Like my research interests, my writing has ranged over many different areas over the years, and appeared in many different outlets, from national newspapers to specialist magazines.

Among the themes I’ve pursued are:

  • Why do so many “breakthroughs” fade over time ? I became interested in this when I started working on Bayesian methods back in the early 1990s. By the late 1990s I was highlighting the potential role of the abuse of textbook statistical methods by researchers. At the time, academic institutions showed little interest in this possibility. That has now changed with the emergence of the “Replication Crisis”.
  • What’s the evidence for that ? Despite its reputation, science has always had its share of grand claims based on suspect evidence.  Two examples I started to write about over 20 years ago are the plausibility of earthquake forecasting and the reliability of animal models in drug testing. Again, attitudes towards both have changed markedly in recent years. Today few geophysicists believe there are precursors that can reliably warn of the time, place and severity of major earthquakes, while results based on animal models now routinely come with caveats about extrapolation to humans.
  • Warning: science is done by humans. The methods of science have helped debunk countless misconceptions, and is the best set of tools we have for understanding reality. But in the end, they’re still wielded by humans, and thus susceptible to everything from simple mistakes and outright deception to fashions, intransigence and ambulence-chasing. My columns for BBC Focus, The National (UAE), The Financial Times and others often focus on examples of these human foibles and their impact on the scientific enterprise.