What changes would you make to the current publishing system?

[Dason]

I've read a few articles lately on how certain people feel about the current peer-reviewed publishing system. In short: they don't like it. For example this is a recent blog post by Andrew Gelman. He has more to say on the issue but that was his most recent post on the topic.

I'm inclined to agree that the system seems a little bit broken and if we started from scratch at this moment it would hopefully look quite a bit different than what it is today.

So my question is... if you had the power to change the way articles get reviewed and/or published what changes would you make? Would you keep things the same or would you make a radical overhaul?

[hlsmith]

Cut-Down the number of medical journals, I believe there are 14,000-20,000 indexed journals. Every disease doesn't need 30 journals. This would increase the quality of the select journals. Also, all journals should have a statistician/epidemiologist on staff or available for refereeing every manuscript. If the manuscript was not reviewed by one of these individual or someone with commenserate experence, it should be noted in the publication.

[trinker]

google scholar is another. The rich keep getting richer off it because it sends out the top ranked and many are lazy and so the top ranked keep getting cited more. I'd like to see a different ranking algorithm used.

I'd like to see more community owned journals that are edited by the field. Something like wiki only more guided (I think I stole this idea from something I read recently). I'd also like to see a process where articles are submitted to a bank and the journals fight over the articles rather than the other way around (I think I stole this idea from something I read recently). Journals control way too much of the theory and content of the field (at least in mine). This would do away with top tiered journals but could level the playing field, speed up turn around on submission, and actually allow the field to affect theory, methods and content rather than editors with a specific agenda.

[duskstar]

Interesting post.

I submitted to a journal recently where the readers of the article voted it up and down and that (I think) determined where it appeared on the site. It was an interesting idea, their arguement was that peer reviewers arent always experts, and sometimes thats more the readers. Although it was peer reviewed, it was reviewed more for quality of writing and whether statistics was sound.

On a side note (slightly off tangent) I am not a fan of the ridiculous fees to make work open access. A lot of us don't get that cost covered by our grants, and so we scrimp around trying to find money any time work is accepted (I like my work to be open access, and so when the journal isnt open access I try to make it available that way). I wonder how often work isnt published because of that cost (I went to a talk here once where a lecturer implied his Masters students were put off publishing their work).

[lorb]

What duskstar said. The fees and costs that come with publishing in any well-cited journal are definitely ridiculous.

[bugman]

I agree with all. I recently submitted a paper and was told to allow free access it would cost 1,500 dollars for each paper!! Things need to change.

[derksheng]

For some fields, there needs to be a reputable central body that collects the intention of the experimenter and the planned number of trials/size of the experiment and what exactly the experiment entails BEFORE the experiment begins.

Then when published a note can be made that this was done on the front page of the paper.

Then there can be meta-analyses that only look at this sample and have, by construction, 0 file drawer effect (although the publication bias effect might still exist, and the experimenter might just cancel the experiment if it's showing negative results..).

This is needed where claims of the efficacy of some treatment or the reality of some effect are based on the aggregation of 10s or 100s of small experiments. e.g. parapsychology.


Kind of orthogonal to the purpose of the OP, though.

[Jacov]

Actually, I like the peer-reviewed process. But the final prices 25-35$ for an article and 2000$ for open access is ridiculous. How can they justify these prices? I mean the authors are doing all the editorial work.

For example, I always submit in Tex. But somehow Springer always manages it to put errors in the text and formulas. How could that be? So we often need 2-3 rounds for proof reading. I also review articles and believe me sometimes you need more than 30-40 hours to understand in detail the submitted articles. This work is all done by the authors and then readers have to pay such unbelievable prices.

[Dragan]

The problems I have with traditional journals are mainly the lengthy review process and the lack of open access. I recently published an article in a Wiley-Blackwell journal (Statistica Neerlandica) and they told me it would cost $3000 to make the article open access.

Over the past year, I have had success with the open access Hindawi journals (and the ISRN journals, which also Hindawi). In just this current year (2012) alone, I have managed to get 11 articles accepted. Now, Hindawi wants between $500 and $800 for an article to be published ---but they do all the work i.e. typesetting, etc. I think they do a very good job.

