andrew_gelman_stats andrew_gelman_stats-2010 andrew_gelman_stats-2010-395 knowledge-graph by maker-knowledge-mining
Source: html
Introduction: I’m a physicist by training, statistical data analyst by trade. Although some of my work is pretty standard statistical analysis, more often I work somewhere in a gray area that includes physics, engineering, and statistics. I have very little formal statistics training but I do study in an academic-like way to learn techniques from the literature when I need to. I do some things well but there are big gaps in my stats knowledge compared to anyone who has gone to grad school in statistics. On the other hand, there are big gaps in most statisticians’ physics and engineering knowledge compared to anyone who has gone to grad school in physics. Generally my breadth and depth of knowledge is about right for the kind of work that I do, I think. But last week I was offered a consulting job that might be better done by someone with more conventional stats knowledge than I have. The job involves gene expression in different types of tumors, so it’s “biostatistics” by definition, but the
sentIndex sentText sentNum sentScore
1 Although some of my work is pretty standard statistical analysis, more often I work somewhere in a gray area that includes physics, engineering, and statistics. [sent-2, score-0.182]
2 I have very little formal statistics training but I do study in an academic-like way to learn techniques from the literature when I need to. [sent-3, score-0.21]
3 I do some things well but there are big gaps in my stats knowledge compared to anyone who has gone to grad school in statistics. [sent-4, score-1.009]
4 On the other hand, there are big gaps in most statisticians’ physics and engineering knowledge compared to anyone who has gone to grad school in physics. [sent-5, score-1.058]
5 Generally my breadth and depth of knowledge is about right for the kind of work that I do, I think. [sent-6, score-0.447]
6 But last week I was offered a consulting job that might be better done by someone with more conventional stats knowledge than I have. [sent-7, score-0.836]
7 The job involves gene expression in different types of tumors, so it’s “biostatistics” by definition, but the specific questions of interest aren’t specialized biostats ones (there’s no analysis of microarray data, for instance). [sent-8, score-0.354]
8 I’ve only ever done five small consulting jobs, and I’ve charged as little as $80/hour (working for some ecologists who didn’t have any money) and as much as $250/hour (consortium of insurance companies). [sent-12, score-0.582]
9 Picking a number out of the air, I’m charging $150/hour. [sent-13, score-0.164]
10 Of course one way to think of it is: would I rather have spent three hours last night working on this project for $450, or would I have preferred doing whatever else I would have done instead but not making any money? [sent-15, score-0.542]
11 (My wife is out of town and I hadn’t made plans, so I probably just would have read or watched TV). [sent-16, score-0.236]
12 By that standard I am charging a fair rate, I was happy enough working on this last night. [sent-17, score-0.35]
13 But I also have to put in some time this weekend, when I might feel differently: I’ll probably be giving up something more enjoyable this weekend. [sent-18, score-0.092]
14 On the other hand, I think that from the company’s perspective, at least in this particular instance, they are getting a fantastic deal. [sent-20, score-0.185]
15 Having spoken with the people they’ve had looking at the data up to now, I am definitely much better at this than they are! [sent-21, score-0.082]
16 But on the other hand, some universities offer stats consulting for $80-$100/hour, although this is surely not the free-market rate. [sent-24, score-0.468]
17 For the future it would be good to have a better idea of how to set a rate. [sent-25, score-0.088]
wordName wordTfidf (topN-words)
[('company', 0.213), ('consulting', 0.21), ('fantastic', 0.185), ('stats', 0.183), ('knowledge', 0.178), ('money', 0.169), ('gaps', 0.164), ('charging', 0.164), ('phone', 0.129), ('engineering', 0.127), ('grad', 0.127), ('instance', 0.119), ('training', 0.117), ('quickly', 0.114), ('hand', 0.112), ('gone', 0.11), ('hours', 0.107), ('biostats', 0.106), ('physics', 0.105), ('consortium', 0.1), ('done', 0.096), ('tumors', 0.096), ('breadth', 0.096), ('ecologists', 0.096), ('eh', 0.096), ('working', 0.095), ('little', 0.093), ('enjoyable', 0.092), ('last', 0.091), ('work', 0.091), ('specialized', 0.09), ('future', 0.088), ('charged', 0.087), ('compared', 0.083), ('depth', 0.082), ('spoken', 0.082), ('school', 0.082), ('anyone', 0.082), ('weekend', 0.08), ('gene', 0.08), ('watched', 0.08), ('wife', 0.079), ('answered', 0.078), ('job', 0.078), ('else', 0.077), ('town', 0.077), ('whatever', 0.076), ('reflection', 0.076), ('although', 0.075), ('biostatistics', 0.075)]
simIndex simValue blogId blogTitle
same-blog 1 0.9999997 395 andrew gelman stats-2010-11-05-Consulting: how do you figure out what to charge?
