Joe Bak-Coleman

@jbakcoleman.bsky.social

1257

Followers

328

Following

477

Posts

Associate Research Scientist at Columbia journalism. Harvard BKC affiliate. Comp. soc. science, collective behavior, stats

Posts Replies Media

Reposted by

Joe Bak-Coleman

Today, this brief paper of mine was published in the Journal of Responsible Innovation. I'll do a little thread when my annual leave is over! #STS #metascience

Scandal in scientific reform: the breaking and remaking of scienceurl6649.tandfonline.com This perspective explores the Scientific Reform Movement and its links to scandalized claims, such as ‘science is broken’. It delves into the pivotal role of scandal in shaping and sustaining this ...

Publishing across disciplines, there is remarkable variation in the norms surrounding what constitutes a competing interest in peer review. In animal behavior reviewing/editing a recent co-authors work is very taboo, but seems fairly common in social science.

Reply to

Ira Hyman

For a short term gig, I’m currently scheduling a doctors appointment to verify that I’m sufficiently healthy and have the required vaccines work at my duty station, a desk in Manhattan or at my home in Brooklyn. I truly don’t mind but it’s going to be a fun chat with the doc.

My rabies vaccine is out of date by a couple years. Fortunately it’s been a while since a squirrel was biting people in the park.

Fortunately my access to rapid medical care for a rabies bite is somewhat faster in the city than it was in the field.

Reply to

the wrecked and abandoned vessel

Between the first and second cup.

Waking up to having a manuscript accepted is almost as good as a cup of pour over.

Reply to

David M Rubin

Your point is well taken.. the speed of sequencing and cost these days is absolutely insanely good these days.. That said, it's still generally more expensive and time-intensive than adding 1 to a MTurk survey. If it's an experiment, there will often be some intensive prep beforehand as well.

Reply to

Ira Hyman

Oh absolutely. I think my main concern here is the assertion that replication is easy (or even possible) across science is conditioned on replication-on-mturk rather than everything else. Certainly would be hard to replicate a multi-decadal longitudinal dataset.

Reply to

Ira Hyman

Of course there are easy and difficult studies in any discipline, but I think it's reasonable to note that some areas/disciplines tend to implicitly vary in the constraints, costs and effort required. For example, field biology entails travel, permits, seasonality, finding the species, etc...

I’ve said this before, but there is a deep need in metascience to understand the practicalities of science across disciplines and areas of research before asserting what should be done.

Listening to a podcast on Replication and this comedically backwards bit caught my ear: "in biology, you could do a lot of experiments without much time and resources going into them... in psychology, replicating things is not as trivial as sitting on a lab bench".

I've done bench, lab, and field work in biology and conducted experiments in the social science. I was *astounded* at how much easier data collection is when you can upload a .qsf and get 1000 response for a few thousand bucks inside an afternoon.

Hell, my first set of experiments involved a summer building a flow tank, 4-6 hours setup and running a trial, 2 hours of staining and imaging then all that again for tracking on custom written tracking software. Two full days with two people for one sample. $300k in equipment and fish to start.

It wasn't all that complex as these things go but the best posed lab to replicate it would have burned a year and $50K to replicate it. We used some legacy equipment (particularly an impeller motor) that might take quite a bit of time scouring ebay and paying someone wanted.

Why does this pigeon look like landed gentry?

Feeding ecology of the common wood pigeon (Columba palumbus) in a major European city | doi.org/10.1098/rsos... | Royal Society Open Science | #ornithology

Reply to

Jay Patel

I like to think rigor is more than something you can capture in a taxonomy of emoji

More seriously each of these are choices that impact inference in distinct ways and neither inherently nor universally good or bad, not equally feasible across the questions science asks and importance of those questions. A badge tosses all this context and domain knowledge away.

Reposted by

Joe Bak-Coleman

Writers: Challenge the agreement/contract that your publishers present you with. They'll start off saying this is our contract, take it or leave it, but it's negotiable. Every contract is. 'Boilerplate' means this is where you start, not where you finish. And learn what 'subsidiary rights' means.

Agents can be worth their weight in gold, if you’re able to sign with a reputable one.

Not for everyone, and I’m sure there are horror stories and caveats I don’t know about. As a scientist wading in, it’s been a godsend.

Reply to

Everyday.Human Derek

Above my pay grade for the butterfly but good luck. Glad you've enjoyed the book so far.

Reposted by

Joe Bak-Coleman

#statstab #133 p-Curve and p-Hacking in Observational Research Thoughts: "p-values in observational research may need to be empirically calibrated to be interpretable", or you know, stop using them in obs research 🤷‍♂️ #research #pvalues #observational journals.plos.org/plosone/arti...

p-Curve and p-Hacking in Observational Researchjournals.plos.org The p-curve, the distribution of statistically significant p-values of published studies, has been used to make inferences on the proportion of true effects and on the presence of p-hacking in the pub...

Reply to

Mark Rubin

I think this is a problem with meaning two or more different things when scientists say false positive. Ioannidis has definitely made this mistake before!

Reply to

Carl T. Bergstrom

Honestly any ioannidis paper that isn't about his CV or COVID is a breath of fresh air.

Reply to

Mark Rubin

Honestly someone putting out that many papers has to eventually have a good take.

Reply to

Steve Haroz

With enough within-subjects samples, that assumption of normality becomes very Principle of Maximum Entropy. Consider me unprovoked.

Really formative paper for me seems to have fallen. Imo this is an example of good sleuthing, measured critique, and raises important questions for reconciling visual odometry research following this paper. arxiv.org/abs/2405.12998

The miscalibration of the honeybee odometerarxiv.org We examine a series of articles on honeybee odometry and navigation published between 1996 and 2010, and find inconsistencies in results, duplicated figures, indications of data manipulation, and...