Become a fan of Slashdot on Facebook

 



Forgot your password?
typodupeerror

Comment: Non-Falsifiability (Score 3, Interesting) 157

by dorpus (#49230679) Attached to: Strange Stars Pulse To the Golden Mean

Similar claims have been made about how human anatomy allegedly conforms to mathematical constants. But when we make actual measurements of individuals, nobody fits the constants perfectly. What is the allowed margin of error? One can make just about any number be close to some "elegant" mathematical constant -- pi/2, pi^2, e/phi, whatever.

Similarly, today I just judged a paper about childhood obesity submitted to a scientific journal. Childhood obesity is confounded with low socioeconomic status, so how do we separate the two? Of course, children of lower socioeconomic status have poorer outcomes in terms of health, occupation, and mortality. (Incidentally, the children with the worst outcomes in terms of future health, income, and mortality are the underweight kids who look like walking skeletons. Most scientific papers on obesity exclude that population.)

Comment: Re:Narrowing the context (Score 1) 60

by dorpus (#48956177) Attached to: Test Shows Big Data Text Analysis Inconsistent, Inaccurate

Yes I did. There were a few thousand responses that fit on a single spreadsheet, and after an hour spent coming up with buckets and keywords for them, I couldn't find any exceptions from the above. I'll keep an eye out for future changes, though I doubt they will change much. I know the hospitals and their problems.

Comment: Narrowing the context (Score 2) 60

by dorpus (#48952985) Attached to: Test Shows Big Data Text Analysis Inconsistent, Inaccurate

I analyzed the free-text field on hospital surveys. A simple keyword search gave me very reliable results on what the patients were complaining about -- they fell into the categories of bad food (food, cafeteria, diet, tasted, stale), dirty rooms (dirty, rat, blood, bathroom), rude staff (rude, ignore, curt), noise (noise, loud, echo, hallway), TV broken (TV, Television, "can't see"). So if the context is narrow enough, even simple searches work.

I agree that more broadly worded questions require more sophistication. I've looked at word combinations and so forth, though I haven't really needed to use them yet in analyzing health care data. We would not trust a computer to parse a full doctor's report, no matter how sophisticated the software; that will require manual inspection, often by multiple people to agree on a consensus interpretation.

Comment: Spontaneous Quitting (Score 1) 178

by dorpus (#48886251) Attached to: New Nicotine Vaccine May Succeed Where Others Have Failed

Substance addicts will often spontaneously quite their habit when the pain of continuing the habit becomes greater than the pain of quitting. A year ago, I had severe stomach pains and was hospitalized for 3 days. I figured the chewing tobacco was upsetting my stomach, so I went cold turkey. As it turned out, it had nothing to do with the tobacco -- it was intestinal colitis. Anyway, I'm off of nicotine permanently now.

Comment: Except when scientists do it (Score 1) 497

by dorpus (#48875249) Attached to: Science By Democracy Doesn't Work

When mathematicians vote on whether to accept a new theorem, when psychiatrists vote on which diseases should be included in the latest version of DSM, when NIH panels vote on whether to fund a grant. No, science couldn't possibly be run by the tyranny of the mob that refuses to believe in ideas that are too new and radical.

Comment: I was in the same boat (Score 2) 280

by dorpus (#48612573) Attached to: Ask Slashdot: How Should a Liberal Arts Major Get Into STEM?

I graduated at the end of the Cold War ('93), so an engineering degree was worthless -- all the companies were laying off their engineers as quickly as possible. Combined with the fact that the engineering jobs I interned for or heard about were not very interesting (managing a chemical factory?), I got a liberal arts degree. I went into IT for about 10 years, but in the long run I just didn't care that much about the mechanics of computers. I eventually got a PhD in biostatistics after taking the prerequisite courses. Statistics has let me get into various different research projects without having to overspecialize. I work for a hospital system now and do different research studies every day.

Comment: Aren't we supposed to be dead already? (Score 1) 28

by dorpus (#48266535) Attached to: CERN Looking For Help Filling In the Gaps In Photo Archive

Growing up in the 1970s, there were many books by reputable scientists who were absolutely sure that the Earth will have mega-famines and enter an Ice Age by the year 2000. They were absolutely certain that no economic growth will ever occur in China or India, the Cold War will last a thousand years, and air pollution will require everyone to wear oxygen masks at all times.

Comment: What if it will kill 100,000 people instead? (Score 2, Interesting) 186

by dorpus (#47333937) Attached to: Larry Page: Healthcare Data Mining Could Save 100,000 Lives a Year

I work as a statistician for a hospital chain. We already do data mining and have interventions for our sickest patients. Our experience, consistent with the medical literature, has shown that badgering patients with whatever "preventative" interventions increase hospitalizations and other costs. These programs persist because of a statistical illusion of regression to the mean -- people tend to be enrolled in such programs when their health is at a nadir, then they stabilize therafter. It makes it appear as if the intervention reduced utilization. In fact, a proper comparison shows that it actually increases utilization. Does Google think that spamming millions of people with robo-calls about eating apples will improve anything?

Nothing will ever be attempted if all possible objections must be first overcome. -- Dr. Johnson

Working...