Introduction: I’m a physicist by training, statistical data analyst by trade. Although some of my work is pretty standard statistical analysis, more often I work somewhere in a gray area that includes physics, engineering, and statistics. I have very little formal statistics training but I do study in an academic-like way to learn techniques from the literature when I need to. I do some things well but there are big gaps in my stats knowledge compared to anyone who has gone to grad school in statistics. On the other hand, there are big gaps in most statisticians’ physics and engineering knowledge compared to anyone who has gone to grad school in physics. Generally my breadth and depth of knowledge is about right for the kind of work that I do, I think. But last week I was offered a consulting job that might be better done by someone with more conventional stats knowledge than I have. The job involves gene expression in different types of tumors, so it’s “biostatistics” by definition, but the
2 0.1969218 1597 andrew gelman stats-2012-11-29-What is expected of a consultant
Introduction: Robin Hanson writes on paid expert consulting (of the sort that I do sometime, and is common among economists and statisticians). Hanson agrees with Keith Yost, who says: Fellow consultants and associates . . . [said] fifty percent of the job is nodding your head at whatever’s being said, thirty percent of it is just sort of looking good, and the other twenty percent is raising an objection but then if you meet resistance, then dropping it. On the other side is Steven Levitt, who Hanson quotes as saying: My own experience has been that even though I know nothing about an industry, if you give me a week, and you get a bunch of really smart people to explain the industry to me, and to tell me what they do, a lot of times what I’ve learned in economics, what I’ve learned in other places can actually be really helpful in changing the way that they see the world. Perhaps unsurprisingly given my Bayesian attitudes and my preference for continuity , I’m inclined to split the d
3 0.18637763 1596 andrew gelman stats-2012-11-29-More consulting experiences, this time in computational linguistics
Introduction: Bob wrote this long comment that I think is worth posting: I [Bob] have done a fair bit of consulting for my small natural language processing company over the past ten years. Like statistics, natural language processing is something may companies think they want, but have no idea how to do themselves. We almost always handed out “free” consulting. Usually on the phone to people who called us out of the blue. Our blog and tutorials Google ranking was pretty much our only approach to marketing other than occassionally going to business-oriented conferences. Our goal was to sell software licenses (because consulting doesn’t scale nor does it provide continuing royalty income), but since so few people knew how to use toolkits like ours, we had to help them along the way. We even provided “free” consulting with our startup license package. We were brutally honest with customers, both about our goals and their goals. Their goals were often incompatible with ours (use company X’
4 0.1668328 835 andrew gelman stats-2011-08-02-“The sky is the limit” isn’t such a good thing
