Sign error in SLV fear index

[ccjolley]

I was just adding some sanity checks to fear_correlations.Rmd to ensure that all three indices were pointing in the right direction, and it turns out that El Salvador's index was reversed. I don't know if this affects the conclusions that @calbano drew when she ran this, because she didn't set a random seed at the time. Worth double-checking with the random seed set (now on line 16).