communication. community. cognition.
Posts tagged Statistics
Headlines, ripped straight from the Razr
Aug 11th
From today’s New York Times:
Twitter has been described many ways. At its best, it has been called a revolutionary political tool and a low-cost marketing machine. At its worst, it has been dubbed a waste of time.
Now, two researchers are calling it a hedonimeter, a device that measures happiness.
Peter Sheridan Dodds and Christopher M. Danforth, a pair of statisticians from the University of Vermont, are hoping to harness the stream of messages flowing through the popular micro-blogging platform at any given moment to read public opinion and sentiment in real-time.
Sounds a bit similar in concept to a piece I wrote two years ago.
Yes, the application was an April Fools’ joke, but the concept isn’t.
Hit and Run Journalism
Apr 10th
Nick Adenhart was a pitcher for the Los Angeles Angels, killed last night in a car crash. It was billed as a hit-and-run accident, but those under the intense deadline pressures of modern media are always groping to find some deeper meaning.
It is a senseless death. Literally. Trying to make sense of random occurrences is an exercise in futility and frustration. At the far end, it is a dangerous exercise. Our minds often aren’t capable of writing off the random as just that, and in placing significance on meaningless correlations, we perpetuate ideas that just are not valuable.
Lisa Fletcher is the reporter who covered the story Friday morning for Good Morning America. She’s the ‘face’ of the piece, and in no way do I hold her entirely accountable. My dander is up over the legion of producers and news managers who pushed a story, and even vetted one that has zero significance. Not Adenhart’s death – but the ridiculous tangent of “hit-and-run” accidents being on the rise.
Here’s a quote from Fletcher’s piece:
“California has the highest number of hit-and-run deaths in the country, but other states are seeing an increase in these accidents, many of which go unsolved.”
Let’s examine that for just a moment.
A “hit-and-run” accident is a crime, and is reported as such. The accident, in and of itself, might be the result of a separate crime. In other words, the driver might be drunk, or might be driving with a suspended license. Those are crimes that may or may not have contributed to the collision.
The “hit-and-run” itself is nothing more than leaving the scene of an accident. So there are “hit-and-runs” where no crime was committed initially, but the driver left the area out of fear or panic or confusion. Which means we have many different kinds of hit-and-run accidents.
- A true accident where a panicked driver fled. (1 crime)
- A true accident where an unlicensed driver fled. (1 minor violation, 1 crime)
- A true accident where a driver who should not be driving (suspended license) also flees the scene. (2 distinct crimes)
- An alcohol or substance-related accident where the driver flees. (2 crimes)
- An alcohol or substance-related accident where the driver had a suspended license and flees. (3 distinct crimes)
You see where this is going. There are simply too many possibilities present to lump all of these together as “hit and run accidents” with no context. It’s a meaningless distinction, because the tragedy is that people are dead, not that a participant in the accident ran away.
In her report, Fletcher followed the statement about California with a couple of anecdotal incidents from the east coast, bearing no statistical context. She then stated that 37,000 Americans every year die in hit-and-run accidents.
Tell me, how do you prevent a hit-and-run? Wouldn’t you rather prevent the accident in the first place?
And why is it that some states “lead” others in hit-and-run accidents? Might it be that California is the most populous? Let’s assume for a moment that the people vetting this story aren’t completely brain-dead, and they factored in “per capita” or “per hours on the road?” What factors might play a role in a high incidence of hit-and-run accidents?
- A state with stiffer penalties for those driving with suspended licenses would provide greater incentive for someone to “make a break for it.”
- A state with a less lenient judicial track record for DUIs would make it more likely that a driver will flee.
- A state with a higher percentage of drunk drivers on the road might be likely to have more hit-and-runs, because of impaired judgment.
- A state with a higher percentage of undocumented illegals would have a higher incidence of leaving the scene.
In the above, we have one factor that actually has something to do with the accident (drunk driving), and three that have to do with behavior linked to wanting to avoid police/judicial contact. And of those, we have two that indicate that in states where enforcement is uniform and there’s zero tolerance for DUIs, there is a greater incentive to run.
That’s right. In states where you know you’ll be punished severely for driving while intoxicated, there is more incentive to run.
