A long time ago I calculated correlation coefficients for how closely correlated the changes in the Democratic % nationwide are correlated to changes in the % Democrat for each state.
Of course this is a tad more precise as bellwether since its a "bellwether" to actual changes in movements in the voting habits of the state.
Alabama and Mississippi have opposite correlations largely due to 1964 which is weighing down a lot of the Southern states.
This is 1960 to present.
PA 0.966067054
IN 0.95341152
NM 0.949044516
MO 0.946832887
NJ 0.942439247
MI 0.934318638
MD 0.933170349
CO 0.927456992
NY 0.923773398
CT 0.919558356
NE 0.911658126
DE 0.910226962
VA 0.910198852
NH 0.905413893
WA 0.904983944
KS 0.901958647
WI 0.901263071
NV 0.900218139
OR 0.89605574
AZ 0.894099031
RI 0.87412351
IL 0.872103429
HI 0.861837923
CA 0.860305372
FL 0.857749611
TX 0.842579042
OK 0.837252298
ID 0.833592726
IA 0.828201135
ME 0.824423507
OH 0.81955975
VT 0.809179288
MA 0.796208464
MN 0.794198757
NC 0.779418497
MT 0.77638117
UT 0.768767821
KY 0.742380603
AK 0.705862352
WV 0.684463347
WY 0.674054027
SD 0.641501564
TN 0.631647735
AR 0.556386947
SC 0.476540412
LA 0.455813242
GA 0.433122723
MS -0.128090695
AL -0.250714962
Interesting. I'd like to see something that focuses on recent elections (say, from 1972 onwards) to reduce some of the weird stuff the South did.