A Comparison of Distribution Models for Ranked Architectural Sizes at Tikal: An Empirical Assessment of an Archaeological Economic Indicator

This preprint presents an empirical comparison of distribution models for an archaeological economic indicator ranking constructed from architectural group size at Tikal (Guatemala). Using a single rank–value dataset (n = 501), it fits and compares three candidate models—a power-law (Pareto) model, an exponential model, and a connected two-part model that combines a power-law regime with an exponential regime separated by an estimated cutoff rank. The results show that the ranking is not adequately captured by a single distribution form; instead, it exhibits a composite structure in which distributional properties change across ranking regions, implying that inequality at Tikal was continuous and structurally patterned. In addition, the preprint addresses a practical issue in archaeological description: how to relate a continuous economic-indicator structure to categorical hierarchies used in reporting and comparison. As a worked example, it provisionally proposes a nine-level hierarchy based on structural attributes and demonstrates how continuous quantities can be translated into classificatory practice without negating distribution-analytic results. For reproducibility, the study provides detailed preprocessing and estimation procedures in Appendices A–C and makes the rank–value dataset and integrated analysis code publicly available on Zenodo (doi:10.5281/zenodo.18500507).