This emphasis on “hit-and-run” accidents is one based on emotion, not fact. Stories where bereaved family members have “no justice” and “no one to hold accountable” are quite sexy. It makes the journalistic endeavor seem more noble than exploitative. It brings a community together to solve this crime, and bring the offender to justice.
And it’s just plain wrong, because it stirs up anger and resentment at the wrong thing.
Be angry about the accident. Be angry about the fact someone fled the scene. Don’t confuse the two, because the higher rate of “hit-and-run” accidents might in fact be a sign that the rest of the justice system is working.
Passing Trains
Oct 28th
The following messages came in through my Twitter (and Facebook) stream within a 15-minute span Monday morning:
Laura Howe: wow! metro is really quiet this am. hardly anyone on board for a monday morning.
Rick Murray: Want proof people are driving less: Metra has added a car to this train = 250 more folks doing the green thing downtown.
Two trains, two cities, two divergent opportunities to make an observation. That’s what we do, gather input and try to reorganize it into some semblance of understanding. It’s easier to remember conclusions than the steps it took to get there. It’s easier to remember the final score of a game than a comprehensive list of the plays. The trick is to remember the information that is significant.
Finding the Factor
In geometry, we have theorems – postulates that aren’t true by definition, but have been proven true by experience and derivation from other known truths. We fail when we skip to a new derivation without enough fact to support it. And we do that far too often.
Take Rick’s assertion. Is it indeed a sign of the times? Are there more commuters avoiding high gas prices? Is he basing this on a personal observation or just reading a news release from the operators of Chicago’s Metra?
Now look at Laura’s message. The Metro (DC) is eerily quiet. Does that entail a drop in ridership? Or is there a federal holiday underway, not uncommon for a Monday? Or are more people taking cars?
Truth is, we don’t know. We don’t have enough information. Such a message (or observation) remains strictly anecdotal out of context. Maybe Metra is adding a car in anticipation of behavior that will not materialize. Who knows?
Shortcuts Can Miss the Truth
All too often, we over-apply Occam’s Razor and end up buttressing our previously-held beliefs with new ‘evidence’ that might not apply. Take Laura’s case. How many reasons might there be for such a small number on her train:
- Holiday
- More people are driving cars than before
- Previous train was later than normal, giving a smaller “gap of opportunity” for riders
- Connecting train was late
- Erroneous report that trains were closed
- Laura got on an unusually empty car
Again, who knows? But human nature tells us that when we carry a powerful narrative, we’ll tend to lean in that direction. If you are just convinced that we’re in the worst economy ever, you might even believe no one is on the train because everyone else is at home after a layoff. If you’ve just heard about a flu outbreak, you might blame it on that. Or, if you know there is a parade that day, you might assume people are attending that event instead.
The misapplication of Occam’s Razor occurs when we choose to believe the simplest explanation that matches our preconceived worldview, instead of just the simplest.
And the simplest isn’t always correct – just more likely to be.
Now – for fun – please chime in with your ideas. Come up with your own possible explanations for why Laura’s train might have been unusually empty. Or, share your theories about why Metra is adding a car that might have nothing to do with “going green.”
[tags]Ike Pigott, Occam’s RazR, statistics, occam’s razor, postulates, theory[/tags]
Base Assumptions
Jul 16th
Conventional wisdom can be a good guide, but why elevate it to canon and law? How often do we go with the prevailing thought instead of challenging it?
This came to mind in the 11th inning of Tuesday night’s All-Star game. Baseball is a sport that is steeped in tradition, and it’s hard — even when you have the statistical evidence in front of you — to challenge a base assumption. Here’s the situation:
With two outs in the bottom of the 11th, Carlos Quentin is at the plate. Dioner Navarro is on third base, Michael Young is on second. Conventional wisdom tells us that with two outs, the runners should always be running. Sure enough, Young charged forth the moment the pitch was delivered — and as the ball bounced to the shortstop, Young almost waltzed right into a tag for the third out! I’m here to say that conventional baseball wisdom is stupid, and here’s why:
- Young was no longer a factor in the game.
- With two outs in extra innings, either Navarro scores or he doesn’t.
- Young cannot be forced out at third base.
