The first step in identifying the clusters was to create a simple scale that totaled the number of times a respondent said the trends were good for society, bad for society or made no difference.