Introduction: A few months ago, in response to some reflections from Steve Hsu, I wrote the following: I’ve long been glad that when I went to college in the 80s, I was at MIT rather than Harvard. Either place I would’ve taken hard classes and learned a lot, but one advantage of MIT was that we had no sense–no sense at all–that we could make big bucks. We had no sense of making moderately big bucks as lawyers, no sense of making big bucks working on Wall Street, and no sense of making really big bucks by starting a business. I mean, sure, we knew about lawyers (but we didn’t know that a lawyer with technical skills would be a killer combination), we knew about Wall Street (but we had no idea what they did, other than shout pork belly prices across a big room), and we knew about tech startups (but we had no idea that they were anything to us beyond a source of jobs for engineers). What we were all looking for was a good solid job with cool benefits (like those companies in California that had gy
5 0.13314332 578 andrew gelman stats-2011-02-17-Credentialism, elite employment, and career aspirations
Introduction: Steve Hsu has posted a series of reflections here , here , and here on the dominance of graduates of HYPS (Harvard, Yale, Princeton, and Stanford (in that order, I believe)) in various Master-of-the-Universe-type jobs at “elite law firms, consultancies, and I-banks, hedge/venture funds, startups, and technology companies.” Hsu writes: In the real world, people believe in folk notions of brainpower or IQ. (“Quick on the uptake”, “Picks things up really fast”, “A sponge” …) They count on elite educational institutions to do their g-filtering for them. . . . Most top firms only recruit at a few schools. A kid from a non-elite UG school has very little chance of finding a job at one of these places unless they first go to grad school at, e.g., HBS, HLS, or get a PhD from a top place. (By top place I don’t mean “gee US News says Ohio State’s Aero E program is top 5!” — I mean, e.g., a math PhD from Berkeley or a PhD in computer science from MIT — the traditional top dogs in academ
6 0.13282938 2084 andrew gelman stats-2013-11-01-Doing Data Science: What’s it all about?
7 0.12187258 390 andrew gelman stats-2010-11-02-Fragment of statistical autobiography
8 0.11852171 1611 andrew gelman stats-2012-12-07-Feedback on my Bayesian Data Analysis class at Columbia
9 0.11549133 706 andrew gelman stats-2011-05-11-The happiness gene: My bottom line (for now)
10 0.11315348 1832 andrew gelman stats-2013-04-29-The blogroll
11 0.11079161 1003 andrew gelman stats-2011-11-11-$
12 0.10968453 53 andrew gelman stats-2010-05-26-Tumors, on the left, or on the right?
13 0.10837598 1605 andrew gelman stats-2012-12-04-Write This Book
15 0.10419194 1658 andrew gelman stats-2013-01-07-Free advice from an academic writing coach!
16 0.10387537 231 andrew gelman stats-2010-08-24-Yet another Bayesian job opportunity
17 0.10336955 1207 andrew gelman stats-2012-03-10-A quick suggestion
18 0.10331348 970 andrew gelman stats-2011-10-24-Bell Labs
19 0.1015863 2137 andrew gelman stats-2013-12-17-Replication backlash
20 0.10099117 740 andrew gelman stats-2011-06-01-The “cushy life” of a University of Illinois sociology professor
topicId topicWeight
[(0, 0.234), (1, -0.087), (2, -0.042), (3, 0.019), (4, 0.042), (5, 0.05), (6, 0.061), (7, 0.044), (8, 0.006), (9, 0.025), (10, -0.045), (11, -0.029), (12, 0.026), (13, -0.02), (14, -0.053), (15, -0.02), (16, 0.05), (17, -0.029), (18, 0.026), (19, 0.052), (20, 0.022), (21, -0.012), (22, 0.002), (23, 0.054), (24, -0.038), (25, 0.013), (26, -0.021), (27, -0.029), (28, -0.034), (29, 0.052), (30, 0.026), (31, -0.004), (32, 0.009), (33, 0.006), (34, 0.029), (35, 0.002), (36, 0.007), (37, 0.065), (38, -0.064), (39, 0.0), (40, 0.022), (41, -0.023), (42, -0.044), (43, 0.091), (44, -0.03), (45, 0.001), (46, -0.041), (47, 0.011), (48, -0.011), (49, -0.018)]
simIndex simValue blogId blogTitle
same-blog 1 0.97094011 395 andrew gelman stats-2010-11-05-Consulting: how do you figure out what to charge?
