Picture of textbook on Bayesian Meta-Analysis

Excited to get my hands on this new meta-analysis text by @rlgrant.bsky.social @bma-net.bsky.social

Love it. The cookbook is also great too! 🍝

Last day of the holiday’s for me, and I put the finishing touches on our new herb vertical garden. Pretty happy with how this one turned out.

Reposting this for folks in a different timezone

My research unit at #MonashUni just joined BSky! Check out our intro 📣 and give us a follow if you’re interested in #whatworks in social welfare – and the methods we apply to find out

Love this!

I really enjoyed Al Roth’s Who Gets What - and Why. Market Design has never looked so good. Also, Heilbroner’s The Worldly Philosophers…

Jim's Greek Tavern in Collingwood is an institution if you have a crowd, or are just hungry! Gray and Gray in Northcote is a personal favourite.

Huge fan of your team's work @michael-sanders.bsky.social! The What Works Centre's are doing a great job funding some important research in this space, we're really grateful for @foundationsww.bsky.social's for funding our review.

Thanks for reading! Shout out to two of my awesome co-authors who are also on BSky: @aronshlonsky.bsky.social and @balbers.bsky.social. Both who whom do and share excellent work on #whatworks, #evidencesynthesis and #impSci

14/14

Many existing programs for care leavers haven't been causally evaluated — a major evidence gap. Its great to see new rigorous studies emerging though, like @michael-sanders.bsky.social & team's recent work (too recent for our review alas) on Staying Put and Lifelong Links in England 👏

13/14

Implementation Science could help improve the delivery of some interventions. Rather than looking for a shiny new *intervention*, there’s a big opportunity to focus on improving the quality of “usual services” provided to care leavers using a continuous quality improvement approach.

12/14

What can we do about this?

The expansion of extended care will likely continue as is easy to scale and may have important short to medium-term benefits. However, it is unlikely to improve long-term outcomes for care leavers when provided alone. Additional support is likely needed.

11/14

Extending care may be beneficial. However, on its own, it is unlikely to solve the many complex challenges that care leavers face as it appears to mostly delay, rather than prevent, negative outcomes.

10/14

Why? It’s possible that interventions are either delivered at the wrong time, with inadequate quality, or at an insufficient intensity or focus to make a meaningful difference to the trajectory of care leavers.

9/14

Although some individual interventions may offer a marginal improvement compared to services as usual or other programs of a similar nature, the bigger issue is that collectively current policy and practice does not provide care leavers with the support and resources they need to thrive.

8/14

What does this mean? The existing evidence is insufficient to recommend any particular policy or intervention…

7/14

We conducted 20 small meta-analyses for two different types of transition support programs. Only 2 showed results favouring the intervention. However, we have very low confidence in these results.

6/14

We found 14 eligible studies. From these, we calculated 152 effect sizes. For ~65% (98/152), there was no detectable difference between intervention & comparison groups. When differences were detected, they were typically very small. 📊

5/14

We looked for studies using either RCTs or QEDs to measure the impact of programs on range of outcomes:
🏠 Housing & homelessness
🧠 Health & wellbeing
🎓 Education
📈 Economic & employment
⚖️ Criminal & delinquent behaviour
⚠️ Risky behaviour
🤝 Supportive relationships
🪡 Life skills

4/14

Support for care leavers falls (broadly) into two categories:
1️⃣Transition Support Programs — include independent living programs, coaching & peer support, transitional housing etc.
2️⃣ Extended care policies — policies enabling continued care beyond 18 with sustained funding & practical support

3/14

Care leavers are more likely to face a range of major challenges compared to their peers. While there's lots of research is happening in this space, the last systematic review that sought to find and assess all (causal) program evaluations was done in 2006. Time to have another look 🔍

2/14

Sage Journals: Discover world-class research Subscription and open access journals from Sage, the world's leading independent academic publisher.

New Paper from my PhD research 📣 — Interested in what works to support young people leaving out-of-home care? 🤔 Check out a #systematicreview that’s just been published (open access) in TVA.

Here's a 🧶 highlighting what we found...

journals.sagepub.com/doi/10.1177/...

1/14

Couldn't agree more. I recently read Richard Evans' Third Reich Trilogy (highly recommended btw) and he frequently quotes Klemperer throughout. What a story...

Great 🧵 @milescorak.bsky.social 👏 FWIW, showing results replicate in different contexts — ideally through a systematic review + meta-analysis, can be more compelling than a single study. Implementation is often neglected and showing that the idea is (easily) scalable, is often of interest too...

Hi 👋 I'm an applied guy working in social welfare, mostly child welfare, youth homelessness/employment. I work with linked admin data to build better evidence — by trying to do less terrible causal inference — using #RStats (I like DAGS) and have recently jumped on the bayesian meta-analysis 🚂

I feel like Babylon Berlin ticks a few of those boxes…

Great idea! Can you please add me too? 😀

DAG dabbler here 🙋‍♂️