‘I think you’ll find it’s a bit more complicated than that” is the title of both the latest book by @bengoldacre and his #londonthinks talk tonight @ConwayHall
Strangely that title is my conclusion from both of the last two Ethical Society and Humanist talks I saw last week – so much so that having drafted a piece on those two together, I can now see me enlarging it to cover all three talks in one post.
[Post Note : Scrap that. Not as advertised. Old TED Talk “Better Data” on Statins et al. Cheap laughs and no new content, but good points included:
- Psychology of statistics and down-side risk is real problem even for official practitioners and management of medical options – perverse decisions in terms of optimum outcomes.
- Big data approach is real option for public health, despite being largely marketing jargon in so many fields. Again perversely it is easy to justify one-off expensive clinical trials with lower value (even negative value) outcomes, than to do cheap easily-randomised long-term statistical analyses on “freely” available GP records – because of privacy rights risks perceived with the easy option. Wrong view of risks again.
- (It’s not about making “data” publicly available, but making available public data into proper “information” for decision-making. Since there are many ways to manipulate statistical data, we need to be able to trust who is turning data into information for us, ‘because psychologically it’s not intuitive for us, Joe public, to interpret data directly without a proper statistics & risk lens. I’m guessing Ben’s agenda is to sell the industry such a lens and trust is pretty much my earlier “authority” topic – ie who says?.)
But not the talk advertised. Left before the Q&A. Oh, and I now see the new book is just a collection of Ben’s previous posts and articles, hence why no new content. Be really interested to know the context of the title quote.]