Rob Taylor
Psychologist, statistician, and writer
Expertise
Motivation
Statistics is one of the most powerful tools for understanding the world — and one of the most poorly taught. Too often it’s presented as a collection of procedures to follow rather than a coherent way of thinking about uncertainty. The mechanics get taught but the intuition doesn’t.
Much of my writing exists to change that, at least in a small way. Every post here is written with a single goal: to help you understand not just how a method works, but why it works and when you should trust it. The mathematics is taken seriously, but it’s always in service of understanding — never an end in itself.
If you leave a post feeling like something that was previously opaque has genuinely clicked, that’s the goal achieved.
Background
My path to statistics was not a straight line. I trained originally as a cognitive psychologist, completing a PhD focused on psychophysics, visual memory, and perception. Much of that work involved building and fitting mathematical models to understand how humans process sensory information — which meant getting very comfortable with probability, inference, and the question of what data can and cannot tell you.
From academia I moved into applied data science, working for New Zealand Police where the stakes of analytical decisions are very real. That experience — taking statistical thinking out of the controlled environment of a research lab and into messy, consequential, real-world problems — sharpened my understanding of where methods work, where they break down, and where good judgement has to fill the gap.
I recently returned to academia and am building models of food systems.
On Bayesian inference
If there is one thread running through most of what I write, it is probability as a language for uncertainty. Bayesian inference is not just a set of methods to me — it is a coherent framework for reasoning under uncertainty that I find both intellectually compelling and practically powerful.
A recurring theme here will be building Bayesian intuition from the ground up: what it means to have a prior, what it means to update on evidence, and why that way of thinking changes how you approach data problems.
Short CV
2025 –
Research Fellow, Massey University
Building food system models and web applications with the Sustainable Nutrition Initiative.
2019 – 2025
Data scientist, New Zealand Police
Applied statistics and modelling where analytical decisions carry real consequences.
2015 – 2019
Postdoctoral research, visual working memory
Computational models of visual memory precision and capacity. Neural coding models.
2011 – 2015
PhD, cognitive psychology
Psychophysics and Bayesian signal detection models. Examined how feedback influences judgements of auditory stimuli.
Selected publications
2022
, Tomić, I., Aagten-Murphy, D., & Bays, P. M. Working memory is updated by reallocation of resources from obsolete to new items. Attention, Perception, & Psychophysics, 1–15. PDF
2020
Schneegans, S., , & Bays, P. M. Stochastic sampling provides a unifying account of visual working memory limits. Proceedings of the National Academy of Sciences, 117(34), 20959–20968. PDF
2020
, & Bays, P. M. Theory of neural coding predicts an upper bound on estimates of memory variability. Psychological Review, 127(5), 700. PDF
2018
, & Bays, P. M. Efficient coding in visual working memory accounts for stimulus-specific variations in recall. Journal of Neuroscience, 38(32), 7132–7142. PDF
2018
Bays, P. M., & A neural model of retrospective attention in visual working memory. Cognitive Psychology, 100, 43–52. PDF
Get in touch
Questions, suggestions, or you just want to talk statistics — I’d love to hear from you.