NewMR
Sign up Donate

Blog

Analysis, AI updates and practical guidance for market researchers and insights professionals, written by Ray Poynter and guests.

Featured Why don’t probability samples win elections? Why can probability samples outperform opt-in samples on benchmarks yet fail to predict election results more accurately? Ray Poynter examines the evidence and the roles of modelling, non-response and shared polling errors. October 2026 · Ray Poynter
Why don’t probability samples win elections?
AI adoption needs a strategy AI AI adoption needs a strategy Individual AI use can save time, but turning those gains into organisational results takes leadership, clear roles and… 15 October 2026 How Did My Predictions for 2026 Hold Up? AI How Did My Predictions for 2026 Hold Up? Last December I shared my forecasts for 2026 in a NewMR presentation, and in January I added a… 10 October 2026 The Pew Study Is Interesting, But It Does Not Settle Whether Synthetic Data Works AI The Pew Study Is Interesting, But It Does Not Settle Whether Synthetic Data Works Pew’s new study of synthetic respondents is one of the strongest public tests yet, but it should not… 2 October 2026 The other existential risk for market research AI The other existential risk for market research Market research is debating whether AI will replace researchers, research agencies, panels, and much of the work people… 12 September 2026 When several AI services fail at once: the continuity question for insights teams AI When several AI services fail at once: the continuity question for insights teams Several leading AI services were disrupted within the same few hours on Thursday. OpenAI reported around two hours… 4 September 2026 What Is So Worrying About Recursive Self-Improvement? AI What Is So Worrying About Recursive Self-Improvement? Recursive self-improvement, or RSI, sounds like science fiction. RSI is when AI improves itself, then the improved AI… 5 August 2026 AI Update – 3 August 2026 AI AI Update – 3 August 2026 Here are three big themes from last week';s AI news. AI moves from conversation to action: three developments… 3 August 2026 Open-weight AI: four ways to use it and why access matters AI Open-weight AI: four ways to use it and why access matters Open-weight AI is often discussed as though releasing the model automatically makes AI accessible. In practice, access depends… 1 August 2026 When the model vanishes overnight: the lessons from the Fable ban and its implications for your AI supply line AI When the model vanishes overnight: the lessons from the Fable ban and its implications for your AI supply line Anthropic's two most capable models disappeared for everyone in a single evening. For research teams now building on… 20 June 2026