Welcome, Guest. Please login or register.
Did you miss your activation email?
February 20, 2019, 01:06:02 am
HomePredMockPollEVCalcAFEWIKIHelpLogin Register
News: New features added! Click here for more information. Click here to configure new features.

+  Atlas Forum
|-+  Presidential Elections - Analysis and Discussion
| |-+  2016 U.S. Presidential Election (Moderators: TJ in Oregon, Virginiá)
| | |-+  NJ: Academics and Model-makers on Hillary Clinton's chances (search mode)
Pages: [1] Print
Author Topic: NJ: Academics and Model-makers on Hillary Clinton's chances  (Read 5180 times)
YaBB God
Posts: 4,547

Political Matrix
E: -1.16, S: -4.35

View Profile
« on: February 17, 2015, 10:11:11 am »

You can't have a "model" of who will win a Presidential race.  

Anyone knows that campaigns, candidates and unpredictable events matter to the outcome.  So, if your model doesn't account for those variables, and it can't, it's worthless.

At the same time, the data points for any model are lacking.  Presidential elections only happen once every 4 years.  Plus, you can't really include data points about Presidential elections that happened outside of the current paradigm.  So, you're basically left with an n size of 8-10.  The result is ultimately just a mix of banal truisms and conjecture.  A bad economy is bad for the incumbent party... wow what a stunning insight.

Over the course of time, we've seen the best single predictor of the outcome is public polling.  At this point though, it's also of limited reliability because so much can change.  And, if you test Hillary Clinton vs. someone without name recognition, it's not really fair.  So, there's nothing you can really say that's worth a damn about predicting the election.  Hillary Clinton is obviously the favorite, but there are no guarantees.  
Pages: [1] Print 
Jump to:  

Login with username, password and session length


Powered by SMF 1.1.21 | SMF © 2015, Simple Machines