Whenever I do a peer review, I do all my annotations in the PDF, then waste time rewriting it all into a text review. Double work and so much information lost. Why can’t we just review in the PDF? Just me?
#academicsky #academia #scisky 🧪
Take it as a compliment. Appearantly, they did not find anything real to criticize.
I'm still early career, so no spamming yet 😄
Well, at least the review is quickly done this way... I would lose trust in the whole manuscript if this happened to me.
This is not unusual in my perception.
They add you to their system first so that they can contact you via their system, I guess?
REMINDER: I am looking for two PhD students to work with me on super exciting field experiments in the Economics of Education, joining an extremely vibrant environment in Munich, @ifoeducation.bsky.social & @econmunich.bsky.social 🤩🚀
Deadline in 11 days ⏰
#EconSky #EconJobs #AcademicJobs #JobAlert
Academia really knows how to use bad software...
It won't be perfect, but using AI to detect fake references is a good idea. It's very tedious to do manually. At the same time fake paper / fake citations are increasing in numbers.
As an author, I also like the constructive reviews the most.
If reviews are not actionable for the authors, we would spend a lot of time for only :thumbs up:
In the end, only one thing matters, and that's quality. If AI produces high-quality reviews, we should absolutely use it.
However, we have to consider the side effects. For example, it's useless when some reviewer only uses AI without own input. The editor can do that himself...
I've seen high-quality MDPI publications. Lots of authors publish there thinking these are normal journals.
I don't think you can exclude them generally.
That's the most shocking part of it, how willingly some people waste other people's time...
True! Writing high-quality code is such an integral part of good research nowadays. In and outside of CS.
Worth investigating! But I'm not fully convinced. The side effects will be drastic. People will stop reviewing if there is no payment. Other researchers will make good money by reviewing a lot. Many AI created reviews will be created to make more money. Especially the last point is a problem...
True! Preregistration is essentially non-existent in my field (computer science/engineering).
I'm very interested in the publication system and how to improve it. A broader implementation of these ideas is probably a good step forward.
The reviewing system seems to break apart. Eighth months is just ridiculously slow...
I'm currently building reviewersearch.net. I hope it will help to make things more efficient and faster.
Thank you!
However, as far as I understand it, research screener is a tool for writing review papers, not for peer reviews. Still an interesting approach.
Yeah, probably ChatGPT. However, usually, the citation is complete bogus and can never be found.
Thank you! I hope that my tool can also offer that in the long-term. A good quality database of reviewers, how quick they answer, how good their reviews are, etc.
I didn't know covidence yet. Thanks for sharing! I'm actually building something related at the moment : reviewersearch.net
Finally submitted my PhD thesis!
Now, I have time to do some coding for reviewersearch.net again.
And probably some days of rest are also a good idea? Weird feeling to finally have this done.
#AcademicSky #PhDone #Academia #Science #PhDChat #HigherEd #AcademicChatter
BTW do you know @jobrxiv.bsky.social already?
My personal lesson learned: Don't change your topic over and over again to find the perfect topic. Stick with the first good-enough idea and get it done! (PhD)
Yes, that is a common thing, at least in Germany. The Prof already has someone in mind but is forced to advertise publicly. That's how such situations happen...
Great work! I really like the idea of fast reviewing and will certainly consider this when improving reviewersearch.net. However, I think that a single week is not necessary. Two to four weeks would still be a great improvement without the pressure on the reviewers/editors.
The first article in RJIM!
For-profit publishers are increasingly dominating scientific publications. $$APCs are growing and your work is paywalled, so they can make profit margins larger than Google’s.
So 2 dozen researchers and I created RJIM…
#medSky #EMIMCC
www.rjmedicine.org/publications...
Seems to be a common problem, and there must be a technological solution. One solution is to stop publishing our email in our papers (which then gets crawled by some algorithm). It doesn't make sense anyway to have a short-lived email in a publication. Better would be to use the ORCID for this.
One risk here is certainly that people will submit AI-generated reviews to make quick money...
I am skeptical about direct payment. There are too many potential negative side effects and false incentives.
But some kind of renumeration is certainly required.
How do you recognize them? Just from the overall style? Or are they just garbage quality-wise?