Fred L. Bookstein

Prof. DDr.

Professor Emeritus

External links

Profile in ResearchGate



 About me

My main academic interest is in the foundations of reasoning from numerical evidence across the natural sciences, in particular the ways that statistical methods do or do not help us get from arithmetic to understanding in fields like evolutionary biology. Early in my career I created two techniques for representing the information content of organismal forms, the Bookstein shape coordinates [I did not name them that] and the thin-plate spline. These techniques have been widely adopted across many branches of organismal biology, but their basis in biomathematics was not examined deeply enough until much later.  For the last twenty years I have been occupied in rebuilding the links between theoretical biology and the multivariate statistics of these coordinates and splines.  There has resulted a pair of textbooks ("Measuring and Reasoning", 2014, and "A Course in Morphometrics for Biologists", 2018) and several difficult articles on the assumptions behind some inappropriately popular approaches such as Procrustes analysis, principal components analysis, and statistical significance testing.

This collection of iconoclastic yet principled insights has led me into a haphazard collection of application domains.  In America I am most identified with the neuroanatomy of fetal alcohol syndrome as seen in brain images, particularly brains of men convicted of murder and at risk of a death sentence (which, here in America, is still a possibility). A new interest here is in the underlying logic of the analyses of risk associated with the subduction earthquakes along the Juan de Fuca fault and the consequent tsunamis that afflict the Pacific coast near my Seattle home every 300 to 500 years.  In Europe I have been pursuing extensions of the thin-plate spline statistics to incorporate specific bioengineering hypotheses, such as nasal airflow or pelvic flexion.  Other areas of interest include the role of statistical graphics in the communication of complex multilevel findings and the role of strong statistical inference (pattern analyses going beyond ordinary statistical significance testing) in the analysis of multivariate longitudinal human data.  I am also very interested in teaching the limits of statistical inference in evolutionary biology more generally, with a goal of re-establishing the centrality of explicitly biotheoretical arguments such as were pursued a hundred years ago at the Vienna Vivarium, the direct ancestor of our present department.

 Selected publications

Bookstein, F. L. (2018). A course in morphometrics for biologists: Geometry and statistics for studies of organismal form. Cambridge University Press.

Bookstein, F. L. (2014). Measuring and reasoning: Numerical inference in the sciences. Cambridge University Press.