Many studies have highlighted how ethnic and religious diversity shape important political
outcomes, yet the proper measurement of heterogeneity remains a source of scholarly dispute.
We address this long-standing issue by identifying the key problems with existing approaches
and proposing a novel remedy that is easy-to-interpret and user-friendly. After introducing
the intuition with examples, we then demonstrate how it satisfies two essential mathematical
principles that the two current approaches both violate. We provide a simulation study that
shows ELF and POL both generate significant bias and can drastically mislead researchers,
while HET correctly recovers the true parameters. We also use real time-varying data, and
provide new calculations for HET, ELF and POL on the ethnic and religious composition
of all countries from 1946-2020. A new measure of heterogeneity is long overdue: HET
provides a novel, consistent and comparable measure that researchers can easily utilize and
interpret in their own studies.