What topics does Chris Said's blog cover?

Updated

Chris Said's blog is a personal site written by a data scientist at Propel, and it covers three broad areas: statistics, technology, and science. The homepage describes it as "mostly about statistics, technology, and science," and the post archive backs that up — you'll find practical statistics write-ups alongside essays on AI, public health, and scientific integrity. If you're looking for applied statistics explanations or data-driven commentary on science and tech policy, this is the relevant scope.

Statistics and data analysis

This is the blog's core strength. Posts tend to explain a specific method or statistical fact and why it matters in practice.

  • A/B testing and experiment design — A three-part series, "Optimizing sample sizes in A/B testing" (Part I: General summary, with Parts II and III), published 10 Jan 2020.
  • Variance and scaling — "Variance after scaling and summing: One of the most useful facts from statistics" (18 May 2019).
  • Empirical Bayes — "Empirical Bayes for multiple sample sizes" (03 May 2017).
  • Causal inference — "Instrumental variables analysis for non-economists" (13 Mar 2021).
  • Measurement pitfalls — "You're probably measuring your treatment effect incorrectly" (28 Mar 2021).
  • Cross-entropy — "Things that confused me about cross-entropy" (26 Dec 2020).

A recurring theme is correcting common analytical mistakes, so the statistics posts are aimed at practitioners rather than pure theory readers.

Technology, AI, and data infrastructure

Technology posts range from AI risk to the practical plumbing of data work.

  • AI — "The El Chapo model of AI containment" (12 May 2023).
  • Data pipelines — "Coronavirus and fragmented data pipelines" (14 Apr 2020).
  • Site/tooling — "Upgrading to MathJax 3.0 after the Kramdown update in Github Pages" (05 Sep 2020).

Science, public health, and research integrity

A large share of the archive is science commentary, often with a quantitative angle.

  • Pandemic analysis — "Why low sensitivity antigen tests are better than slow PCR tests" (15 Jul 2020), "Contagiousness sensitivity: The metric that could control the pandemic" (29 Nov 2020), "Waning immunity and the little paradox of R0" (01 Jan 2022), and "Pandemic virus prediction: All risk and minimal benefit" (18 Dec 2021).
  • Research misconduct — "The case for criminalizing scientific misconduct" (17 Jun 2024), "Scientific whistleblowers can be compensated for their service" (18 Aug 2024), and "Bounty hunters for science" (25 Aug 2024).
  • Health and behavior — "Small correlations can drive large increases in teen depression" (10 May 2022) and "Teens, loneliness, and the social media paradox" (14 Aug 2021).
  • Miscellaneous science — "Everything I've learned about solar storm risk and EMP attacks" (18 Jun 2020) and "A map of laboratory-acquired infections" (01 Nov 2022).

How to browse the archive

The homepage lists "Selected Blog Posts" near the top, then a "Full Blog Archive" with dated entries running from 2016 through 2026. The selected list is the fastest way to find the statistics pieces; the full archive is where the science and policy essays live. Posts are dated, so you can also read chronologically to follow how topics shift over time.

Who it's for

  • Data scientists and analysts — for method explanations and experiment-design guidance.
  • Science and tech policy readers — for quantitative takes on public health, AI, and research integrity.
  • General readers — the essays are written accessibly, though the statistics posts assume some familiarity with the concepts.

The author also links to his Twitter and LinkedIn from the homepage if you want to follow new posts.