Introduction: I’m a physicist by training, statistical data analyst by trade. Although some of my work is pretty standard statistical analysis, more often I work somewhere in a gray area that includes physics, engineering, and statistics. I have very little formal statistics training but I do study in an academic-like way to learn techniques from the literature when I need to. I do some things well but there are big gaps in my stats knowledge compared to anyone who has gone to grad school in statistics. On the other hand, there are big gaps in most statisticians’ physics and engineering knowledge compared to anyone who has gone to grad school in physics. Generally my breadth and depth of knowledge is about right for the kind of work that I do, I think. But last week I was offered a consulting job that might be better done by someone with more conventional stats knowledge than I have. The job involves gene expression in different types of tumors, so it’s “biostatistics” by definition, but the
2 0.89399266 1596 andrew gelman stats-2012-11-29-More consulting experiences, this time in computational linguistics
Introduction: Bob wrote this long comment that I think is worth posting: I [Bob] have done a fair bit of consulting for my small natural language processing company over the past ten years. Like statistics, natural language processing is something may companies think they want, but have no idea how to do themselves. We almost always handed out “free” consulting. Usually on the phone to people who called us out of the blue. Our blog and tutorials Google ranking was pretty much our only approach to marketing other than occassionally going to business-oriented conferences. Our goal was to sell software licenses (because consulting doesn’t scale nor does it provide continuing royalty income), but since so few people knew how to use toolkits like ours, we had to help them along the way. We even provided “free” consulting with our startup license package. We were brutally honest with customers, both about our goals and their goals. Their goals were often incompatible with ours (use company X’
Introduction: Ever since I got this new sound system for my bike, I’ve been listening to a lot of podcasts. This American Life is really good. I know, I know, everybody knows that, but it’s true. The only segments I don’t like are the ones that are too “writerly,” when they read a short story aloud. They don’t work for me. Most of the time, though, the show is as great as everyone says it is. Anyway, the other day I listened to program #466: Blackjack . It started with some items on card counting. That stuff is always fun. Then they get to the longer story, which is all about a moderately rich housewife from Iowa who, over a roughly ten-year period, lost her life savings, something like a million dollars, at Harrah’s casinos. Did you know they had casinos in Iowa and Indiana? I didn’t. Anyway, the lady was a gambling addict. That part’s pretty clear. You don’t lose your life savings at a casino by accident. The scary part, though, was how the casino company craftily enabled her to
4 0.86256671 835 andrew gelman stats-2011-08-02-“The sky is the limit” isn’t such a good thing
Introduction: A few months ago, in response to some reflections from Steve Hsu, I wrote the following: I’ve long been glad that when I went to college in the 80s, I was at MIT rather than Harvard. Either place I would’ve taken hard classes and learned a lot, but one advantage of MIT was that we had no sense–no sense at all–that we could make big bucks. We had no sense of making moderately big bucks as lawyers, no sense of making big bucks working on Wall Street, and no sense of making really big bucks by starting a business. I mean, sure, we knew about lawyers (but we didn’t know that a lawyer with technical skills would be a killer combination), we knew about Wall Street (but we had no idea what they did, other than shout pork belly prices across a big room), and we knew about tech startups (but we had no idea that they were anything to us beyond a source of jobs for engineers). What we were all looking for was a good solid job with cool benefits (like those companies in California that had gy
5 0.84243226 1003 andrew gelman stats-2011-11-11-$
Introduction: Felix Salmon relates the story of an economics Nobel Prize winner getting paid by a hedge fund. It would all seems pretty silly—sort of like Coca-Cola featuring Michael Jordan in their ads—except that hedge funds are disreputable nowadays and so it seems vaguely sleazy for a scholar to trade on his academic reputation to make free money in this way. It falls roughly in the same category as that notorious b-school prof in Inside Job who got $125K for writing a b.s. report about the financial stability of Iceland—and then, when they came back to him later and asked how he could’ve written it, he basically said: Hey, I don’t know anything about Iceland, I was just taking their money! That said, if a hedge fund offered me $125K to sit on their board, I’d probably take it! It’s hard to turn down free money. Or maybe not, I don’t really know. So far, when companies have paid me $, it’s been to do something for them, to consult or give a short course. I’d like to think that if
6 0.84187371 578 andrew gelman stats-2011-02-17-Credentialism, elite employment, and career aspirations
7 0.83921152 1597 andrew gelman stats-2012-11-29-What is expected of a consultant
8 0.8326838 968 andrew gelman stats-2011-10-21-Could I use a statistics coach?
