predictions: data adulteration is the coming thing
when state sponsored agencies lose on level playing fields, they seek to slant the fields of the future. we must not let them.
much has been made (including by certain notoriously loquacious internet felines) of the emergence of a new class of distributed swarm sourced data analysts whose adjacent expertise has flowed into new fields cross pollinating them and challenging their assumptions and authority.
a great deal has been learned by a great many people.
while we have spoken at length about the learnings from the analysis, we have not delved overmuch into that which was apprehended on the other side of this ball. and i think we should, for the realizations of the challenged about these new challengers shall determine their responses and if you think recognition that the cloistered clerisies of academia and policy cannot win in a straight up fight is going to cause them to lay down their power, prestige, and prerogative and admit they were outmatched, boy are you in for a surprise.
because that’s not how it works.
those who scaled to the tops of the greased poles of these lofty, credentialed perches have more cards yet to play and they too know the time honored dictums of winning fights and ends justifying means:
“if you find yourself in a fair fight, your tactics suck.”
so do not expect fair. expect tactics. and anticipating them is the first step to countering them.
so what did they learn?
i think the chief learning, fear, and driver are all essentially the same:
“with the emergence of vast and highly capable militias of non-affiliated data analysts, false narratives cannot be sustained atop good data.”
there are too many threads to pull, took much work that will be done, and too widespread a means to sharpen, validate, and communicate findings.
they cannot win on this field. so they won’t play there.
the game has shifted and will shift further.
the battle will be won or lost on data suppression and adulteration.
you cannot outwit or out-analyze the bazaar from within the cathedral. the only solution is to keep the mass in latin and the holy words away from the congregants. i suspect data embargo and tampering (already rife) will become absolute gold standard operating procedure in short order.
longtime gatopal™ bachman, ever the puckish provocateur, asks some useful questions:
it’s certainly not lack of funding.
it’s a lack of desire to get at real outcomes.
and this is an important dictum:
mostly, when you see this sort of enduring wrongness and incapacity to handle simple tasks with adequacy, it’s not because no one can solve the problem; it’s because someone wants it that way.
if the CDC wanted real mask studies or lockdown assessments or RCT’s on a wide variety of early treatment and prophylaxis regimens, we’d have them.
their continued non-existence amidst trillions in expense and largess can ONLY mean one thing: they did not want them.
the united states is the richest, most technologically advanced country in the history of countries and technological advancement.
can anyone tell me with a straight face that if we were determined to count election votes in real time with something approaching zero fraud or ineligible vote casting that we could not do so?
of course we could. and the fact that we don’t stands irrefutable accusation.
someone does not want it that way.
the CDC has lots of data it’s not sharing. series after series of useful info (like UK cases per 100k by age and vaxx status) keep getting cancelled when the conclusions that may be drawn from them grow too counter to policy. when too many vaxxed and boosted land in hospitals from covid, they change the definition of “hospitalized for covid” to reduce the count just as they rigged the definitions of vaccinated in the drug trials (dose 2+14 days only) and rigged the per 100k counts in states by imputing the number of unvaxxed using old census data and subtracting the number of known vaxxed from it resulting in persistent undercounts of the unvaxxed and thus higher prevalence rates.
from alberta to albuquerque, the fix is in and the data, especially the US data, is being turned into trash such that it may not be accurately assessed by outsiders.
if you want to analyze it, be my guest. we’ve already made the outcomes forgone by the way we slanted and tainted it.
if this were just simple error and incompetency, the error would be two tailed. sometimes it would make it look worse, sometimes better. the fact that the “bank error” always seems to be in the bank’s favor speaks volumes about the deliberateness of the manipulation.
the deck is stacked with made up cards.
and if you think this is bad, you ain’t seen nothing yet. many of these modeling and distortion tactics were learned or adopted from the climate science gang. and if you want to see the real masters of data adulteration, that’s the team to watch.
the “carbon scoring” they use for ESG measures and lease prices is invented from whole cloth.
the “cost per kilowatt” of their favored renewables plans is slanted to the point where you cannot stand on it.
even the global temperature data they use to speak of tenths of a degree of decadal rise in heat has error bars 2 degrees wide that are entirely slanted to read warm and even in spite of that, they cheat and cool the past to make the rise look more prominent.
many years ago i undertook a large scale study in the climate space. i had no agenda apart from “find out how real this is and make a determination about whether getting into carbon trading in a big way would be a good idea for a large investment entity.”
i thought this was going to be easy. i thought the science more or less was settled and the data and models sound.
i was very, very wrong.
after reading 100’s of papers and speaking to many of the top scientists in the space on every side of the issue, i came away staggered at what a towering, tottering mountain of error bars the space was. the claims were outlandish, mostly unfounded, and nothing like the level of validation you’d want to make such important policy. the models had no predictive ability and even simple issues like “do clouds provide positive or negative feedback to temperature rises?” were not understood.
but what REALLY stunned me was the measurement. the data itself is absolute, stunning garbage, everyone serious in the space knows it, and no one cared. because the slant was all in one direction and that was the direction that enhanced crisis and therefore money.
one of the first big crowdsourced verification projects of government science i ever saw was in this space. it was run by anthony watts, a great guy and a prolific publisher of ideas and data. it was called SURFACESTATIONS.ORG.
the mission was simple:
there are 1221 weather/temperature monitoring stations in the US called the USHCN network. this is where reported US temperatures over time come from.
and literally no one had EVER gone out to see if the sites were accurate or sited in accordance with CRN guidelines. this is a big deal because the thermometers are greatly affected by surroundings. a reading in the middle of a parking lot of black asphalt can be dramatically different from a grassy field 300 yards away. this is not climate change, it’s local environmental factor.