Fortunately, to cover these publishing costs, my University library has an institutional repository offering permanent and free access to research and scholarly publications produced by faculty (both tenure or non-tenure track) and graduate students. All I need to do is submit my (open access) publications to the repository’s administrator and the University reimburses me for the amount of money I paid.

You may want to check with your University to see if they have a similar program in place.

[Jacov]

in just this current year (2012) alone, i have managed to get 11 articles accepted.

wow !!!!!!!

[victorxstc]

If I had the power, I would eliminate those "open access" journals which are created for the sake of making money. They easily accept articles without profound review, and their main goal is to make money. So they are biased. Unfortunately Medline and ISI are trick-able, so they can be fooled by some methods. Their criteria (especially Pubmed's ones) are not stringent. They need a high number of submitted articles in the first place, and I can say they don't care too much about the context. I have seen too many weak journals now indexed in Pubmed or Pubmed Central.

Every week I receive lots of ads from newly emerged open access journals, many of which manage to get into Pubmed. So thats becoming a trouble.

2, I agree with guys here that all journals need at least one "good" statistician (a good statistician, as I have seen in person journal statisticians who even are famous, but know nothing compared to some guys here)... Imagine that they judge the quality of statistics while they know literally nothing.

3, if I was a journal editor (or someone who could make policies for editors of different journals), I would ask for raw data from every author, and would put that in the journal's online repository. This way the future investigators could compare their results easily with the results of the current article, or could judge its statistics (and send letters and correct the statistics or bring on better results/conclusions), or could incorporate those raw data in the metaanalyses. Besides, further systematic reviews could manage to gather a very large sample based on the raw data from various studies, which I think can be great. The disadvantage was that those who had lost their raw data could face some difficulties. Maybe they could send their articles to journals with weaker grades.

4. As said before, I would add a new type of article. Currently, an original research has to bring something new to the literature. Therefore, many editors are inclined to reject the confirmatory articles while they are very important. So the authors have to write many jargon in order to make a confirmatory article look like a new one, which this is both discouraging to them, and is boring and also distracting to the readers. So a new type of article in which it would look like a short communication (with word limits as small as 2000 or 1500 words) but it would be categorized as original study might be good, so that the authors could briefly report their findings and shortly compare them with the findings of the exploratory studies, without adding boring / repetitious material.

5. I think (and many other too) that impact factor and H index suck. We have a perfect journal in our field which publishes 3000 to 4000 pages a year with IF = 1.5, while there is a not-bad journal which published about 200 pages a year with IF = 1.6. Therefore, the latter is ranked higher. I think other factors should be created / modified or emphasized instead of IF, to avoid this. There are many ways to increase the IF without necessarily increasing the quality and real impact of the journal on the science.

[Dragan]

5. I think (and many other too) that impact factor and H index suck..

I agree.

Moreover, the way Google Scholar "counts" the number of an author's citations is absurd.

[trinker]

I'd also make everyone accept tex files

[trinker]

Maybe we'd have less cheating (as there seems to be a great deal of discussion around this lately) if an article could be published if the results weren't significant. Not sure how to work out the details but we'd maybe take more chances with what we investigate and uncover some nifty things. True science looks at failures too (IMHO).

[bryangoodrich]

I think there are meritorious changes already coming along, such as open-access publishing and publishers like PLoS. Didn't a bunch of universities also recently make it so their researchers have to pay the fee when publishing so that their articles will be publicly available through the publisher? A lot of people don't like the broken system and avenues beyond it and just as trustworthy are being developed alongside the use of today's technology. I also recall hearing about an article aggregator or storehouse that focuses exclusively on "failures." That way, people can post the results of their work that didn't amount to anything successful. In that way, people can not waste time making the same mistakes. Along with the creative common licensing and other options out there, I think we're already on the right track. With the increase in online education and training that is a growing phenomena, I think we're going to see some pretty incredible changes to the way people learn and information is disseminated over the rest of this decade. Unfortunately, the vast majority of the population aren't well equipped to assimilate that information, and I think we're going to need to see new ways in both the way we deliver information itself and in the ways we learn to operate in society, but that's a broader issue altogether.