My rating system is based on data compiled from 15 sources and uses a scale of 3 categories (Safe/Likely/Weak) for each party (Republican/Democrat/Independent). The rating is based on a numerical score where a likelihood of victory less than 67% is Weak, 67-83% is Likely, and greater than 83% is Safe. To fit within this site's rating system, the Likely category is split in half so that a likelihood of 75% or more is "Strong" and less than 75% is "Lean." No races are rated as Toss Up.

My rating system is based on data compiled from 15 sources and uses a scale of 3 categories (Safe/Likely/Weak) for each party (Republican/Democrat/Independent). The rating is based on a numerical score where a likelihood of victory less than 67% is Weak, 67-83% is Likely, and greater than 83% is Safe. To fit within this site's rating system, the Likely category is split in half so that a likelihood of 75% or more is "Strong" and less than 75% is "Lean." No races are rated as Toss Up.

My rating system is based on data compiled from 13 sources and uses a scale of 3 categories (Safe/Likely/Weak) for each party (Republican/Democrat/Independent). The rating is based on a numerical score where a likelihood of victory less than 67% is Weak, 67-83% is Likely, and greater than 83% is Safe. To fit within this site's rating system, the Likely category is split in half so that a likelihood of 75% or more is "Strong" and less than 75% is "Lean." No races are rated as Toss Up.

My rating system is based on data compiled from 13 sources and uses a scale of 3 categories (Safe/Likely/Weak) for each party (Republican/Democrat/Independent). The rating is based on a numerical score where a likelihood of victory less than 67% is Weak, 67-83% is Likely, and greater than 83% is Safe. To fit within this site's rating system, the Likely category is split in half so that a likelihood of 75% or more is "Strong" and less than 75% is "Lean." No races are rated as Toss Up.

My rating system is based on data compiled from 11 sources and uses a scale of 3 categories (Safe/Likely/Weak) for each party (Republican/Democrat/Independent). The rating is based on a numerical score where a likelihood of victory less than 67% is Weak, 67-83% is Likely, and greater than 83% is Safe. To fit within this site's rating system, the Likely category is split in half so that a likelihood of 75% or more is "Strong" and less than 75% is "Lean." No races are rated as Toss Up.

My rating system is based on data compiled from 12 sources and uses a scale of 3 categories (Safe/Likely/Weak) for each party (Republican/Democrat/Independent). The rating is based on a numerical score where a likelihood of victory less than 67% is Weak, 67-83% is Likely, and greater than 83% is Safe. To fit within this site's rating system, the Likely category is split in half so that a likelihood of 75% or more is "Strong" and less than 75% is "Lean." No races are rated as Toss Up.

My rating system is based on data compiled from 10 sources and uses a scale of 3 categories (Safe/Likely/Weak) for each party (Republican/Democrat/Independent). The rating is based on a numerical score where a likelihood of victory less than 67% is Weak, 67-83% is Likely, and greater than 83% is Safe. To fit within this site's rating system, the Likely category is split in half so that a likelihood of 75% or more is "Strong" and less than 75% is "Lean." No races are rated as Toss Up.

August 2014 Update: Colorado from Strong D to Lean D, Connecticut from Lean D to Lean R, Florida from Lean D to Lean R, Georgia from Strong R to Lean R, Hawaii from Strong D to Lean D, Kansas from Strong R to Lean R, Wisconsin from Strong R to Lean R<br /> <br /> May 2014 Update: Illinois from Lean D to Lean R, Pennsylvania from Lean D to Strong D

Illinois Lean D to Lean R, Pennsylvania Lean D to Strong D

My election assessment method doesn't map exactly to the terminology used on this site. We rate races as "Safe," "Likely," and "Weak" for one party or another (no toss ups). I merged Safe and Likely into strong confidence and classified Weak as lean. I set all the percentages as >50% because we don't track that in our method.