Of the millions of science "studies" produced each year, how much are they to be believed? A look at some trends in science reporting - fake or real?