9 0.82950681 592 andrew gelman stats-2011-02-26-“Do you need ideal conditions to do great work?”
10 0.82841939 2080 andrew gelman stats-2013-10-28-Writing for free
11 0.82714164 988 andrew gelman stats-2011-11-02-Roads, traffic, and the importance in decision analysis of carefully examining your goals
12 0.8236779 1536 andrew gelman stats-2012-10-16-Using economics to reduce bike theft
13 0.82276726 732 andrew gelman stats-2011-05-26-What Do We Learn from Narrow Randomized Studies?
14 0.81051463 1707 andrew gelman stats-2013-02-05-Glenn Hubbard and I were on opposite sides of a court case and I didn’t even know it!
15 0.80990589 2053 andrew gelman stats-2013-10-06-Ideas that spread fast and slow
16 0.80597913 668 andrew gelman stats-2011-04-19-The free cup and the extra dollar: A speculation in philosophy
17 0.80515516 970 andrew gelman stats-2011-10-24-Bell Labs
18 0.80049819 937 andrew gelman stats-2011-10-02-That advice not to work so hard
19 0.79948521 582 andrew gelman stats-2011-02-20-Statisticians vs. everybody else
20 0.79851347 525 andrew gelman stats-2011-01-19-Thiel update
topicId topicWeight
[(3, 0.014), (15, 0.067), (16, 0.047), (21, 0.049), (24, 0.145), (36, 0.015), (45, 0.022), (58, 0.013), (62, 0.044), (66, 0.016), (69, 0.01), (86, 0.015), (89, 0.022), (95, 0.022), (97, 0.016), (99, 0.368)]
simIndex simValue blogId blogTitle
1 0.98681599 1139 andrew gelman stats-2012-01-26-Suggested resolution of the Bem paradox
Introduction: There has been an increasing discussion about the proliferation of flawed research in psychology and medicine, with some landmark events being John Ioannides’s article , “Why most published research findings are false” (according to Google Scholar, cited 973 times since its appearance in 2005), the scandals of Marc Hauser and Diederik Stapel, two leading psychology professors who resigned after disclosures of scientific misconduct, and Daryl Bem’s dubious recent paper on ESP, published to much fanfare in Journal of Personality and Social Psychology, one of the top journals in the field. Alongside all this are the plagiarism scandals, which are uninteresting from a scientific context but are relevant in that, in many cases, neither the institutions housing the plagiarists nor the editors and publishers of the plagiarized material seem to care. Perhaps these universities and publishers are more worried about bad publicity (and maybe lawsuits, given that many of the plagiarism cas
same-blog 2 0.98591721 395 andrew gelman stats-2010-11-05-Consulting: how do you figure out what to charge?
