% points, to 71.5%, about 10.1% points below the long-run average
• For the finished stage, utilization decreased 0.1% points to 73.8%, about 3.7% points below the long-run average

If crude production rates drive primary, semi-finished and finished stages, then the more historically comparable rates of crude manufacturing will perhaps carry through into the later stages of production in the next three manufacturing reports of the fourth quarter of 2010.

*Note that for the purposes of this Federal Reserve Statistical Release, the industrial sector is comprised of manufacturing, mining, electric and gas utilities, as well as the logging, newspaper, periodical, book, and directory publishing industries.