Michael Lenczner
@mlencCanadian working in nonprofit data & tech. Thinking about information infrastructures. Bookmarks: https://t.co/l6Gne1twkp
Similar User
@CivicTechTO
@TraceyLauriault
@futureofgood
@jwmcconnell
@kentdaitken
@AllanNorthcott
@JHvanGeuns
@EnviroFundersCA
@JPBervoets
Nine wealthy Canadian families and foundations have made Canada's largest-ever philanthropic commitment to climate action, pledging $405-million over the next decade to accelerate the shift to a low-carbon economy. @TrottierFdn going all in with $150M! theglobeandmail.com/business/artic…
fsg.org/resource/webin… Mark Kramer, Co-Founder & Board Chair "...people can solve their own problems. We don’t need to develop solutions for them. We do need to empower them. We need to give them support from their peers, and we need to give and encourage government to give…
“Avoiding things that dont work is as important as funding things that do.”
Great to see a Denver city councilman quoting an RCT of a basic income pilot meant to reduce homelessness that didnt work as justification for not making the program permanent. The final vote failed, 6-6. Avoiding things that dont work is as important as funding things that do.
THIS 👇
When I took office, I promised to make government work more effectively and efficiently for the good people of Pennsylvania. We delivered. It took 8 weeks to get a business license here in PA. Now, it takes 3 days. It took 12 weeks to get an initial teacher certification. Now,…
It's happening. #learningagendas
A big day for researchers interested in UK economic policy! @hmtreasury has published its first ever “Areas of Research Interest” document, setting out the research questions the Treasury would most like to see answered.
Want to work with me and are excited about all things knowledge graphs and product? We have an opening for the Product Manager for Wikibase Suite. Apply! wikimedia-deutschland.softgarden.io/job/50824818/P…
Free the data
🧵 spain has some of europe's best neighbourhood-level income data, but it's buried in the statistical office website i scraped and structured the data to make it accessible with one line of code check out my new R package: pablogguz.github.io/ineAtlas/
Say I wanted to write a pop thing about how hard it is to do "evidence based policymaking" (or grant making) because of: 1. low replication and code quality 2. low power 3. selection on statistical significance 4. a culture of indifference to 1-3 What are good (real) examples?
🚀Today I launch Policycraft, a free monthly newsletter curating fresh and insightful thinking on the practice of public policymaking in Canada 🗞️ (1/3 )
🙁 high key
Seriously, read this thread, look at the preregistration, look at the papers. This is low-key academic fraud by one of the world’s most famous economists. You can’t bury the null on the preregistered outcome and then do posthoc analysis as if it was your intention all along.
I don’t know how many RCTs are getting done, but thinking that they could suffer from the file-drawer problem is brain-itchy. 🫠
I love this (the research, not the result) because I worry that long-run RCT follow-up studies suffer from publication bias, in which only successful RCTs get followed up. Admin data could help run the follow-ups without the cost. Nice to see an example!
😬😬😬 Now *that* was bold thinking. #socialinnovation #systemschange #techforgood
Remember those null effects of One Laptop Per Child? They don't get better in the long run! publications.iadb.org/en/laptops-lon… from @SantiagoCueto6, Beuermann, @JulianCristia, and others
Gov't agencies love these elaborate construct-a-data-set GUIs / shopping cart widgets when they'd almost certainly be more useful to their real users if they just gave you a link to a big CSV & a codebook
#opengov opportunity
You could see that for yourself, but... that leads me to the next problem. Transparency. I defy you to actually *find* these agreements. The summaries are relatively easy to find on the CMHC website. But finding the full agreements - that takes effort.
NBER posted an RCT of the Chicago Heights (CHECC) preschool program. Quick take: The study doesn't report findings for its prespecified primary outcome - grade 3 test scores for full sample - & instead highlights positive results of an unreliable (post-hoc) subgroup analysis. 1/6
This session is giving me a clear example of what professional malpractice would look like in the realm of data and information sharing systems.
United States Trends
- 1. Travis Hunter 8.964 posts
- 2. Heisman 4.252 posts
- 3. Northwestern 5.728 posts
- 4. Carnell Tate 1.448 posts
- 5. Sheppard 2.817 posts
- 6. $CUTO 8.090 posts
- 7. Colorado 67,1 B posts
- 8. Jeremiah Smith N/A
- 9. Arkansas 27,2 B posts
- 10. Ewers 1.227 posts
- 11. Shedeur 3.166 posts
- 12. Jahdae Barron N/A
- 13. Denzel Burke N/A
- 14. Wrigley 3.558 posts
- 15. #Buckeyes N/A
- 16. Renji 7.090 posts
- 17. #SkoBuffs 3.205 posts
- 18. #HookEm 2.345 posts
- 19. #collegegameday 5.809 posts
- 20. #Caturday 9.548 posts
Who to follow
-
Civic Tech Toronto
@CivicTechTO -
Tracey P. Lauriault
@TraceyLauriault -
Future of Good
@futureofgood -
McConnell Foundation / Fondation McConnell
@jwmcconnell -
Kent Aitken
@kentdaitken -
Allan Northcott
@AllanNorthcott -
Hacked account...
@JHvanGeuns -
Environment Funders Canada
@EnviroFundersCA -
JP Bervoets
@JPBervoets
Something went wrong.
Something went wrong.