"So what I need is a way of measuring how similar two ranked lists are… And good news! this metric exists and is provided by the Spearman’s Correlation Coefficient for Ranks (SCCR) -BTW, the Kendall correlation coefficient does the job as well-.
For each year, I computed the SCCR value with all other years split by gender. I created a scattered plot where the size and the transparency of each point is determined by the SCCR value. Intuitively if we have a look at the diagonal, bigger opaque points together form a cluster where the trend persists and places with almost no color in this diagonal represent trend interruption."
There should be a way to apply this to the MPC voting. The order could be the dove to hawk and the change in personnel would be the equivalent to new members of the committee.
No comments:
Post a Comment