Election day pits pollsters as well as politicians
Posted by mark in politics, Uncategorized on November 2, 2010
Sunday’s St. Paul Pioneer Press reported* an astounding range of predictions for today’s election results for Governor of Minnesota. The Humphrey Institute showed Democrat Dayton leading Republican Emmer by 41 to 29 percent, whereas Survey USA (SUSA) respondents favored Dayton by only 1 percent – 39-38! The SUSA survey included cell-phone-only (CPO) voters for the first time – one of many factors distinguishing it from their competitor for predicting the gubernatorial race.
What I always look for along with such predictions is the margin of error (MOE). The Humphrey Institute pollsters provide these essential statistical details: “751 likely voters living in Minnesota were interviewed by telephone. The margin of error ranges between +/-3.6 percentage points based on the conventional calculation and +/-5.5 percentage points, which is a more cautious estimate that takes into account design effects, in accordance with professional best practices.”** Note that the more conservative MOE (5.5%) still left Dayton with a significant lead, but just barely at 12 points (vs 5.5%x2 = 11% overlap of MOEs).
Survey USA, on the other hand, states their MOE as +/- 4%. They provide a very helpful statistical breakdown by CPO versus landline, gender, age, race, etc. at this web posting. They even include a ‘cross-tab’ on Tea Party Movement – a wild card in this year’s election.
By tomorrow we will see which polls get things right. Also watching results with keen interest will be the consultants who advise politicians on how to bias voters their way. Sunday’s New York Times offered a somewhat cynical report on how these wonks “Nudge the Vote”. For example, political consultant Hal Malchow developed a mailer that listed each recipient’s voting history (whether they bothered to do so, or not), along with their neighborhood (as a whole, I presume). Evidently this created a potent peer pressure that proved to be 10 times more effective in turning non-voters into voters! However, these non-intuitive approaches stem from randomized experiments, which require a control group who get no contacts (Could I volunteer to be in this group?). This creates a conundrum for political activists – they must forego trying to influence these potential voters as the price paid for unbiased results!
“It’s the pollsters that decide. Well, a poll can be skewered [sic #]. I can go out and get you a poll on anything you want and probably get the results that I want just in how I conduct it.”
— Jesse Ventura, professional wrestler (“The Body”) and former governor of Minnesota
# Evidently a Freudian slip – him being skewered on occasion by biased polls. 😉
* “Poll parsing” column by David Brauer, page 15B.
Brain-bending thoughts on a coffee experiment
Posted by mark in design of experiments, Uncategorized, Wellness on October 24, 2010
The Stat-Ease training center here at our world headquarters in Minneapolis features a wonderful single-cup brewing system that you can see demoed here. When we are not holding a workshop, I sometimes sneak in to steal a cup late in the day. By then I am reaching my limit, so I brew a “half-calf” at the half-cup setting. Being a chemical engineer, I calculate that, in this case, half of half makes a whole, that is, coffee with the normal concentration of caffeine. Does that make sense?
Making a tasty and effective cup of coffee is a huge deal for knowledge workers who need to keep their heads in gear from start to finish of every single day. One of our workshop students, a PhD, has been picking my brain about testing coffee blends on her staff of scientists. She proposes to do a mixture design such as I did on varying types of beers (see Mixture Design Brews Up New Beer Cocktail—Black & Blue Moon).
Obviously overall liking on a sensory basis should be first and foremost for such an experiment on coffee – a 5 to 9-point scale works well for this.* However, the tricky part is assessing the impact of coffee for accelerating information processing and general problem-solving, which I hypothesize depends on level of caffeine. I wonder if an online “brain training” service, such as this one developed by neuroscientists at Stanford and UCSF, might provide a valid measure.
The down side of doing a proper test on whether coffee improves cognitive skills will be the necessity of reverting to the base line, that is, every morning getting up and trying to function without the first cup.
“A mathematician is a machine for turning coffee into theorems.”
— Alfréd Rényi
*Turn your volume down (to not hear the advert) and see this primer on sensory evaluation by S-Cool– a UK educational site for teenagers.
Caltech prof puts finger on what turns on consumers
Posted by mark in Consumer behavior on October 20, 2010
The November/December issue of American Scientist provides an intriguing heads-up from Caltech neuroeconomist Antonio Rangel on impulse buying. His group randomly presented items to students in varying ways – by name, picture and with the actual object. Buyers paid 50 percent more for the real goods! However, when the item could be seen under glass but not available to be held, this difference in desirability disappeared. Expect to see even more irresistible items showing up at the checkout counters to entice you into an impulse purchase. : (
“Next time you’re stuffed and the waiter wheels around the dessert cart, know that the odds are against you. Just cross your fingers that the chocolate cakes under a glass dome, to help you resist the urge.”
