我发现川粉并非不信民调
只是相信他们喜欢的民调
比如这家
据说2016神准
但是被左胶538列为c级booksin
那538把这家列成c总有理由吧
于是我看了下2018 mid term这家参议员的poll
Arizona Senate - McSally 47.6 vs. Sinema 50.0 D+2.4 (WRONG PREDICTION)
Nov 4 – 5 Trafalgar Group (R) McSally 47, Sinema 45 McSally (R) +2.0
bias R +4.4!
Florida Senate - Scott 50.0 vs. Nelson 49.9 R+0.1
Nov 4 – 5 Trafalgar Group (R) Nelson 47, Scott 49 Scott (R) +2.0
bias R +1.9
Indiana Senate - Braun 50.8 vs. Donnelly 44.9 R+5.9 (WRONG PREDICTION)
Jul 31 – Aug 7 Trafalgar Group (R) Donnelly 51, Braun 39 Donnelly (D) +12.0
bias D +17.9!!
Michigan Senate - James 45.8 vs. Stabenow 52.3 D+6.5
Nov 2 – 5 Trafalgar Group (R) Stabenow 53, James 44 Stabenow (D) +9.0
bias D +2.5
Missouri Senate - Hawley 51.4 vs. McCaskill 45.6 R+5.8
Oct 29 – Nov 4 Trafalgar Group (R) Hawley 48, McCaskill 44 Hawley (R) +4.0
bias D +1.8
Montana Senate - Rosendale 46.8 vs. Tester 50.3 D+3.5
Nov 2 – 5 Trafalgar Group (R) Tester 50, Rosendale 49 Tester (D) +1.0
bias R +2.5
Nevada Senate - Heller 45.4 vs. Rosen 50.4 D+5.0 (WRONG PREDICTION)
Oct 29 – Nov 1 Trafalgar Group (R) Heller 49, Rosen 46 Heller (R) +3.0
bias R +8.0!!
North Dakota Senate - Cramer 55.1 vs. Heitkamp 44.3 R+10.8
Oct 23 – 24 Trafalgar Group (R) Cramer 54, Heitkamp 45 Cramer (R) +9.0
bias D +1.8
Texas Senate - Cruz 50.9 vs. O'Rourke 48.3 R+2.6
Nov 3 – 5 Trafalgar Group (R) Cruz 52, O'Rourke 43 Cruz (R) +9.0
bias R +6.4!
West Virginia Senate - Morrisey 46.3 vs. Manchin 49.6 D+3.3
Jul 13 – 16 Trafalgar Group (R)* Manchin 50, Morrisey 40 Manchin (D) +10.0
bias D +6.7!
这家期中选举参议院做了10个州的poll，错了3个
选前（11月）做的7个里面错2两个，几乎是三分之一的错误率
尤其以nevada的d+5做到r+3最为离谱
cruz的+2.6在选前做到+9也不遑多让
仔细观察，这家的手法似乎4这样:
如果有一方明显领先的州
选前很久瞎丢个数字出来，简直就是信手胡作
关键的拉锯州，毫无例外的在选前偷偷给共和党从两趴到八趴修正带风向
大概4这样
单纯从结果讨论，完全没有刺激川粉的意思
感恩～