From my email, a note about epidemiology

This is all from my correspondent, I won’t do any further indentation and I have removed some identifying information, here goes:

“First, some background on who I am.  After taking degrees in math and civil engineering at [very very good school], I studied infectious disease epidemiology at [another very, very good school] because I thought it would make for a fulfilling career.  However, I became disillusioned with the enterprise for three reasons:

  1. Data is limited and often inaccurate in the critical forecasting window, leading to very large confidence bands for predictions
  2. Unless the disease has been seen before, the underlying dynamics may be sufficiently vague to make your predictions totally useless if you do not correctly specify the model structure
  3. Modeling is secondary to the governmental response (e.g., effective contact tracing) and individual action (e.g., social distancing, wearing masks)

Now I work as a quantitative analyst for [very, very good firm], and I don’t regret leaving epidemiology behind.  Anyway, on to your questions…

What is an epidemiologist’s pay structure?

The vast majority of trained epidemiologists who would have the necessary knowledge to build models are employed in academia or the public sector; so their pay is generally average/below average for what you would expect in the private sector for the same quantitative skill set.  So, aside from reputational enhancement/degradation, there’s not much of an incentive to produce accurate epidemic forecasts – at least not in monetary terms.  Presumably there is better money to be made running clinical trials for drug companies.

On your question about hiring, I can’t say how meritocratic the labor market is for quantitative modelers.  I can say though that there is no central lodestar, like Navier-Stokes in fluid dynamics, that guides the modeling framework.  True, SIR, SEIR, and other compartmental models are widely used and accepted; however, the innovations attached to them can be numerous in a way that does not suggest parsimony.

How smart are epidemiologists?

The quantitative modelers are generally much smarter than the people performing contact tracing or qualitative epidemiology studies.  However, if I’m being completely honest, their intelligence is probably lower than the average engineering professor – and certainly below that of mathematicians and statisticians.

My GRE scores were very good, and I found epidemiology to be a very interesting subject – plus, I can be pretty oblivious to what other people think.  Yet when I told several of my professors in math and engineering of my plans, it was hard for me to miss their looks of disappointment.  It’s just not a track that driven, intelligent people with a hint of quantitative ability take.

What is the political orientation of epidemiologists?  What is their social welfare function?

Left, left, left.  In the United States, I would be shocked if more than 2-5% of epidemiologists voted for Republicans in 2016 – at least among academics.  At [aforementioned very very good school], I’d be surprised if the number was 1%.  I remember the various unprompted bashing of Trump and generic Republicans on political matters unrelated to epidemiology in at least four classes during the 2016-17 academic year.  Add that to the (literal) days of mourning after the election, it’s fair to say that academic epidemiologists are pretty solidly in the left-wing camp. (Note: I didn’t vote for Trump or any other Republican in 2016 or 2018)

I was pleasantly surprised during my time at [very, very good school] that there was at least some discussion of cost-benefit analysis for public health actions, including quarantine procedures.  Realistically though, there’s a dominant strain of thought that the economic costs of an action are secondary to stopping the spread of an epidemic.  To summarize the SWF: damn the torpedoes, full steam ahead!

Do epidemiologists perform uncertainty quantification?

They seem to play around with tools like the ensemble Kalman filter (found in weather forecasting) and stochastic differential equations, but it’s fair to say that mechanical engineers are much better at accounting for uncertainty (especially in parameters and boundary conditions) in their simulations than epidemiologists.  By extension, that probably means that econometricians are better too.”

TC again: I am happy to pass along other well-thought out perspectives on this matter, and I would like to hear a more positive take.  Please note I am not endorsing these (or subsequent) observations, I genuinely do not know, and I will repeat I do not think economists are likely better.  It simply seems to me that “who are these epidemiologists anyway?” is a question now worth addressing, and hardly anyone is willing to do that.

