Hacker News new | past | comments | ask | show | jobs | submit login

I've tried to solve that challenge a few times now. Here's what I've learned so far, from most efficient to least:

* Ask an expert; this is a simple question for them. A university library's reference desk has worked well for me. For example, looking for books on Egypt, I contacted a university library referece desk and was connected to the library's Arabic Studies reference librarian, who had a Ph.D. in the field from an elite university. What was a challenging question for me was probably offhand knowledge for him. (Of course, respect their time (be prepared and patient) and the fact that you are not entitled to their services, but he seemed excited that a member of the public was interested and serious.)

* Look for scholarly articles reviewing publications in the field. I use Google Scholar but I get the impression JSTOR is better if you have access (I've never used it and know little about it). In each case I've read several articles to triangulate what is widely accepted and what is particular to that author, and to fill gaps in each author's coverage. However, I find the review articles themselves fascinating, almost essential background on what I want to read, and tend to get lost in them. Lacking expertise myself, I still feel uncertain about the conclusions I draw - what don't I know? what context am I missing? - there is no substitute for expertise.

* Look at class syllabi at universities. This has been slow going: I have to find the classes and the syllabi (if available), and then hope it includes more than an unannotated list of books.

* Do not: Use general web resources. It seems like the general web and the scholarly world are two separate silos of knowledge. I find 100x more valuable information in the scholarly articles, and random people on the web are often badly misinformed (this includes Amazon reviewers, a great source of complete nonsense that sounds good until you know better).

Overall, it makes me appreciate the resources available to university students, who have free and easy access to experts in a huge variety of fields, all paid to help them.




Very astute, especially the library reference idea. In the same vein:

- Avoid news brought up by social networks - people like and share what's emotionally engaging, so the story writers have a very strong incentive to embellish. Majority of the "news" on social media is either bogus or empty.

- Major newspapers check credentials of their guest authors. Those guest articles aren't always correct, but signal/noise is way higher.

- Articles in peer-review journals are usually right, especially the bigger journals with more eyes on them.

- Approaching a librarian is not as hard as it may seem. Here's UW. for example: http://www.lib.washington.edu/about/contact

- Apparently there's a "24/7 Reference Cooperative", a nation-wide librarian Q&A service: http://wiki.questionpoint.org/w/page/13839418/24%207%20Coop%...


However, for the biomedical sciences, you should assume on the order of half of the scholarly, published in refereed journals papers are incorrect. See e.g. that essay by a Lancet editor, and the general history of this, which I've been watching since the late '70s.

Why? I think it's because this stuff is hard, so many variables, publish or perish, and the fact that while referees for any field can filter out a lot of guff, they have strict limits in their ability to determine correctness. That comes when the results of a paper are used as the basis of further research in other labs.


+1 for looking into peer reviewed publications, specifically review articles.

Not only are they concise, but they are highly referenced and well annotated. As long as you are willing to dig into the specific terms and concepts relevant to the subject yourself, you will find a very high S:N ratio in these papers.

The best part is that you can easily find yourself down 4-5 different rabbit holes by the time you are done because of the easy to follow references.




Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact

Search: