Linus Van Pelt
Sr. Member
Posts: 2,145
|
|
« on: January 15, 2014, 10:27:35 PM » |
|
It should be kept in mind, when communicating in venues other than the Atlas Forum, that the definition of "trend" used on here is an idiosyncratic usage of Dave Leip's, and not the usual definition used by statisticians and the rest of the world. The Atlas definition of "trend" is just as short-term as "swing" - calculated from only two elections - but it's relativized to the national average (so by mathematical definition the nation as a whole doesn't "trend" at all). By contrast, what is usually meant by "trend" (as measured by a linear regression, for example), is longer-term than just two data points, but not relativized to any larger average such as the whole country.
That being said, it does look like you were clear in your discussion here is specifying that what you meant was trend relative to the national average.
|