As an opening gambit, I’d like to propose that we pay epidemiologists more.  (And one of my correspondents points out they are too often paid on “soft money.”)  I know, I know, this plays with your mood affiliation.  You would like to find a way of endorsing that conclusion, without simultaneously admitting that right now maybe the quality isn’t quite high enough.

Comments

Civil engineering doesn't pay very well. Neither do math teachers. Does this mean we should discount more the opinions from someone who dropped out of an epidemiology program because they are broke? Bill Gates has the most money and his opinion is to shut the down the entire global economy. He also votes for Democrats too and probably hates Trump too.

There are many other jobs for math majors other than teacher.

They're all on the left and ALSO have low IQ's?

Epidemiology doesn't seem like a good field to be in because it is so easy for the average person to wave off any advice as scare-mongering for bucks or influence and then downplay any potential problem as "just a flu". The only time your opinion matters is when a once a decade epidemic hits a small part of the world or a once a century pandemic hits the whole planet. Until then how do you justify your value to the economy?

> Until then how do you justify your value to the economy?

In one sense we are lucky with COVID-19. It has spread basically everywhere and will cost *trillions*, but we'll survive. After seeing this unfold, it's going to be much easier to justify higher wages and more investments for epidemiology & immunology (a medical field that is also unglamorous and underpaid)

"how do you justify your value to the economy?"

Obviously very badly.

Trillions have been spent to deter a nuke or two hitting the US launched on a rocket built by a country with 100 million people or less, with 90% not having more than 8th grade education.

How bad would a nuke dropped on the US actually hurt the US economy?

But epidemics are a common occurance and Ebola is clearly not out of the question in many nations.

But do we spend 10% defending against a pandemic as scary as Ebola as spent defending against an Iranian or DPRK nuke dropped on the US?

Speaking of epidemiological modeling - is there a reason why it can't work based on attempting to simulate, say, an entire city interacting between each other rather than attempting to use various mathematical methods? Obviously this would require insane computational resources, but seems doable with a supercomputer and it's not like you need the simulation to run in realtime anyway.

Or is that hindsight thinking now that we know how bad the existing models are at predicting anything?

@Myst: what you propose is called agent-based modeling, does not require 'insane' computing power beyond a PC, and usually reduces to the analog formulas derived from closed form
first principle solutions, e.g., the agent- based, digital Black-Scholes formula in options trading. It's actually easier to derive this in digital form than analog form. About six months ago I wrote in C# a Monte Carlo simulation, using a Cholesky decomposition, on the "fallacy of time diversification" that AlexT has blogged about, with some shocking predictions (you can go bust with a balanced portfolio of stocks, even in a permanent bull market; rare but happens)

Bonus trivia: I endorse the OP, it's something I could have wrote. Note the OP dropped out of science and went to work for Wall Street. Smart move the way present society is structured.

I'm talking of "SimCity" level modeling, where we attempt to simulate what every single citizen is doing in the city, rather than using Monte Carlo simulations.

Myst,

Ray is correct. There are tons of epidemiological models that are basically network and agent based models.

Read: Prof Matthew Jackson Social and Economic Networks, or his recent book, Human Networks. For modeling, you might want to read Scott Page's book on models, Model Thinking.

There's not a lot of meat on this bone.

Q: What is an epidemiologist’s pay structure?
It would be more honest to say this person doesn't know. They only have anecdotes. "Presumably there is better money to be made running clinical trials for drug companies." Presumably this person is guessing.

Q: How smart are epidemiologists?
It would be more honest to say this person doesn't know. They only have anecdotes. GRE scores are usually presented in the form of numbers. None were given. "looks of disappointment" on your professors' faces is not data.

Q: What is the political orientation of epidemiologists?
Tired of repeating myself. In a country with secret balloting, yet this person knows how everybody else voted.

So this person took a few classes on the topic in school some time ago but is now an expert in a field they've never worked in nor graduated in. There's literally no data, just surface impressions and shallow anecdotes. They don't even have the best kind of anecdotes--the secondhand stories from actual epidemiologists in the field who they know of. I'm calling bullshit on this one.