and you can see the way this plays out:
the CRN has a set of strict and specific guidelines for station siting. this is a quick overview:
Climate Reference Network Rating Guide - adopted from NCDC Climate Reference Network Handbook, 2002, specifications for siting (section 2.2.1) of NOAA's new Climate Reference Network:
but no one had any idea as to whether and to what extent sites complied. anthony and a large group of volunteers decided to find out. using FLIR cameras, measuring tapes, digital photography, and a lot of good, old fashioned shoe leather they went out and surveyed 1007 of the 1221 sites.
what they found is nothing short of shocking.
categories 1 and 2 are deemed acceptable by CRN. less that 8% meet spec.
over 70% are category 4 or 5, injecting a one sided variance of >2 degrees C (and over 5 degrees in the 5’s).
running 3 degrees hot in a system trying to measure tenths of a degree on decadal timescales means your data is junk. it runs outlandishly hot. and there is no way to adjust or control for it, especially if you have never surveyed these sites yourself. (and even to this day, they have not)
how can this possibly be used in earnest to measure temperature?
conclusions drawn from this data are forgone. the systemic upward bias in it is so high that it dwarfs any signal.
(it also gets used as calibration sets for satellite systems, so this is not independent in the manner often portrayed and satellite instrument management is surprisingly (at least to me) inaccurate and impressionistic.
if they wanted good data, they could have it. if they wanted to publish good data, they could. but they don’t. they publish the USHCN data that carried with it impossible to adjust for warming bias.
and here’s what makes this especially galling: they HAVE the data. it’s called the US climate reference network. it’s run by the NOAA.
Each station is positioned in a pristine site which is expected to remain free from development over coming decades. Each station may include the following sensors: triple redundant air temperature sensors, precipitation sensors, wind speed sensors, and ground temperature sensors. Stations have been placed in rural environments in order to avoid possible urban microclimate interference. The contiguous U.S. network of 114 stations was completed in 2008. There are two USCRN stations in Hawaii and deployment of a network of 29 stations in Alaska continues. 
it’s data only goes back to 2004, but in those 18 years it shows NO meaningful US warming nor warming trend.
quite striking, no?
what’s even more fascinating to me is that while they publish this data, it’s not in a format that 99% of people can read. there is no graph, no chart. it’s all buried in csv etc. there is, buried a couple of pages deep, what purports to be a “visualizations” link but i tried it in 3 browsers and cannot get it to work.
i got this chart from an aggregation done by anthony watts on his wonderful site wattsupwiththat.
meanwhile, the “scare charts” known to be diverging from the much higher quality “reference” data that is literally produced by the same agency
pause and really think about that.
you have 2 datasets, one designed to be accurate and precise and one known to be riddled withy one sided error that accentuates warming trends.
you choose the latter to publish and foreground and as the basis for most “scholarly” research in the space.
you taint an entire body of endeavor and inquiry in full knowledge that you’re doing so because most others don’t know you’re doing it.
this is not how science is done.
this is how cronies are capitalized.
this kind of science runs on grants and sinecures, public positions and academic tenure and so, just as anywhere else, follow the money.
words you will never hear:
“we the experts empaneled to study the crisis have determined that it is not, in fact, a crisis. please accept the return of the rest of our funding and put it to better use.”
and climate is the king daddy of cash. the money that flows though ESG, green energy, green mandates, climate studies, EV’s, wind, solar, carbon tax, and every other form of “mitigation” predicted on the idea that climate is changing, man is causing it, and that not only would that be a bad thing (debatable) but that stopping it is better/cheaper/more plausible than adapting to it (a difficult and implausible claim to support) makes covid look like a lemonade stand.
the incoming SEC driven climate reporting standards will both be cruelly expensive and deeply dangerous as a new pretext for punitive pigouvian taxation and showers of subsidy. it will enable powerful new levels of corporate steerage and control by unaccountable technocrats
zero carbon is the original zero covid. it’s the same poorly calibrated impulse to ignore costs and trade offs and presume the efficacy of interventions for power and profit.
and it will be driven by the same misrepresentation of fundamental data. “the USHCN says so” is just “you tested positive for covid at 40 Ct PCR 22 days before you died of cancer, mark it covid death.”
this is how the data monks of technocracy trap you.
and it’s how they will seek to use definitions and datasets to make accurate analysis by outsiders impossible.
and it’s why they won’t let you near it or inside their closed loop echo chambers of “friendly peer review.”
and it’s why we need another kind of scientific revolution.
if we are to have federal data, then it must be independent data.
the agencies must be open, observable, auditable, and utterly separate from any and all policy arms.
once, perhaps, this would have posed challenges. now it’s trivial.
the only way to do real science is to have real data and the only way to assure such data in the face of the adulteration incentives of government agencies is full and fulsome sunshine everywhere.
open all sources, all methods, all assumptions.
make peer review real and open ended.
peer assessment is not the finish line, it’s the starting line. repetition, replication, refutation, and revision are the road to truth. everything else is skullduggery, superstition, and credentialism driven fraud pretending to be the data and practice of scientific method.
we have a citizen army of analysis. let’s use it.
break open the cloisters and let both sunshine and citizens alike inside.
and if “not wanting it that way” is more important than the truth, well, then why on earth would anyone want to trust people like that?
you’re done here. we have no place for you.
you just told you that lying and control take precedence over veracity and reality.
this is truly a binary:
you’re either on the side of open, honest inquiry or you’re the other guys.
stop settling for less from those who would dictate the course of your lives to you by dominating and distorting the nature of and inputs to science.
human flourishing resides in choices and trade-offs and no one can assess such without good data on which to base decision.
free the data, and you lives and livelihoods will follow.