Astrological Synastry in a Bipartite Network: An Exploratory Analysis Using a Demographic Null Model

This preprint presents an exploratory analysis of astrological synastry in a bipartite network of 494 F–M dyads. It documents an explicit statistical procedure for comparing observed angular summaries with those expected under a null model that preserves the network’s degree sequence and weights alternative configurations according to two age-related demographic characteristics. Four MCMC chains of bipartite networks were generated using 2-switch edge swaps. Angular data were represented by sine and cosine components across several harmonics, forming a 1,296-dimensional vector. Three global statistics were examined: Mahalanobis distance, energy distance, and max-T. Their upper-tail empirical p-values were approximately 0.039–0.048 for Mahalanobis, depending on shrinkage; 0.100 for energy distance; and 0.196 for max-T. A subsequent Mahalanobis analysis of six pair–harmonic blocks was more extreme relative to the evaluated null networks. However, these blocks were selected using the same exploratory data, so the p-value does not account for that selection and the result is not confirmatory. The preprint also reports the available MCMC diagnostics and their limitations, including sample dependence and Monte Carlo resolution. This work does not seek to affirm or reject astrology in general. Its aim is to contribute more rigorous, testable, and reproducible methods for evaluating specific questions about synastry, while distinguishing exploratory analyses from future independent confirmatory tests. Individual-level data are not included because they contain personal information; any release would require prior verification of applicable permissions and ethical and legal requirements.