Kyle Onda

@cjwich @gregspierce I don’t disagree with anything you’re saying. I’m on the board of a water utility that often refuses to do anything of the sort because of state laws. My comment to AR was to remind him of an issue to surface these with people who can put this on the policy agenda

Always welcome more empirical work on community water system rate setting, this time by @cjwich @rff, especially using publicly-available @EFCatUNC data

@gregspierce @arelkhattabi Remember that one pager for NC we talked about @arelkhattabi

@gregspierce Also this is legally suspect in many states, including our respective two

Although Footnote 2 gives me a lot of heartburn, but no one reads threads on here anymore, so why bother elaborating. And noting the role of enterprise funds, especially in contexts beyond the SE U.S., certainly matters.

If you're keeping score at home, new @USGS_Water national water system boundary file dropped. Differences from @internetofh2o @GoSimpleLab @EPIConservation's boundary tool described ⬇️. Soon to be joined by a 3rd set of national shapefiles from @EPAwater...

@AdanBecerraPhD @unboxpolitics @jenniferdoleac @DinaPomeranz its a somewhat growing field to do large RCTs that measure mediators and do SEM- or IV0-type modeling on top of them. (doi.org/10.1186/s13063…) full factorial mediator trials are probably just too expensive for now, but in principle I could see demand for them growing

Some people falsely believe that papers using RCTs can only test whether something works or not, but cannot shed light on the underlying mechanisms. What are great papers to use as examples to illustrate that this is not the case? #econtwitter #followerpower

@AdanBecerraPhD @unboxpolitics @jenniferdoleac @DinaPomeranz to your point, mediators need to be manipulable in a potential outcomes framework for it to make sense to use an RCT for, and these are very rare...particularly in biomedical work, and maybe slightly less so in econ/social. ie doi.org/10.1111/rode.1…

@AdanBecerraPhD @unboxpolitics @jenniferdoleac @DinaPomeranz factorial designs are relatively common. eg sciencedirect.com/science/articl… Specifically with this mediator setup research question, no, not beyond designs that attempt to remove placebo-type "mediator" effects (e.g. #s2" target="_blank" rel="nofollow noopener">journals.plos.org/plosone/articl…) 1/2

@unboxpolitics @AdanBecerraPhD @jenniferdoleac @DinaPomeranz Right so factorial RCTs aren’t that out of the ordinary. And they could reveal some bundled simultaneous causal mechanism with enough theory/assumptions. But what they don’t do is help figure out a mediator. You need to do trials manipulating mediators independently.

@adnosk @AdanBecerraPhD @jenniferdoleac @DinaPomeranz Yeah that makes sense. I think a more interesting example of the kind of design I have in mind is this paper. x.com/unboxpolitics/…

@unboxpolitics @AdanBecerraPhD @jenniferdoleac @DinaPomeranz makes sense thx. I think the disconnect might be about the placebo concept. 1st-order mediators like time availability and disposable income seem uninteresting for many behavioral outcomes, similar to med. treatment placebo effects. So trtmt arms about that should be std practice

@adnosk @AdanBecerraPhD @jenniferdoleac @DinaPomeranz The most likely channel would be non-cognitive skills. I'm spitballing a bit here but it seems like an apprenticeship could lead one to develop discipline, attentiveness, motivation to learn, and/or problem-solving skills, and any one of these traits could lower criminality.

@unboxpolitics @AdanBecerraPhD @jenniferdoleac @DinaPomeranz Then This is an RCT of a jobs program (which is an IV for human cap), with a quasi placebo arm. This does not clarify a causal mechanism with much granularity. Genuinely curious ,what’s the hypothesized causal mechanism from human cap to reduced crime conditional on dispos. Inc?

@adnosk @AdanBecerraPhD @jenniferdoleac @DinaPomeranz Happy to clarify! One could equalize pay by using an unconditional cash transfer to net out the difference in pay between apprenticeship vs gig economy jobs. This would hold disposable income constant while allowing human capital development to vary.

@unboxpolitics @AdanBecerraPhD @jenniferdoleac @DinaPomeranz Like, what does it even mean, in practical implications, to set the pay equal between the human capital and unskilled jobs program, if a causal mechanism from human capital to decereased crime probably includes increased pay?

@unboxpolitics @AdanBecerraPhD @jenniferdoleac @DinaPomeranz A paralell critique to what @AdnBecarraPhD is talking about, is basically that doing a complex RCT designed for a mediator boils down to using your treatment as an IV for the mediator. Are you really sure about the exclusion assumption?

Thanks @chapelhillgov & @CarrboroGov for proclaiming Nov. 16 "Care to Share Day"! Our Care to Share program in partnership with @IFCforSS helps members of our community with their water bill when experiencing hardship. Donate at owasa.org or: ifcweb.app.neoncrm.com/np/clients/ifc…

@gregspierce NC equivalent is GS 160A-314. So hard to deal with.

Prop 218 is an enormous equity obstacle, part 3,491

@BradWilcoxIFS I wonder if the mechanism is informational…makes it easier to make a coherent budget together…or more about the actual interdependence. My wife and I never had an actual joint account but we both have each others banking info and see it all together in the same web app.

Couples w/ joint checking accounts are happier. But correlation or causation? Causation. Fascinating new *experimental* study assigned some newlyweds separate & other newlyweds joint accounts. Joints did much better in their marriages. academic.oup.com/jcr/advance-ar…

Attending @foss4g? Interested in the latest advancements in open #waterdata and standards? Don't miss these two sessions with @neb_webb

@gregspierce I feel this so hard

Felt pretty good about responding to a funding proposal prompt on length of project: "A lifetime, but also, 2 years"

Really pleased that this is out on the street. OGC Engineering Report: "Hydrologic Modeling and River Corridor Applications of HY_Features Concepts" docs.ogc.org/per/22-040.html

Congrats to a great friend and researcher!