Kendall Tau.. is a alternative

Spearman's rho is a standard rank correlation, defined here:

http://www.statsref.com/HTML/index.html?rank_correlation.html

You want a measure of agreement, not correlation, I think. Think carefully about the question behind the data. Since the options are ordered, then there will be N unique values in each column. You can ask if the predicted value follows the same order as the real world (Kendall's tau-b) but you can also ask if they agree (intraclass correlation or coefficient of concordance). The precise question will depend on the meaning of the data, so the answer is beyond statistics into the realm of science.