Head on over to LinkedIn to read my thoughts on co-leadership. In summary, i'm in favour!
https://www.linkedin.com/feed/update/urn:li:activity:7155152912518713344/
https://www.linkedin.com/feed/update/urn:li:activity:7155152912518713344/
@MarkKelson
- 130 Followers
- 137 Following
- 51 Posts
Professor of Statistics for Health丨Turing Fellow丨Interim co-director of IDSAI | Uni of Exeter, UK丨Reproducibility | Trials & physical activity丨Reading: The Quincunx
Our policy evaluation report for the Home Office was published just before Xmas (a mere 118 min read!). We found promising signs of effectiveness using synthetic control groups.
The headlines picked out this as noteworthy "Since funding began, an estimated 3,220 total (or 8 per 100,000 persons) hospital admissions for any violent injury had been prevented in funded areas. ".
Come to Exeter to do a fascinating 4-year PhD with me and Russ Jago and Tom Monks - "Using modelling approaches to reduce inequality in physical activity interventions".
Exeter is a great place to live and the University campus is beautiful!
It is open to UK *and* international students!
The closing date is 5.00pm on 1st November 2023. The stipend is £18,622 p.a. for 2023/24 (updated each year).
Details here: https://www.exeter.ac.uk/study/funding/award/?id=4863
I'm on a policy evaluation course and spent some time looking through Hansard yesterday. Found the first mention of the term "statistically significant" from 1952 regarding proportions of admissions to psychiatric hospitals! Made by Sir Reginald Bennett. https://en.wikipedia.org/wiki/Reginald_Bennett
Ben LongdonJust been looking at the CIA’s Simple Sabotage Field Manual: A Timeless Guide to Subverting Any Organization with “Purposeful Stupidity” https://www.openculture.com/2015/12/simple-sabotage-field-manual.html (HT @MarkKelson) - think workplaces have long been infiltrated!
Read the CIA’s Simple Sabotage Field Manual: A Timeless Guide to Subverting Any Organization with “Purposeful Stupidity” (1944)
I’ve always admired people who can successfully navigate what I refer to as “Kafka’s Castle,” a term of dread for the many government and corporate agencies that have an inordinate amount of power over our permanent records, and that seem as inscrutable and chillingly absurd as the labyrinth the character K navigates in Kafka’s last allegorical novel.
My feasibility study on Kinship Connected has been published on the What Works for Children's Social Care website. We propose some ways forward in evaluating this intervention. https://whatworks-csc.org.uk/research-report/kinship-connected-the-feasibility-of-a-pilot-randomised-controlled-trial-investigating-outcomes-for-children-in-kinship-care/
Kinship Connected: The feasibility of a pilot randomised controlled trial investigating outcomes for children in kinship care - What Works for Children's Social Care
Summary Kinship Connected is a programme developed by Kinship, the leading kinship care charity. The programme offers one-to-one support to kinship carers for at least six months via a dedicated project worker, and develops and maintains kinship carers’ support network through community-based peer support groups. Research illustrates that Kinship Connected has been positively received by […]
I've started listening to Radio 4 (yeah, yeah) and the Desert Island Discs segment came on. I'd never actually listened to it before.
It was utterly charming and enjoyable, but am i right in saying that they don't actually play the chosen songs in their entirety?!? What the hell?
Dorothy Bishop@MarkKelson
Gosh, thanks! Not very confident about this, but have been worried that in my field, people seem convinced you should only use one outcome, when, with appropriate corrections, multiple outcomes will give more power.
My sense is that there's been too much focus on avoiding false positives, and not enough attention to false negatives.
Gosh, thanks! Not very confident about this, but have been worried that in my field, people seem convinced you should only use one outcome, when, with appropriate corrections, multiple outcomes will give more power.
My sense is that there's been too much focus on avoiding false positives, and not enough attention to false negatives.
@MarkKelson I should add that I owe a great debt to the reviewers of 1st version - Daniel Lakens and Kristin Sainini - as this is F1000, you can read their reviews and see just how much they influenced this more recent version.
I know peer review gets a lot of criticism but sometimes it works very well. I've been lucky with this one I think
I know peer review gets a lot of criticism but sometimes it works very well. I've been lucky with this one I think
Lovely paper (from twitter https://twitter.com/deevybee/status/1613923572677857281?s=20&t=bcr4y4nFW9T7iTDSjv3tIw) from @deevybee introducing the MEff to a wider audience. Nice intuitive idea from genetics to adjust the alpha level depending on how *correlated* your multiple outcomes are. If uncorrelated, it reverts to Bonferroni, but if correlated you pay less of a penalty. Paper here https://f1000research.com/articles/10-991/v2
Dorothy Bishop on Twitter
“In a bold move, I challenge CONSORT: "Having several primary outcomes incurs the problems of interpretation associated with multiplicity of analyses and is not recommended." https://t.co/TmepoaxBLA V useful, critical reviewer comments for v1! Hope I have responded adequately.”