– Christopher Intagliata, Scientific American reporter (from text of podcast here)
PS. While researching Rangel, I came across this article in the Caltech News on another study he did on buying behavior – in this case the propensity to confuse price with quality for wine. Here you can learn why Rangel is a “neuro” economist – he directly measures brain activity, which provides more reliable measures of consumer response than what they might admit when asked how they feel about something. See how MRI signals change in student brains confronted with money, trinkets versus snacks (can you guess which turns them on the most?) at this web-page detailing research by the Rangel Neuroeconomics Laboratory.
Yankees leverage wins by throwing money at their players
Posted by mark in sports, Uncategorized on September 26, 2010
Today’s New York Times sports section provided this intriguing graphic on “putting a price tag on winning”. Their hometown Yankees stand out as the big spenders by far. It paid off in wins over the last decade – the period studied. However, if you cover up the point depicting the Yanks, the graph becomes far less compelling that salary buys wins – mainly due to counteractive results enjoyed by two low-payroll teams: The Minnesota Twins and the Oakland Athletics.
I found similar patterns and, more importantly, data to reproduce these, in this study of MLB Payroll Efficiency, 2006-2008 by Baseball Analyst Rich Lederer.
No offense to Rich or the NY Times – it is the damn Yankees (sorry but I am weary of them defeating the Twins every post-season) who are the blame for this flaw in drawing conclusions from this data: One point exerts undue leverage on the fit, which you can see on this diagnostic graph generated by Design-Expert® software.
However, after doing the obvious thing – yanking the Yanks from the data, the conclusion remains the same: Higher payroll translates to more wins in Major League baseball. Here are the stats with/without the Yankees:
- R-squared: 0.41/0.34
- Wins per $ million of payroll (slope of linear fit): +0.13/0.16
In this case, a high leverage point does not exert the potential influence, that is, the end result does not change due to its location. If you’d like to simulate how leverage impacts fit, download this educational simulation posted by Hans Lohninger, Associate Professor of Chemometrics at Vienna University of Technology.
Minnesota’s ’08 Senate race dissed by British math master Charles Seife
Posted by mark in Basic stats & math, politics, Uncategorized on September 20, 2010
Sunday’s New York Times provided this review of Proofiness – The Dark Arts of Mathematical Deception – due for publication later this week. The cover, seen here in Amazon, depicts a stats wizard conjuring numbers out of thin air.
What caught my eye in the critique by Steven Strogatz – an applied mathematics professor at Cornell, was the deception caused by “disestimation” (as Proofiness author Seife terms it) of the results from Minnesota’s razor-thin 2008 Senate race, which Al Franken won by a razor-thin 0.0077 percent margin (225 votes out of 1.2 million counted) over Norm Coleman. Disestimation is the act of taking a number too literally, understating or ignoring the uncertainties that surround it; in other words, giving too much weight to a measurement, relative to its inherent error.
“A nice anecdote I like to talk about is a guide at the American Museum of Natural History, who’s pointing at the Tyrannosaurus rex. Someone asks, how old is it, and he says it’s 65 million and 38 years old. Sixty-five million and 38 years old, how do you know that? The guide says, well, when I started at this museum 38 years ago, a scientist told me it was 65 million years old. Therefore, now it’s 65 million and 38. That’s an act of disestimation. The 65 million was a very rough number, and he turned it into a precise number by thinking that the 38 has relevance when in fact the error involved in measuring the dinosaur was plus or minus 100,000 years. The 38 years is nothing.”
– Charles Seife (Source: This transcript of an interview by NPR.)
We Minnesotans would have saved a great deal of money if our election officials had simply tossed a coin to determine the outcome of the Franken-Coleman contest. Unfortunately, disestimation is embedded in our election laws, which are bound and determined to make every single vote count, even though many thousands in a State-wide race prove very difficult to decipher.
Stats reveal real ace of Twins pitching staff
Twins fever is running rampant now in Minnesota. The home baseball club is leading their division and riding a wave of popularity with a new stadium and a home-town hero (Joe Mauer – a fellow graduate of my high school in Saint Paul). Over dinner this week with a colleague and a master statistician in town for a visit the talk turned to the Twins and who now should be considered their ace pitcher – Carl Pavano or Francisco Liriano. Although appreciative of Pavano’s consistently good performance over the entire year, I felt that Liriano has come on stronger in the second half of the season. Having reached a stalemate, the three of us agreed that the data might tell the story.
