This study offers a method for the measurement of political cleavages by decomposing the measure of diversity in the opinions of survey respondents, in within and between parts across sub-populations formed through demographic categories. A synthetic population is generated by adopting as a seed the responses to a questionnaire from a non-representative sample. Two different kinds of cleavage are considered: the first measures how far apart two finite and logically adjacent sub-populations are in terms of the opinions surveyed through the questionnaire; the second measures how much a G categorical grouping variable is relevant to observing divisive opinions. In the conclusions, an extension of the method is offered for sample estimates.