-1 @Alexis - Your post simply says anecdotal evidence is not definitive evidence. Yet if you understood statistics you would know you cannot prove anything to 100% using statistics: any correlation can, even with six sigma confidence, be due to chance alone. So your own criticism rebuts you.

Pointing out that one can never reject a null hypothesis with p=1.0 is irrelevant to Alexis’ comment. If you understood statistics you would know that one person providing anecdotes is 1) A very poor indicator of trends at a population level and 2) not even statistics.

Both pay and intelligence are things people don't share freely. You can make an educated guess but it's not necessarily correct. How often will a professor tell you their exact pay and all of the side money they make? Or what they scored on the GRE?

However often people wear their political orientation on their sleeves.

It's a well-known observation with mounds of data (e.g. survey and otherwise) that academics are OVERWHELMINGLY Left-wing. You don't need to break the rules of secret balloting to know how people vote - especially those in institutions of the Progressive Cathedral. You don't honestly think that there are a bunch of MAGA-hat-wearing medical students roaming the halls of Harvard, do you? Especially when it comes to a field that lends itself to impulses of socialist central planning like epidemiology - they get to play god with society and wave around their fancy degrees in everyone else's faces and say, "Shut up and listen to me!"

They're little tyrants with big egos in lab coats - let's not pretend otherwise.

No, no, no: We must ask about hypotheses, not about people!

Ask these same questions of academic economists.

I doubt that epidemiologists have changed in the last few years. They’ve likely always overreacted to medical threats. They’ve likely always had little sense of proportion when it comes to policy and the value of life.

The real question is how they acquired so much power in this instance.

Let's have an academic death match. Who's the best at formal modeling this? Physicists, chemists, epidemiologists, economists, financial quants, climate scientists, biologists, or computer scientists? That's really what this is about isn't it? The econ guys here think they are the best because they take into account human behavior. Epidemiologists think they are the best because this is exactly what they study. Physicists, computer scientists, and quants think they are best at everything. No need for GRE scores and other weak forms of gatekeeping. Put up your models and data on Kaggle or Github. You have the whole shutdown to work on this.

Sounds like the start of a good bar joke :). But the lawyer or doctor would say: where's the money to be made from modeling it right? Other than fame, what's the point?

The post “ What does this Economist think about epidemiologists” got some reactions on Twitter.
https://twitter.com/ct_bergstrom/status/1249328619781939209?s=21

not just some...the butthurtdedness from epidemiologists is hilarious...

Private sector pay is better? News Flash! Error estimates are lacking? And, gosh, since we have global pandemics resulting in shut-downs every year, it's hard to understand why we can't estimate the errors of mostly ad hoc models. RIght?

Does anyone seriously believe that academic epidemiologists' models would beat a deep, liquid prediction market?

In February when the stock market was hitting daily new highs you could say the world's most visible prediction market failed to see this coming.

Is anyone else going to point out the entire thread of questions is unspeakably rude and meaningless? To follow up with this anecdotal email is ridiculous.

Tyler, you're better than this. Please stop and read some books.

Yes, it felt a little bit like an attack on the epidemiologists themselves ( how smart are they ? what’s their political affiliation? How much do they get paid ?) rather than a critique/evaluation of the output of the discipline.

In the other post: "I do not know the answers to the questions raised here, but I do see the debate on Twitter becoming more partisan, more emotional, and less substantive."

So Ty, what exactly did you think was going to happen when you questioned their intelligence and who they voted for?

Hopefully this results in a link to an epidemiologist blog by someone who is more informed about this stuff?
Either that, or the people who could actually answer these questions will not feed the troll.

( how smart are they ? what’s their political affiliation? How much do they get paid ?)

Commenter who noted this was all a rather pointed way to highlight that these questions are lobbed at economists got it right. What's good for the goose...

Economists deserve it though. They miss every single crisis.

Don’t they predict 7 of every 5 recessions?