I found everything I needed to make my case for Liriano at ESPN’s statistics site for Major League Baseball. Here is the pitching “split” for the second half of the MLB season so far for Liriano vs Pavano; respectively:
- Win-Loss: 7-0 vs 6-4
- WHIP (walks and hits per inning pitched): 1.27 vs 1.32
- ERA (earned run average): 2.22 vs 3.41
Pavano is good, but Liriano is my pick as the current ace of the Minnesota Twins pitching staff. Why argue with words? Let the data speak.
Harvard economist advises students of all ages to learn some statistics
Posted by mark in Basic stats & math on September 5, 2010
In this Sunday New York Times “Economic View” column, Harvard professor N. Gregory Mankiw advises that those who wish to pursue this “dismal science” take one or more courses in statistics while in college. He sees a dearth of knowledge on this subject in his first year students.
“High school mathematics curriculums spend too much time on traditional topics like Euclidean geometry and trigonometry. For a typical person, these are useful intellectual exercises but have little applicability to daily life. Students would be better served by learning more about probability and statistics.”
— N. Gregory Mankiw
I’m with him on learning more about stats but not at the expense of less geometry and trig, which come in very handy for anyone pursuing an engineering career. Also, budding economists could benefit from a little knowledge of period functions such as sine waves. It seems to me that what goes around comes around.
Quantifying statements of confidence: Is anything “iron clad”?
Posted by mark in Basic stats & math, Uncategorized on August 19, 2010
Today’s “daily” emailed by The Scientist features a heads-up on “John Snow’s Grand Experiment of 1855” that his pioneering epidemiology on cholera may not be as “iron clad” as originally thought. A commentator questions what “iron clad” means in statistical terms.
It seems to me that someone ought to develop a numerical confidence scale along these lines. For example:
- 100% Certain.
- 99.9% Iron clad.
- 99% Beyond a shadow of a doubt.
- 95% Unequivocal.
- 90% Definitive.
- 80% Clear and convincing evidence.
- 50% On the balance of probabilities.
There are many other words used to convey a level of confidence, such as: clear-cut, definitive, unambiguous, conclusive. How do these differ in degree?
Of course much depends on how is making such a statement, many of whom are not always right, but never in doubt. ; ) I’m skeptical of any assertion, thus I follow the advice of famed statistician W. Edwards Deming:
“In God we trust, all others bring data.”
Statistics can be very helpful for stating any conclusion because it allows one to never have to say you are certain. But are you sure enough to say it’s “iron clad” or what?
Blah, blah, blah…”quadratic”
Posted by mark in design of experiments, Uncategorized on August 15, 2010
This add by Target got my attention. It reminded me of my futile attempt to get my oldest daughter interested in math. For her the last straw was my overly-enthusiastic reaction to her questioning me why anyone would care about quadratic equations. Perhaps I over-reacted and lectured on a bit too long about this being a very useful approximating function for response surface methods, blah, blah, blah…
Opportunistic eating a problem for new college students
A study recently published online by the Journal of Adolescent Health concludes that female (but not male!) students in dorms with dining halls gained significantly more weight than those who were forced to go out to eat. The lead author, Kandice Kapinos (University of Michigan), took advantage of the practice of colleges to randomly assign dorm rooms. Specifically, she and her study team examined the weight gain of 388 freshmen at Marquette University. (See a few more details in this press release.).However, they relied on self-reporting rather than direct observation and measurement, which creates some doubt about the validity of their conclusions. After all, people have been known to fudge about their weight. 😉
Nevertheless, based on observation of three daughters and two sons who went off to college, I believe that Kapinos et al are really on to something. This was sealed in my mind from the observation of my youngest girl, who soon will start her third year in a biochemistry program. She said it really is very simple – the female students hang around the dorm dining hall for social reasons, during which times they naturally munch on stuff and pack on the pounds. I think for the good of their students it would be wise of schools not to put cafeterias in the dorms.
“I think perhaps that women are a little more socially oriented at college, and one social thing they engage in might be, ‘Let’s go get a snack,’ Men may not tend to do that as much.”
— Wayne Westcott, Senior Fitness Executive for the YMCA in Quincy, Massachusetts
By the way, putting on weight after going away to college — the proverbial “freshman 15” – is bound to happen, I think; and for both men and women. I gained 15 pounds after moving into a dorm at Michigan State University. It had a dining hall in the building. However, I’ll bet I’d have added weight just the same even if the meals were served elsewhere on campus.