Speaking of Causality, from Kim Zussman
I used unemployment data to calculate the month-to-month change in unemployment:
"chg" = (this month UNRATE%) - (last month UNRATE%)
Partitioned into Dem and Rep presidencies, each compared to zero:
One-Sample T: D chg, R chg
Test of mu = 0 vs not = 0
Variable N Mean StDev SE Mean 95% CI T P
D chg 325 -0.0086 0.2276 0.0126 (-0.0334, 0.0162) -0.68 0.496
R chg 420 0.0216 0.2063 0.0100 ( 0.0018, 0.0414) 2.15 0.032
Under "D", there was no significant change in month-to-month unemployment, but under "R" there was, on average, significant increase from month-to-month (of 0.02%).