Introduction: I’m a physicist by training, statistical data analyst by trade. Although some of my work is pretty standard statistical analysis, more often I work somewhere in a gray area that includes physics, engineering, and statistics. I have very little formal statistics training but I do study in an academic-like way to learn techniques from the literature when I need to. I do some things well but there are big gaps in my stats knowledge compared to anyone who has gone to grad school in statistics. On the other hand, there are big gaps in most statisticians’ physics and engineering knowledge compared to anyone who has gone to grad school in physics. Generally my breadth and depth of knowledge is about right for the kind of work that I do, I think. But last week I was offered a consulting job that might be better done by someone with more conventional stats knowledge than I have. The job involves gene expression in different types of tumors, so it’s “biostatistics” by definition, but the
3 0.98208219 1848 andrew gelman stats-2013-05-09-A tale of two discussion papers
Introduction: Over the years I’ve written a dozen or so journal articles that have appeared with discussions, and I’ve participated in many published discussions of others’ articles as well. I get a lot out of these article-discussion-rejoinder packages, in all three of my roles as reader, writer, and discussant. Part 1: The story of an unsuccessful discussion The first time I had a discussion article was the result of an unfortunate circumstance. I had a research idea that resulted in an article with Don Rubin on monitoring the mixing of Markov chain simulations. I new the idea was great, but back then we worked pretty slowly so it was awhile before we had a final version to submit to a journal. (In retrospect I wish I’d just submitted the draft version as it was.) In the meantime I presented the paper at a conference. Our idea was very well received (I had a sheet of paper so people could write their names and addresses to get preprints, and we got either 50 or 150 (I can’t remembe
Introduction: This material should be familiar to many of you but could be helpful to newcomers. Pearl writes: ALL causal conclusions in nonexperimental settings must be based on untested, judgmental assumptions that investigators are prepared to defend on scientific grounds. . . . To understand what the world should be like for a given procedure to work is of no lesser scientific value than seeking evidence for how the world works . . . Assumptions are self-destructive in their honesty. The more explicit the assumption, the more criticism it invites . . . causal diagrams invite the harshest criticism because they make assumptions more explicit and more transparent than other representation schemes. As regular readers know (for example, search this blog for “Pearl”), I have not got much out of the causal-diagrams approach myself, but in general I think that when there are multiple, mathematically equivalent methods of getting the same answer, we tend to go with the framework we are used
5 0.98139441 1395 andrew gelman stats-2012-06-27-Cross-validation (What is it good for?)
Introduction: I think cross-validation is a good way to estimate a model’s forecasting error but I don’t think it’s always such a great tool for comparing models. I mean, sure, if the differences are dramatic, ok. But you can easily have a few candidate models, and one model makes a lot more sense than the others (even from a purely predictive sense, I’m not talking about causality here). The difference between the model doesn’t show up in a xval measure of total error but in the patterns of the predictions. For a simple example, imagine using a linear model with positive slope to model a function that is constrained to be increasing. If the constraint isn’t in the model, the predicted/imputed series will sometimes be nonmonotonic. The effect on the prediction error can be so tiny as to be undetectable (or it might even increase avg prediction error to include the constraint); nonetheless, the predictions will be clearly nonsensical. That’s an extreme example but I think the general point h
6 0.97982645 506 andrew gelman stats-2011-01-06-That silly ESP paper and some silliness in a rebuttal as well
7 0.97975463 2287 andrew gelman stats-2014-04-09-Advice: positive-sum, zero-sum, or negative-sum
8 0.97924751 902 andrew gelman stats-2011-09-12-The importance of style in academic writing
9 0.97921336 2115 andrew gelman stats-2013-11-27-Three unblinded mice
10 0.97918224 1683 andrew gelman stats-2013-01-19-“Confirmation, on the other hand, is not sexy”
11 0.97871125 1832 andrew gelman stats-2013-04-29-The blogroll
12 0.97849172 2006 andrew gelman stats-2013-09-03-Evaluating evidence from published research
17 0.97753292 2254 andrew gelman stats-2014-03-18-Those wacky anti-Bayesians used to be intimidating, but now they’re just pathetic
18 0.97740638 785 andrew gelman stats-2011-07-02-Experimental reasoning in social science
20 0.97738326 2236 andrew gelman stats-2014-03-07-Selection bias in the reporting of shaky research