AJ,
Agreed. This is a poor topic and reflects poorly on economists, particularly those who have not read anything on network analytics where they would see tons of epidemiological modeling.

I wonder if this discussion is designed to raise the status of economists at the expense of epidemiologists so that when the economists speak at the table as to when and how restrictions will be lifted economists will be the decider.

When you don't recognize that something is complex and needs inputs from everyone you are blind.

I think it's more helpful to summarise epidemiology and critique ideas, not focussing on the researchers.

Epidemiology core insights that (1) Covid-19 has a high enough pre-lockdown R0 to infect most and (2) R0 reductions to values still far above 1 buy you time without stopping most people getting it at the same time, whilst reducing R0 to below 1 causes it to peter out. These seem solid, useful conclusions that many of my smart friends, and their governments seem not to understand.

I perceive lots of uncertainty regarding exact IFR value, and the impact of various policies on R0 reductions. We should welcome brainpower to try and model our way out of this, or even just bootstrapping confidence intervals. Given the lack of analysis by statisticians/economist colleagues all now as obsessed with this topic as Tyler and I, there seems not to be low hanging fruit here, but hopeful some progress can be made. An easiest way to resolve these uncertainties is antibody testing on random population samples, and lab experiments.

A further interesting point of uncertainty that Tyler discussed before is the assumed large net benefit of hospitalisation for survival odds. I would like to see some more evidence confirming this.

Economists can work with epidemiology forecasts to propose tradeoffs of lives saved vs. economic damage. If we can quantify damage of various enforced measures, vs. baseline of consumers just being very scared.

The uncertainty here is very high too, but you may have a utilitarian argument for saying X trillion of GDP is more valuable than y 100k deaths. Would you shout this from rooftops? I think a key reason why epidemiology minimises death is because it's the only socially acceptable thing to say, rather than optimising a political swf. Normal analysis on economic value of a human life is pretty weird, and if you have up to millions of deaths on the left hand side of the scale it all starts to seem gross.

@Oli - nice OT rant, though this sentence fragment is technically incorrect: "whilst reducing R0 to below 1 causes it to peter out". No. Even smallpox will peter out when over 90% of the population gets it and hopefully recovers (herd immunity).

Hi!

What is OT rant?

Also sorry but I don't understand your point about herd immunity? Can you explain?

OT is Off Topic. The R0 point is that once herd immunity kicks in, then by definition R0 will drop below 1.0 and the infection peters out.

(Wikipedia): "The most important uses of R0 are determining if an emerging infectious disease can spread in a population and determining what proportion of the population should be immunized through vaccination to eradicate a disease. In commonly used infection models, when R0 > 1 the infection will be able to start spreading in a population, but not if R0 < 1. Generally, the larger the value of R0, the harder it is to control the epidemic. For simple models, the proportion of the population that needs to be effectively immunized (meaning not susceptible to infection) to prevent sustained spread of the infection has to be larger than 1 − 1/R0.[22] Conversely, the proportion of the population that remains susceptible to infection in the endemic equilibrium is 1/R0." - note R0 changes with time. Once herd immunity kicks in, R0 will drop below 1.0

"Data is limited and often inaccurate in the critical forecasting window, leading to very large confidence bands for predictions"

If this is true, it suggests that none of us should get too fancy (and that paying epidemiologist more isn't going to buy you a lot).

What if all you need to know is "at R0 greater than X, immediate action is required?"

Did the countries that jumped on immediate contact tracing do so because they were great modelers? I think not. I think it's because they already had life lessons.

Is it true that history, successes and failures in 1918, are still a better guide for success today than "computer models?"

I'm beginning to think we needed a serious, but simple, response. The K.I.S.S. principle might apply.

But sadly a serious K.I.S.S. response did not, still does not, fit the American intellectual or political landscape.

As Trump might say: 'K.I.S.S., my ass, my reelection is more important than older people who won't vote for me anyway'.

+1 Don't make it too complicated my head hurts

I did epidemiological research for a while. Most epidemiologists are not doing contagion forecasting - but that seems be what you think it is.