Michael Young should have parked his butt on second, and started drawing stick figures in the dirt just like my son did on his tee-ball adventures! The only thing that could have happened that would have mattered is if he had been tagged out! But hey – after 30 years of being yelled at to run with two outs, who can blame a guy. Unless that guy is making $6,174,974 this season alone! Think for yourself, Michael! Don’t cost your team a game by blindly following rote!
Sports Smarts
Baseball used to be filled with examples of managerial decisions that were based on hunch instead of science. That’s not as true anymore, as the geeks and nerds have compiled massive quantities of data that indicate a different strategy than gut instinct offer. Earl Weaver, the late manager of the Baltimore Orioles was obsessive about his charts, and would choose his lineups and pinch hitters based on leveraging the smallest of advantages. Then he’d ditch the charts entirely and follow the old safe rules, because no one faults you for playing it safe.
Now a whole generation of stats-minded people are running major league teams. I highly recommend the book Moneyball, which looks at how the Oakland A’s used statistics and smarts to outwin and underspend everyone else. You don’t have to be a sports fan to find it fascinating. There are many examples of how the value of stolen bases is over-inflated, how defense is under appreciated, and how the scouts really have no idea what they are looking for!
Extra Points
Baseball isn’t the only sport that confuses tradition with reality. Gregg Easterbrook has taken football coaches to task for clinging to the punt when the numbers clearly indicate that punting is a losing bargain. It’s a thoroughly-researched piece of athlethic academia, as is his ongoing explanation of how teams kill themselves by blitzing too often. The former is a case of coaches playing it safe, and avoiding criticism in case they don’t convert. The latter is an issue of playing to the fans, or placing too much credence on “momentum” instead of just following your averages.
You’ll find a lot of this sort of non-analytical thinking in places other than dugouts and sidelines. It happens anywhere there is a management structure that is “inbred.” Baseball managers are former players, and as such learn the subtlties of the game from other managers. Football coaches typically come from within the offensive line, the position that rates the highest on intelligence tests. Offensive linemen are certainly in a position to appreciate the battles in the trenches, but it’s hard for innovation to take root when everyone in the coaching fraternity pledged the same script.
Game Theory
In a truly free market, innovation is rewarded with success. Adam Smith’s so-called “invisible hand” was merely the recognition that when a bunch of people act in their own self interests, they optimize their activity to follow the rules in place. In economic terms, this means higher efficiency, lower prices, and better service. Those who can’t compete lose out. Monopolies are dangerous to free markets because if there is no reward for innovation, there is no investment in innovation. That’s why they are either regulated or disintegrated.
However, what about those occasions where the game (and those playing) has stabilized at a level that is NOT really the optimum? We’ve highlighted several examples in sports, but what about other industries and professions? Do we not see newspapers and broadcasters clinging to their business models even though the game has changed around them? Are we not living in a political age where it’s now entirely imaginable that a candidate can raise a previously unimaginable sum? Not to mention the generational flip-flops that occur with regards to health and nutrition.
Hold ‘Em
How easy would it be to win at Poker if you had the luxury of looking at your own cards and your opponents’? Information is king, and traditionally those who had access to the information had the advantage. Much of the speculation markets are built on the concept (or rather the wager) that one person knows something the other doesn’t. In this present age, the prevalence of information and the ability to quickly find it has shaved away most of that imbalance. Someone on the other side of the world has access to much of the same data you do — and maybe more. (And being on the other side of the internet is the perfect poker face – you never see the cards either.)
This isn’t a disintegration of markets – it is just a more rapid approach to equilibrium. (Equilibrium is never truly reached, but we get closer much faster.) Want to know exactly how much a Wii Fit is worth? Click on eBay and look at the sales about to close out. Want to know how much your life is worth? Compare insurance quotes, instantly. The closer we get to real-time information in a universal format, the more the rough edges disappear.
Swing for the Fences
The above would be sobering news, but human nature always affords us opportunities to find an imbalance to create a new market. The trick is to not get sucked down the rabbit hole, chasing a never-ending stream of decimal points in search of an advantage. There are still many instances where the collective wisdom of a particular group has calcified and become stagnant. They compete with each other using a set of assumptions and measurements that no longer reflect reality, if they ever truly did. Don’t quibble about inches when there are yards to be gained. The key is to question the assumptions that lock everyone else into patterns of behavior. Challenge enough of those, and you’ll find an answer no one 
else would have expected.
