The importance of the literature review

I came across a recent article entitled Dismissive reviews: academe's memory hole, focusing on the importance of knowing what's been done in the past and the consequences and prevalence of easy dismissal of prior research.

With a dismissive literature review, a researcher assures the public that no one has yet studied a topic or that very little has been done on it. ... firstness claims and dismissive reviews are far more common than they have any right to be. And, when false, they can be harmful. Dismissive reviews assure readers that no other research has been conducted on a topic, ergo, there is no reason to look for it.

The author points out a bunch of examples, noting "[p]erhaps we can empathize with an impoverished Ph.D. student cutting corners to meet a dissertation deadline. In the research fields I know best, however, dismissive reviews are popular with some of the most celebrated and rewarded scholars working at the most elite institutions."

His area of research is educational testing and he notes some of the consequences of the lack of knowledge of the past:

All the aforementioned statements dismissing the research on educational testing were uttered within several years of the 2000 presidential campaign, the only national election in our country’s history in which standardized testing was a major campaign issue. Thus, while the most far-reaching federal intervention in U.S. assessment policy—contained inside the No Child Left Behind (NCLB) Act—was being considered, the most influential research advisors for both major political parties managed to convince policy makers that no research existed to help guide them in their program design. That casually, a century’s worth of relevant research was declared nonexistent. The result? The research-uninformed NCLB Act. ... With the single exception of the federal mandate, there was no aspect of the NCLB accountability initiative that had not been tried and studied before. Every one of the NCLB Act’s failings was perfectly predictable, based on decades of prior experience and research. Moreover, there were better alternatives for every characteristic of the program that had also been tried and studied thoroughly by researchers in psychology, education, and program evaluation. Yet, policy makers were made aware of none of them.

What solutions does he propose? One of his suggestions is more care when writing literature reviews to note where exactly the authors went searching for prior work. One of the things that the author notes makes finding prior research difficult is simply the massive amounts of data being produced.

In 2008, 2.5 million Ph.D.’s resided in the United States alone. They, and others like them now fill more than 9,100 journals for over 2,200 publishers in approximately 230 disciplines from 78 countries, according to Journal Citation Reports. ProQuest Dissertations and Theses—Full Text “includes 2.7 million searchable citations to dissertations and theses from around the world from 1861 to the present day….More than 70,000 new full text dissertations and theses are added to the database each year.

I've managed to find prior work in an area that my research touchs on that multiple experts in that area weren't aware of simply by spending a day or two working my way through Google Scholar search results on related keywords and then briefly skimming possibly-relevant documents.

Another of his suggestions also sounds reasonable to me:

I believe that we should redefine the meaning of “a contribution” to research. Currently, original works are considered contributions, and quality literature reviews are not. But, what of the scholar who dismisses much of the research literature as nonexistent (or no good) each time he “contributes” an original work? That scholar is subtracting more from society’s working memory than adding. That scholar’s “value added” is negative.