Pointing out such lacks, repeatedly, changes nothing. It makes no difference if you work in the field, cite various sources, or basically all commenters say the same thing.

This is an economics blog, so of course it already represents what two people already believe, regardless of any contrary facts.

"Most epidemiologists are not doing contagion forecasting". What do they do then ? (Honest question)

I singled out item one above, but items one to three are all very good. If California "beat the models" is it because item three?

"Modeling is secondary to the governmental response (e.g., effective contact tracing) and individual action (e.g., social distancing, wearing masks)"

https://twitter.com/EricTopol/status/1249481906749444097?s=19

K.I.S.S.

We have gyrated wildly from "everyone has already had the virus so we are close to herd immunity" to "it's mathematically impossible for a virus to persist in a population of which ~10% are immune".

I think perhaps epidemiology is quite complex a problem regardless of practitioner.

I don't think it was the actual epidemiologists who said either of those things.

The potential for very high fractions infected (as a possible upper boundary, not a certainty) did indeed come from academic epidemiology (unless all have forgotten the Oxford study).

Gee, I wonder why more academic epidemiologists don't support the political party that has been bashing academic STEM (and non-STEM, come to think of it) for decades? That doesn't make them left-wing, it just makes them not Republican.

A RINO, obviously. They are simply in denial that objectively, they are just as leftist as Bernie.

It's pretty much the nature of the political spectrum that anyone who is "group effort" goes left and anyone who is "individual achievement" goes right.

Epidemiologists are not going to be right-wing because individual achievement is never a rational answer, for their chosen field.

Economist bifurcate, and try to create "economic models" which try to put either group effort or individual achievement at the top of the hierarchy.

What's unstated here is that, unlike in economics, there's no markets vs. government intervention dichotomy with epidemiology. I detect a frustration among some economists (and observers) with this. Economists with a strong markets preference often challenge the opposing view by (1) attacking the messenger (bias), (2) attacking the messenger's methodology, (3) attacking the messenger's data, and (4) attacking the messenger's conclusions/recommendations. I suppose the markets alternative for epidemiology would be to (a) determine the appropriate course of action by aggregating the conclusions and recommendations of a large sample of epidemiologists to arrive at an average or (b) create a prediction futures market and let that market be the guide for determining the course of action. What we often observe is that those with a strong markets preference are the first to blame government and regulation for whatever problems are encountered or created. The shortage of ventilators is but one example. The reality is that the market for ventilators was manipulated by big business to prevent the development of a moderately priced and ample supply (supply and price being co-dependent) of ventilators. https://www.nytimes.com/2020/04/12/opinion/ventilators-coronavirus.html

Maryland's public health school is ranked 32 by U.S. News. These are the GRE scores of its students:
https://sph.umd.edu/department/bch/admission-statistics
Note: not all of them are epidmiologists.

The data you posted is for an MPH program. Few people with an MPH would try to call themselves epidemiologists. From the same site, a link of sample careers for MPH graduates:
Master's Level Sample Positions
Estimated starting salary range $40,000–$50,000 (the starting range can vary depending on the where the graduate finds a job. Graduates in the MD/DC/VA area can expect a higher starting salary)

Example career postings:

Public Health Educator II in a state health department
Research Program Manager at a state partnership against tobacco
Coordinator of Health Promotion Programs at a private college

This line of thinking is mystifying. As far as I can tell from news coverage, the epidemiologists were right about the direction of C-19, and the associated policy recommendations. True, it appears (hopefully) that many of them were off the mark regarding the magnitude of the virus’s human impact. But the existing impacts have been bad enough to radically reshape global society, so while these errors of scale are intellectually interesting, they don’t have much real world significance. Overall, I’m all for improving quality going forward, particularly since pandemics will likely be more common in the future, but it looks like the epidemiologists successfully performed a key social function during the current crisis.

