Standard error

Most clinical trials answer a question. But is it the right one?The estimands framework is changing how researchers define what they're actually trying to measure, and why it matters more than you'd think. In this episode, we unpack what an estimand is, why the difference between "intention to treat" and "what really happened" keeps statisticians up at night, and how getting precise about your question changes everything downstream.

In this second instalment of a two-part series, we unpack some of the most commonly used – and most frequently misunderstood – statistical terms in clinical trials and medical research.From p-values and confidence intervals to power, effect sizes, and Type I versus Type II error, we explain what these terms actually mean, how they’re used in practice, and where people often go wrong.

In this first instalment of a two-part series, we unpack some of the most commonly used – and most frequently misunderstood – statistical terms in clinical trials and medical research.From p-values and confidence intervals to power, effect sizes, and Type I versus Type II error, we explain what these terms actually mean, how they’re used in practice, and where people often go wrong.

Adaptive clinical trials are often described as “new”, “innovative”, or even “controversial” — but what do they actually mean in practice?In this episode of Standard Error, we unpack the core idea behind adaptive trial designs. Starting from first principles, we explain why adaptive designs emerged, what problems they are trying to solve, and when they genuinely add value over traditional fixed designs.The episode also introduces more advanced ideas — such as seamless designs and response-adaptive randomisation — while being honest about where these approaches help, where they don’t, and why complexity is not always a virtue.This is a practical, concept-driven episode aimed at researchers, clinicians, and students who want to understand adaptive trials without being buried in equations. If you want to understand how modern trials can evolve while they’re running — and why that’s harder than it sounds — this episode is a good place to start.Link to Mike's adaptive trial that was published in The Lancet Psychiatry: https://doi.org/10.1016/S2215-0366(25)00129-4

In this first episode, we step back from methods and results to ask a more fundamental question: what are clinical trials? We unpack the ethical, scientific, and practical reasons trials exist, from uncertainty and equipoise through to patient benefit, decision-making, and health system impact. This episode sets the foundation for the series, framing clinical trials not as abstract statistical exercises, but as tools designed to answer real questions under real constraints.A couple of links related to this episode: https://www.australianclinicaltrials.gov.au/about/what-is-a-clinical-trialhttps://pmc.ncbi.nlm.nih.gov/articles/PMC1884542/pdf/bcp0058-0001.pdf

The Standard Error is a podcast about statistics, clinical trials, and the interpretation of evidence in medicine and health research.Hosted by researchers working in clinical trials and applied statistics, the podcast explores how evidence is generated, analysed, and communicated — and where common misunderstandings arise. Episodes examine core statistical concepts, trial design, and methodological debates, using real examples from published research.Topics include p-values and hypothesis testing, Bayesian and frequentist frameworks, trial endpoints, bias and uncertainty, and how statistical choices influence clinical conclusions.The aim is not to simplify complexity away, but to make it intelligible. The Standard Error is for clinicians, researchers, and anyone interested in how medical evidence is actually produced and interpreted.Music track: A Positive Direction by AylexSource: https://freetouse.com/musicFree Vlog Music Without Copyright SE