Google Scholar

I was using google scholar the other day. Lots of the first page hits are from companies that will sell you their articles, rather than allowing free access, and most of these seemed to be crappy meta-analysis rather than any type of controlled studies. Was this simply my false impression, or is google placing paid sites higher in exchange for advertising fees?

A lot of these sites are for the publishers that control the rights to the articles you want to read. If you have access to a university library or a library that has access to JSTOR and similar databases, you don’t have to pay for the article.

Also, what you assume to be a crappy meta-analysis either is a crappy meta-analysis, or, more likely, the abstract, which is a summary of the entire article.

What I actually meant was that I do not care for meta-analysis, just rehashing someone else’s information.

I am also disgruntled at the pricing structure. Thirty or more dollars for one article seems steep, particularly when I am assuming that I could request an actual reprint of the article for free, or at least that used to be the policy.

The main links in Google Scholar are mostly to the abstract hosted on the official site of the journal that originally published the article. If it is not an Open Access journal (and most still are not), and you are not going through an academic library that has a subscription (or you do not have a personal subscription*) then you are usually going to have to pay, through the nose, for the article. That is not Google’s fault, it is normal academic publishing practice. Non-OA academic journals are often very expensive, and for some reason buying single articles is even more expensive, per article, than subscribing. Few people buy privately anyway, it is all basically a matter of how much the publishers can extort out of academic libraries.

However, many academics these days will make copies of their articles available for free on their own site, or their university site, or some archive. Often (not always) these are “preprints” that differ in formatting, but not significantly, if at all, in content, from the “official” published version. Where these are available (and where it realizes) Google supplies helpful links to the right of its main list of results to the free, online version of the article. If these links don’t work, it can be worthwhile to click the All X versions link, which may show alternative links to free versions. Another trick is to use regular Google to find the author’s home page or departmental page, which may host versions of the article that scholar may have missed. (I once even found a version of an obscurely published article I wanted -Croatian Journal of Philosophy - in Google’s cached version of the author’s deleted home page.)

I don’t know what “crappy meta-analyses” you are seeing. That will depend on your search terms. In practice, many researchers would often prefer to see meta-analyses (which are not necessarily crappy), or review articles than go thorough all the original research papers, and those tend to get more citations, so, other things being equal, Google will put them at the top of its results. One of the best ways of finding the best original studies is to read a recent review article and then, if you want more detail, look at the studies it recommends.

¬¬¬¬¬¬¬
*Actually, although personal subscriptions to academic journals can be a lot cheaper than institutional, library subscriptions, they often do not include online access to the electronic version, just the paper. By contrast, in many cases libraries no longer bother to take a paper version (and some journals no longer even produce one).

Unless you are yourself a subject expert, though, you will not be able to tell teh good original research studies from the crappy ones.

Welcome to the world of commercial academic publishing, where profit rules (although authors and even editors don’t get a penny)! Thanks to the Web, and Open Access journals and policies, it is on its way out, but it is a slow process, and the for-profit publishers are not giving up without a big fight, and without taking as much money with them as they can.

Take advantage of searching. Google Scholar often gives you the option to “See All Versions”. Look at them. You might find one that you can download. Or you might find another article by some of the paper’s authors on a similar topic that you can get for free. If al else fails, you can either go to a public library or university where you can get the artivcle. Or you can write to the authors, mand possibly get a copy. That’s how people used to g4et a lot of copiesd of papers in the Dark Old Days, pre-internet – authors had stacks of copies that they’d send out to requesters. Now all they have to do is e-mail you a file.
I agree that the current system is a disgrace, with outrageous prices being charged for articles. This can’t be justified – they’re electronic copies. Even if you favctored electronic styorage fees into it, there’s no way the articles can cost $18 or $35 a copy. You can be certain that professors aren’t paying that much.

Professors aren’t paying squat (nor are they earning anything directly from writing the articles). It is the university libraries that are paying, and thus, ultimately, taxpayers and students. The pricing incentive structure is all fucked up, with the ultimate consumers (professors and grad students) being insulated from the costs. The result is the lining of the pockets of big publishers in New and Old Amsterdam.

Luckily there is now an alternative, Open Access, system in place, ready to pick up the pieces when the old one collapses under the weight of its corruption.

It is a problem, yes, and we’ve been pointing it out in academic libraries for some time. Factor in journal inflation costs too, and budgets that may have the same amount in them are effectively less.

It’s not something we can fix on our own. We can wave the OAflag all through libraries and academia and until the disciplines accept peer-reviewed OA articles as just as good as peer-reviewed non-OA articles, things won’t change. You really want to fix this? Fix the tenure system at the same time.
As for the OP, what you find is going to depend upon your search. There might be ways to tweak your search terms to get better searches. For example, try including the word methodology in your search - it’s one way in google scholar to help limit things a bit. You’ll still get some of what you don’t want, but it should be better. I don’t know what you’re looking for, but if it’d medical, try PubMedinstead.

The access issues will still be there though.

I don’t know exactly how the search algorithm works, but I’d assume it’s similar to web search: papers are ranked based on how frequently they are cited. And literature review papers are very frequently cited, since it’s an easy way to provide a reference for recent consensus in your field.

Also, if your searching for a broad topic, you’ll get reviews because current primary literature will be focused on much more narrow sub-topics. I.e. “obesity in children” is a huge area of research, so a google scholar search will pull up lots of reviews. PubMed, which lists chronologically, will pull up every recent paper even slightly related to the topic, things like “rare genetic variant exacerbates obesity-related disease in children”.

I don’t think it is really an issue of non-OA publications being thought better of than OA publications, just a matter of long established journals, that have gained a reputation for quality being better thought of than new ones that have not been around long enough to establish a reputation yet. Even despite their newness, there do seem to be a handful of OA journals that have established a strong reputation for quality: ones I happen to know of are Philosophers’ Imprint, PloS Biology, and Journal of Vision. Unfortunately, though, OA journals are maybe too cheap and easy to start, and there are a huge number of crappy ones about (not that there are not also many, many crappy for-profit journals too, both long established and newly founded).

You are not wrong about the tenure system (indeed, the academic training and hiring system in general is thoroughly fucked up), but I think that academic publishing is likely to shift to a predominantly OA model long before that changes.

I tend to find the name of the paper and authors from the site, and then use regular Google to see if they’ve posted it on their own website or gotten online in some other way.

Many listings for scholarly works include a section listing other works which have used the item in question as a source of info.

Sometimes, once you’ve established which un-accessible item you need badly one of the places that are listed as having cited it will be both accessible AND have reproduced the portion you need.

Occassionally, such a citation includes a link to an obscurer location for the original work at which it is accessible.

Depending on the subject matter, there might also be large, official, government-run databases for the preprints (which contain the same information as the published article). For instance, pretty much everything in physics nowadays ends up getting posted on the Arxiv, which is freely readable by everyone. I think that PubMed is similar for medical papers, but I’ve never used that, so I can’t be sure.

Pubmed has two major components. One is a fairly comprehensive index of the biological and medical literature, but it doesn’t include full-text of anything. Then there’s Pubmed Central, which includes a free-access full text and figures for a pretty large chunk of the literature. Basically, all NIH-funded publications are supposed to be publicly accessible a year after the first publication dates, so they’re uploaded to Pubmed Central. Basically you can use it to access anything NIH-funded and published between 2000 and 2011.