This paper adopts a formal model-testing approach to the Peruvian radiocarbon (14 C) record, the site of the first aggregate analysis of this type of archaeological data. Using a large and improved regional dataset of radiometric determinations (n = 1180) from the period 14000-3000 14 C years before present, the study performs a comparative analysis of the demographic trajectories of two subregions, the desert coast and Andean highlands. Against the backdrop of theoretical models of population growth, and controlling for taphonomic factors and sampling biases, the study performs global significance and permutation tests on the data. These provide a necessary measure of statistical confidence that have hereto been absent from the discussion of pre-Columbian demography. Contrary to the findings of prior work, this study of radiocarbon data in Peru reveals that regional trends in the data are statistically indistinguishable. Further testing and comparison to climate archives is able to illustrate sustained population growth over the entire Holocene epoch in this region, with only a few notable exceptions at the end of the mid-Holocene (5000 cal BP). The findings of the analysis are viewed in relation to the cultural and technological changes that indigenous societies experienced in the timeframe in question, and some directions for methodological advances are suggested.