Sign in

Chao Wang

@excel-wang.bsky.social
25 followers 20 following 89 posts

Associate Professor in health and social care statistics at Kingston University. PhD in econometrics.

PostsRepliesMedia
Chao Wang @excel-wang.bsky.social · 23/09/2026
Study design does not necessarily dictate the level of evidence, as often depicted in "evidence pyramid". Context is important. See the example below (youtu.be/3pxudE0GNAo?...). Many people thought Study 2, a cross-sectional study, is the weakest. However Study 2 is arguably the strongest.
110
Chao Wang @excel-wang.bsky.social · 03/08/2026
In reality it’s more like “we can't possibly do that”… until we could 🤔
100
Chao Wang @excel-wang.bsky.social · 31/07/2026
I’m not sure about that. I think @jmwooldridge.bsky.social demonstrated Poisson regression doesn’t have any distribution assumptions.
281
Chao Wang @excel-wang.bsky.social · 14/07/2026
Also this headline "fatal/serious injuries ⬇️ 34%" is hugely misleading. The 34% headline figure refers to the intervention group alone, not being compared with the control group! It's disingenuous to present 34% as the effect of 20mph!
000
Chao Wang @excel-wang.bsky.social · 21/05/2026
Tried a few models and both Gemini 3.5 Flash (Extended, screenshot1) and M365 Copilot GPT 5.5 (Think Deeper, s2) got the correct answer. Some reported Claude Opus 4.7 is also fine. Yes more advanced mode may not always be better, but they tend to be more correct than instant/quick/... mode.
001
Chao Wang @excel-wang.bsky.social · 11/05/2026
Nice benchmark "Which LLM writes the best Stata code?" www.khaledeltokhy.com/benchmarks/
000
Chao Wang @excel-wang.bsky.social · 18/11/2025
P-value has been severely criticised for 7 decades (the paper in the screenshot was published in 1994). I guess it's not going to go away anytime soon...
010
Chao Wang @excel-wang.bsky.social · 17/11/2025
Yes but this could either mean "the deaths are expected to continue to decrease" (without pandemic) or "the deaths are expected to increase due to 'dry tinder effect'" (www.sciencedirect.com/science/arti...). A simple average is not perfect but it's more model-free.
100
Chao Wang @excel-wang.bsky.social · 16/11/2025
Strange to cherry pick two states. How about a comparison of death rates across all blue and red states? Note red states generally had less stringent policies. www.bsg.ox.ac.uk/sites/defaul...
141
Chao Wang @excel-wang.bsky.social · 16/11/2025
I asked ChatGPT how to calculate confidence intervals without researching and sourcing third-party R packages. Here’s the response.
000
Chao Wang @excel-wang.bsky.social · 10/11/2025
Yes but they are all methods trying to match the UK with other countries, which is unnecessary. A more flexible approach is synthetic difference-in-difference, which only creates "parallel trend". See the comparision between SC and SDID www.nber.org/papers/w25532. The true effect may be smaller.
100
Chao Wang @excel-wang.bsky.social · 30/10/2025
The correct word is "he" dictionary.cambridge.org/dictionary/e...
100
Chao Wang @excel-wang.bsky.social · 09/10/2025
Maarten has a paper on that diagnprognres.biomedcentral.com/articles/10....
010
Chao Wang @excel-wang.bsky.social · 17/09/2025
New meta-analysis of effects of 20mph speed limit in the UK shows 𝙣𝙤 statistically significant impact on killed or serious injured (KSI) crashes or injuries on sign-only schemes. t.co/WSvmnbLJ1n
000
Chao Wang @excel-wang.bsky.social · 17/09/2025
Thanks for the link. Looks like the authors agreed with me that the existing research, in particular the "before and after studies" being reliant on, is weak in deriving causal relationships. Also the report only covers the potential benefit but not cost (e.g. journey time), though not their fault.
100
Chao Wang @excel-wang.bsky.social · 17/09/2025
While it looks promising, it appears to be a simple before-after study (aka single-arm trial with baseline), which no expert in causal inference will consider as solid evidence.
110
Chao Wang @excel-wang.bsky.social · 21/08/2025
I wouldn't create an account on Truth because the user base is so small (4Chan? Haven't even heard of it). Also X seems to be more align with the real world (www.pewresearch.org/journalism/2...), but I can understand this may not be useful for some users.
000
Chao Wang @excel-wang.bsky.social · 07/06/2025
Suppose the value of time is £8/hr (assets.publishing.service.gov.uk/media/5a80bd...), then 2min is worth £0.27. The value of prevent a fatal accident is ~£1.7million (assets.publishing.service.gov.uk/media/673f68...) ...
200