There are more than enough base assumptions waiting to be called out, if you only have the guts to sit your butt on second.
Postscript: Michael Young, who nearly made a foolish out in the 11th inning, won the game with an out in the 15th. He hit a sacrifice fly to bring home the winning run.
[tags]Ike Pigott, Occam’s RazR, baseball, football, economics, statistics, moneyball, TMQ, Gregg Easterbrook[/tags]
The Java Economy
Jul 2nd
“Imitation is the sincerest form of television” – Fred Allen
Apparently, the producers at Good Morning America have taken a page from The Daily Show. Last night, Comedy Central re-aired the show with Larry Wilmore’s insightful and informed piece about how a down economy is destroying Beverly Hills; people are only buying new Bentleys every other year, instead of annually, and plastic surgeons are being forced to do actual reconstructive work!
On Wednesday’s GMA, the producers sent Bianna Golodryga (whose name has appeared here often enough that I don’t have to Google for spelling anymore) to intro a piece about the economy in front of a Starbucks. Paraphrased, it’s getting so bad that people aren’t buying coffee, and Starbucks is closing stores!
Let me understand… people are no longer choosing to spend $4 for a cup of coffee? This is proof of a recession? Yes, Starbucks is closing 600 stores, which is 8-percent of the total nationwide. But let me ask a couple of common-sense questions:
- Didn’t we hear about all of those poor mom-n-pop java joints that were suffering because the evil behemoth from the great Northwest was invading?
Actually, economic studies proved that independent coffeehouses did better when a Starbucks came into the neighborhood. It drove up the overall market for premium coffee consumption, and everyone won. So is this Starbucks retreat a sign that people aren’t buying coffee anymore? Or that they’re buying it from somewhere other than a Starbucks? - How much of the Starbucks retreat is really about growing too fast? How many of the 600 stores were within a half-mile of another location? A quarter-mile? Across the street?
We’ll never know. Starbucks is very good at generating attention for itself. It is an iconic market leader, but too many people will walk away from this GMA story confusing the market-leader with the overall market. Much in the same way that US “domestic auto production” is still calculated as though Nissan, Toyota, Mercedes, Hyundai, and Honda don’t have manufacturing facilities in the United States.
I’m not digging on GMA, I just happen to to watch. But it’s clear that the standards that once reigned over network news continue to drop. Truth is a function of both fact and context – and modern television journalism provides too much of the former without the latter.
Update: the analysts apparently agree with me:
Starbucks said the 600 stores are either unprofitable now or are not expected to meet future return thresholds. All of the targeted units are close to another company-operated store, [Chief Executive Peter] Bocian said.
“This is validating some of the critics who said they were opening stores too close to one another,” said James Walsh, an analyst at Starbucks investor Coldstream Capital Management.
Those critics complained that the company had overbuilt in the United States — particularly in major urban areas like Manhattan, where it is not unusual to see several Starbucks in a single city block.
[tags]Ike Pigott, Occam’s RazR, ABC, GMA, Good Morning America, Bianna Golodryga, Starbucks, journalism, economy, The Daily Show, Larry Wilmore[/tags]
Fuming Mad
Apr 25th
Gas prices are up. Yes. But manufacturing a story about an increase in gasoline theft has me fuming at my former profession. This morning, it just happens to be focused on Good Morning America’s Bianna Golodryga and her producers. Lazy, lazy, lazy all the way around.
I’m fishing for a transcript right now, but the gist of the story is:
- Gas prices are up.
- Some people leave without paying
- Some people are siphoning fuel.
- You should get a locking gas cap, $15.
What is spurring this sudden trend? Why, the media! “Gasoline theft” is poised to be the Shark Attack story of the summer. A story that carries great anecdotal weight, but is statistically numb.
Nowhere in the GMA piece did they cite statistics. What they did manage to find was a tearful grandmother who was interviewed through a crack in her front door, who was caught driving off without paying. Nine times, in the last couple of months.
