Appropriate null model for SPF spatial distribution

Determine whether a homogeneous Poisson (random) distribution is an appropriate null hypothesis for the spatial distribution of Selected POSS1-E Features in POSS1 plates and, if not, develop a suitable alternative null model that accounts for artifact-generating processes.

Background

Analyses in prior studies used statistical tests presuming random distributions to claim nonrandom effects (e.g., shadow deficits, alignments).

The authors argue that because the processes generating plate artifacts are not well understood, it is not clear that a random distribution is the correct null hypothesis, indicating a need to establish and justify an appropriate baseline.