Information Growth Rates

People often ask, "How fast is information growing in our society?".  This could be measured in a number of ways, such as the annual counts of books published or the number of Ph.D. degrees awarded.  The "conventional wisdom" is that information is now doubling every two years.¹  But is this really true? For technical information, it seems that the actual number of peer-reviewed publications and/or the number of patent applications are good measures for the rate of growth of information.

Below is a chart the information growth rates in PubMed and Patent Databases over the last 35 years.  In 2000, the data indicates that biomedical information is doubling in 27 years, in 1990 the doubling time 23 years, while in 1987 it was 21 years².  In 2000,  patent filings were doubling at a rate of 13 years in contrast to 1990 when they were doubling every 19 years³.  Patent filings may be growing faster than biomedical information in part because they are fewer in number (it is much easier to double something that is small than double something that is large), or perhaps because there is a large amount of venture capital flowing into startup companies that produce patents to protect their ideas.  Biomedical information is heavily dependent on governemnt funding and grows largely with fluctuations in the NIH budget.  Patent filings are more likely to vary with the state of the economy and industry R&D expenditures.  Due to delays between U.S. patent filings and publication, accurate statistics for the most recent years are unavailable.  Though not shown in the chart, the growth rates for gene sequence information such as is provided by GenBank or UNIGENE is much faster with doubling times of 2-3 years.  This is largely due to the fact that the information is simply being read from the genomes and entered into the databases with only a small amount of "value-added" by humans.

Both exponential and power functions show relatively good fits to the data. The growth rates vary over time, but the exponential character of the curves can be seen.

(click on chart for larger version)

Footnotes

1. Peter Cochrane, "What price serendipity?", Hard Drive column in Daily Telegraph, October 15, 1996.

BT's Head of Advanced Applications and Technology, says mankind's knowledge base is now doubling in less than two years.
2. It is difficult to extend the biomedical information doubling times further back with the existing data because PubMed has significant discontinuities before 1966.
3. The information in the online patent database shows a significant decrease in entries before 1969, so extrapolations prior to this period are questionable.

Sources:

• Information Growth Data Table (requires style sheets)
• InfoGrowth.xls (the Excel spreadsheet with the source data and charts)

Related Discussions

• Doubling of Medical Knowledge by Robert A. Freitas, Jr. (Aug 1998).
• Information "Factoids" by Michael Lesk
• "How Much Information is in the World" by Michael Lesk
• "Mad Library Disease: Holes in the Stacks" Lazerow Lecture, given at the UCLA, 18 April, 1996 by Michael Lesk