re: #50 bloodstar

For what it’s worth, Rasmussen has a notorious house effect slanted towards republican candidates polls before labor day. After labor day, There are changes made to the model to better reflect reality and you’ll see a narrowing of that house affect compared to other polling outfits.

The theory is that Scott Rasmussen wants to drive a narrative (in this case, that the republicans are on a tidal wave to take over the house and senate), but wants to make sure that he gets things right by election time, so he ‘tweaks’ his likely voter model and then you see some interesting shifts in polls. That and he will tend not to poll certain elections once you’re closer to November, which can also help his average (since most people aren’t going to count a poll taken 3 months before the election against you).

A rule of thumb for the Rasmussen house effect, For every 4 weeks before the election subtract 1 point from the republican candidate and add one to the democrat to a maximum of 4.

Rasmussen doesn’t change shit… the Democrats are going to get steamrolled in November… period.