James Evans reports in Science magazine that electronic access to the science literature may be speeding consensus yet narrowing the range of ideas. As summarized at ArsTechnica:
The conclusion of all this statistical work was that, as more and more articles are readily available online, researchers, on average, cite fewer articles. The articles that are cited are newer, and fewer distinct articles receive attention. The results of the explosion of easily available articles, according to Evans, is that “researchers can more easily find prevailing opinion, they are more likely to follow it, leading to more citations referencing fewer articles.” As a side effect of this, a scientific consensus will typically form more rapidly. The other side of this is that papers containing ideas that don’t catch quickly will be forgotten by the scientific world much faster.
That does not seem like a wholly desirable effect. Unfortunately, being outside the bounds of a subscribing institution, the full text of the article is hidden from me for now.