In 1997, IBM's Deep Blue computer beat Garry Kasparov at chess. He had won their first encounter in 1996, with 3 wins, 1 loss and 2 draws (4-2), so the team of programmers and chess experts tweaked the program and in 1997 came out ahead 3.5-2.5, a big achievement for programming because chess is 'creative'.

Another creative task even more relevant to scientists than chess is extracting data from scientific publications to create a database cataloging results from tens of thousands of individual studies. It sounds like it would be easy but much of cataloging would ordinarily be placed in the 'subjective' camp. 

Relational aggression, such as malicious rumors, social exclusion and rejection, are considered something that girls do more often. The movie "Mean Girls" epitomized it to hilarious effect. A trio of scholars used surveys to show that boys are being shortchanged in popular accounts of mean-ness.

620 randomly selected sixth graders were followed through their senior year, filling out an annual survey talking about victimization. Using group-based trajectory modeling the female co-authors determined that boys are actually meaner than girls - or at least they brag about it more on surveys. Boys were more often to call themselves relational aggression perpetrators while girls reported being victims more.

In epidemiology, matching curves are often enough to imply causation and so it is often done, even if there is no evidence to warrant the link.

There has been an increased use of antibiotics and there are an increased number of diagnoses so some epidemiologists looked at those two curves in the same direction and suggested one may be causing the other. A new analysis of about 500,000 children published in BMJ dismisses those claims and finds that exposure to antibiotics during pregnancy or early in life does not appear to increase the risk of asthma. 

Imagine a car that can't get dirty or a business in Ferguson, Missouri that can't be graffiti'ed by unhinged protesters.

It might be possible with a new class of highly fluorinated super-repellent polymer called a “fluoropore”, which mimics the natural ability of lotus plants and cabbage leaves to make water droplets simply roll away - but for oils too. This lotus effect has been used for producing rough surfaces with special chemical properties. “However, this trick does not work for oils – the lotus plant repels water, but no oil,” says Dr.-Ing. Bastian Rapp of the KIT Institute of Microstructure Technology. “Oil-repellent surfaces need to have another chemical structure, fluoropolymers, for this purpose.” 

The Arctic Ocean sea ice cover emerged 2.6 million years ago - and it hasn't changed since. Not in all of the recurring warming cycles we have had and not even in 2006, when pundits predicted it would be melted by 2014.

It wasn't always that way. Between 4 and 5 million years ago, the extent of sea ice cover in Arctic was much less than it is today. Recent IPCC reports believe that the expanse of the Arctic ice cover has been quickly shrinking since the 1970s and that 2012 was the known sea ice minimum in that time.

A growing number of scholars are using social media data to write articles about both online and offline human behavior - it's cheap, it's as accurate as surveys if properly controlled, and no one ever has to leave the office.

But surveys are not science for an obvious reason and yet, in recent years, studies have claimed the ability to predict everything from summer blockbusters to fluctuations in the stock market. They all get mainstream media attention despite obvious evidence of flaws in many of these studies.