Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
If only scientists had proper stat courses...

Does the MCMC computing method allows one to write better papers? More likely to be considered? Or more precise, scientifically speaking?

There are two big benefits to using Bayesian statistics, of which MCMC-type methods are a component:

1. They allow you to incorporate prior knowledge. For well studied systems, that means your results and data are automatically put into the context of the field as a whole. That's...cool.

2. Bayesian estimates - especially Bayesian posterior intervals (the equivalent to the usual confidence interval) actually have the interpretation that people think confidence intervals have. Namely that there's a 95% probability that your estimate is in between the bounds.

Those are statistical reasons why a paper might be "better". It's also a good way to make an ordinary paper somewhat "sexier" which might get it into a better journal, though at times it spooks reviewers.
 
There are two big benefits to using Bayesian statistics, of which MCMC-type methods are a component:

1. They allow you to incorporate prior knowledge. For well studied systems, that means your results and data are automatically put into the context of the field as a whole. That's...cool.

2. Bayesian estimates - especially Bayesian posterior intervals (the equivalent to the usual confidence interval) actually have the interpretation that people think confidence intervals have. Namely that there's a 95% probability that your estimate is in between the bounds.

Those are statistical reasons why a paper might be "better". It's also a good way to make an ordinary paper somewhat "sexier" which might get it into a better journal, though at times it spooks reviewers.
That's interesting. I could never really understand what the fuss was about battling on statistics, but if bayesian statistics could be used as, say, a drop-in replacement for non-parametric testing, I could even be interested in re-analyzing my own data with a much more powerful and interesting test.

But my statistics courses didn't include how to calculate the confidence interval, test power, or the like.

I hate having to say "no link found", and there's quite a bunch of data to be tested against each other in my project that will go unused from a lack of time. To my knowledge, there was no attempt to link all these data together. Sometimes finding a link first without an hypothesis can shed new light upon events we thought we knew.
 
I'm going to buy my first MBP a.s.a.p
I hate the thin design, I love MBP's design. And I'd hate to have no DVD drive
 
Yeah, I guess but external DVD drives annoy me :L lol

Apple believe that physical media are used by a small fraction of users and even those use it only occasionally. From that perspective, removing the optical drive allows several desirable design changes and saves the cost of the optical drive. Those who want an optical drive simply use an external drive.

Apart from the App Store and the iTunes store as a means of delivery of purchases, there is digital delivery from Netflix, Amazon & others. Even "boxed" software sitting on the shelf at a brick and mortar store increasingly consists of a card with a product key and a download URL. There is little point to shipping a CD/DVD when it will be out of date by the time it is purchased.

The fly in the ointment, however, is the fact that there are still more people than the manufacturers care to admit who do not have access to internet connections adequate for substantial downloads. Some vendors are providing USB flash drives loaded with their software.

In any event, it appears the trend is away from optical drives in laptops because of the number of customers saying that their laptops are too heavy and/or too big.

Unlike some, I do not see optical drives completily disappearing. There simply is too great a demand for recordable media.

Cheers
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.