It is 100% about politics, not about truth or science. One set of esteemed prognosticators called economists seek to keep their status elevated above another set of newly esteemed prognosticators called epidemiologists. When you don't like your political opponent you ask for their birth certificate or in this case their GRE scores.

Tyler Cowen is not an economist, even if he pretends to be.

+1 Seems to me that spread predictions and the effects of suppression by epidemiologists have been broadly accurate. Questioning how "smart" they are seems small at this point.

It seems to me that anybody in academia commenting on how 'smart' other persons are is simply trolling a pointless opinion.
Lots of us have too much time on our hands lately.

Big shoulder shrug here, but I'm glad you're seeking out this info, though I'd like to hear critiques from someone with more knowledge and expertise in the field. I didn't find this one very credible or informative.

It's like the interview one of you linked to about the behavioral economist who said something to the extent that to be good at behavioral economics (which is a huge critique of lots of canonical economics work) you have to really know the canon well. I'm not buying that this person meets this criteria.

So based on this, a recent undergrad didn't like his PhD program and found that academics lean left (or at least don't like Trump) and look down at other fields??? Cool.

I'm pretty sure that the average engineering professor is better at math than the average epidemiologist, and that the average economist is probably better at it, too, and better than lots of political scientists and public policy scholars. And the average physicist beats the average biologist. And mathematicians beat all of them. And???? On the margin, maybe that matters.

But this doesn't translate into needing to be dismissive of entire disciplines. There's still knowledge build up over years of study, which someone in a discipline spends incredible amounts of time exploring and examining and evaluating, and likely, the ideas of the best and brightest stand out over time.

I suspect the bigger issues in the field relate to this person's earlier point about data problems.

Should the signaling model of education continue to be endorsed here after our dear host chooses to use “very, very good school” and “another very good school” as descriptors? Anything to add Mr Caplan?

What absolute arrogance. Who is smarter? Who has the best test scores? Amazing. Such insight into your minds. Such rhetoric which hides nothing more than pure Credentialism. It would be funny if it was not so deadly.

it is puzzling how highly trained people, particularly in designing well-thought studies, make sweeping generalizations and disregard heterogeneity of the practitioners in a field when stating their opinions. like many disciplines, the areas of epidemiological research, skill sets and backgrounds of epidemiologists and possible uses of epi toolsets are diverse. imo this discussion focuses only on a particular application of the discipline.

I'm not sure what "smart" means. If epidemiologists can persuade the people and government to do the right things and not the wrong things that's plenty smart enough.

"certainly below that of mathematicians or statisticians"

ha ha ha

i know what he means

but .....

but he is talking about people who do it for a living not the hobbyists and the inspired among us

who do it because we can

(George Theodore)

(the George Theodore reference is fairly obscure.
I am no big fan of the Seinfeld Universe, but in that universe one of Cosmo's dreams is to just go out and run a fire truck, even though he had no experience or training - and he did it ---- this was based in part on the childhood memories of Mets fans of Jerry Seinfeld's age, who were told that George Theodore was just some guy who was not even an athlete, but just decided he would like to play third base for a NYC team for a year or two, on a lark - get it?).

(the fact that so many crazy people on Twitter and the internet have been spectacularly right about the latest coronavirus for a while, and then have turned into dangerous imbeciles, or vice versa, is the underlying factual substrate to my APRIL !# 10"02 PM COMMNT!!!!)

if you could come up with ten or twenty similarly plausible factual substrates to obscure references in say, Euripides or Sulpicia, or Isaiah or Ruth, you would be a great Biblical or classical scholar)

all in all, the epidemiologists as a group have been pretty good on this, for the record, in my humble (not a hobbyist, not inspired) opinion
(they are not responsible for the comically evil power grabs - policemen chasing down solitary runners on isolated beaches, that sort of thing)

one last reference, from Carl Reiner, I think

"you can say you are a comedian, but that is like saying you are an all-star third baseman, if you aren't one, it does not matter what you call yourself"

although whoever said it first said it better

Comments for this post are closed