Running On Empty
Yes, that was a powerful and emotional interview — all four seconds of it that I saw. She really seemed remorseful. Although I’m willing to bet she’s even more remorseful now that the “crack-in-the-door” interview she did with her local affiliate got kicked up to run across the nation. Television storytelling relies on emotion and visual impact. You can’t really tell a story like this one without that personal touch — but this piece is running on empty, empty journalistic calories. The “personal story” should be something that provides a slice of life, a perspective that reinforces the trend. I can tell you that 3,000 people evacuated a neighborhood, and then concentrate on how it affected a couple of those people in human ways. That’s powerful, and a good use of the medium. Giving you the human part with no statistical context is irresponsible.
This story on GMA was driven by a story that ran in a local market somewhere… and that story only made air because of that interview with the grandmother. We have no proof that gas drive-offs are more prevalent, but a broadcast journalist needs a hook (or “news peg”) to hang that story on.
Carjacking the Agenda
In searching for some evidence online, I found a couple of interesting links. This from Reuters:
NEW YORK (Reuters) – U.S. motorists, angered by soaring gasoline prices, are resorting increasingly to theft — a trend that could worsen heading into summer driving season, a national association of fuel retailers said Thursday.
“It is getting bad. When the price of gasoline goes up, the number of drive-offs goes up,” said Dan Gilligan, president of the Petroleum Marketers Association of America, which represents about 8,000 retailers.
So, the Petroleum Marketers Association is weighing in, providing us with the perspective that as prices rise, so does attempted thefts. Why do you think the PMAA might be making itself available for interviews at this particular time?
Gilligan said that some state fuel dealer associations were pressing lawmakers to make it easier to prosecute motorists who fill up and then drive off without paying, while many service stations were starting to require payment up front.
Oh, so they are lobbying for legislation! Read the article, and tell me if you can find any evidence that gas drive-offs are up.
Low-Octane Truth
So where do we get the notion that gas theft is up? I found another resource, this time a fact sheet from the NACS, the Association for Convenience and Petroleum Retailing. (Yes, I realize that it spells A-C-P-R instead of N-A-C-S, but NACS was originally founded as the National Association of Convenience Stores.) According to the NACS:
Nationwide, in 2007, gasoline theft cost the industry $134 million, a sharp decline from the $300 million reported in 2005 and the $237 million reported in 2004. (Theft totalled $122 million in 2006.)
We don’t have any data for 2008, but we’re supposed to extrapolate from this data that we’re going to have a huge year for theft? Even though 2007 was way down from 2004 and 2005? Bear in mind that the dollar-figure for theft is also affected by the price. So a pre-Katrina total of $237,000,000 in 2004 took place when regular unleaded was less than $2.00/gallon. (Great interactive chart here.)
The NACS does provide some context as to why thefts might be down, even if the temptation to steal might be up:
The problem of gasoline theft would have been even greater since September 2005 if so many retailers hadn’t begun to mandate prepay in after Hurricane Katrina when gasoline prices reached record levels of $3.06 per gallon, and when gasoline prices again topped $3 per gallon in every year since then.
So thefts are down because gas stations now won’t even turn on the pump for you until you swipe a card or go inside the store to prepay. And they’ve been doing it this way for more than two-and-a-half years.
Sucking on Fumes
ABC News, like many many other outlets that I did not watch nor will see, is guilty of lazy journalism. There is no news value in the story I saw today, other than the vague notion that I might be able to find YouTube videos of people showing me how to illegally siphon gas out of someone else’s tank. Gas thefts have been drastically down for two years because of retail policy changes, even with higher prices. We don’t have any real peg here, other than from a lobbying group that wants tougher state penalties, and stands to gain from a general public perception that this is indeed a growing problem.
In retrospect, I have no clue how long ago grandma cried to the camera. My gut (and inside knowledge of how the sausage of news is processed) tells me that the Grandma interview has been sitting on the shelf for a couple of weeks at least. It’s not even “new.” (Again, just my supposition.) My gut tells me that some hotshot producer has had that tape sitting on his desk for a couple of weeks, just waiting for enough of a gas-related news-peg to justify dusting it off and putting it on national television. So much of modern news is driven from press releases, and the truth suffers along with us.
The only surprise is that I was not able to readily find a news release from a manufacturer of locking gas caps.
[tags]Ike Pigott, Occam’s RazR, news, journalism, gasoline, crime, ABC News, Bianna Golodryga, marketing, PR[/tags]
