Using and Misusing Statistics
Previous post: War on Humanity
Next post: Flagging Unconstitutional Protectionism
where orders emerge
EconTalk Podcast
Recent Episodes-
where orders emerge
Previous post: War on Humanity
Next post: Flagging Unconstitutional Protectionism
Recent Episodes
Podcast RSS Feed
Full EconTalk Text
{ 21 comments }
Whenever someone's faced with statistics they don't like, they usually tell me that "you can use statistics to argue anything." I usually reply by telling them, "yes, and you can use words to argue anything, too." The last refuge of someone who can't understand a fact is always to dismiss it as useless and unreliable.
Using statistics in a conversation to prove your point puts your opponent at a distinct disadvantage because you force them to hold those statistics as being true when they may not be privy to the data or your specific interpretation of the data.
This is probably why you get the response you do Adam. Your opponent is often in an adversarial position and you are asking them to hold things as truths which is much harder to overcome in an adversarial position.
If you want to see examples of the abuse of statistics, try looking at the whole second-hand-smoke debate. The data and statistics used against second-hand-smoke has gone to the extent of being comical in their assertions.
Second, I believe that many statistics that are presented to the mainstream are offered long before peer-review is completed. Media outlets seem to scan what is published while not waiting until peer review weeds out potential issues.
In areas that we have to hold as being statistically sound (government statistics – whether they are or not), you still have to have the consideration to weigh both the seen and unseen.
You said:
This fall in the average wage rate, however, does not signal that workers' fortunes are declining. In fact, in this case it is evidence of economic health: The economy is sufficiently flexible to provide jobs to workers who haven't yet acquired valuable skills.
This is a weird value judgment. In effect you are arguing against raising standards of living. Your value judgment that it is a good tradeoff to prevent a rise in the standard of living of employed persons to
increase employment. In the long run the bulk of increasing living standards comes from employer providing more capital to their employees. The usual incentive to do this is raising wages. But you are arguing that it is better to keep labor cheap and prevent the improvement in the capital-worker ratio that leads to higher standards of living.
Do you have any evidence that this is a good way to improve overall living standards?
Spencer,
Where in the world are you getting that from Dr. Boudreaux's commentary? All he was saying that if there are more unskilled workers entering the workforce than skilled workers, then the statistics will show median and average wages declining or stagnating, even though the standard of living of everyone may have increased a little bit. It's a sign of health in the economy because jobs would be plentiful in such a situation, where people with few job skills are still desired in the job market.
Spencer accuses Boudreaux of arguing against raising standarnds of living. What article did Spencer read? No where in this article or any other aricle written by Don does he argue against raising living standards. The quote in Spencer's commnet disproves Spencer's arguement: "The economy is sufficiently flexible to provide jobs to workers who haven't yet acquired valuable skills."
This is what contrasts the US with France. In the US, low skilled, low experience workers can more easly find employment to gain experience and aquire skills. This can create an illusion of declining wages or lower productivity, but as Don CLEARLY argues, this is just an illusion and is really a positive sign.
So, Spencer, why are you against rising standards of living?
Also, Spencer, Doc Boudreax was saying if you saw a decline in the average wage, you cannot automatically assume that things were getting worse for people, that drop in wages may have come about as a result of many new entrants into the job market at just about the time you decided to take a measurement, and everybody else would still be earning at about the same as before the measurement, and life would just have gotten a whole lot brighter for the people who just came into the job market.
He is not saying anything about sustained drops in wages across every segment making life better off or anything like the interpretation you took away from the article.
No, No, No — what he is arguing is that it better to keep labor cheap rather then the entire history of the US economy is that scarce and expensive labor is what leads to greater capital spending and raising standards of living.
There is no evidence that lowering wages leads to greater employment. I know the standard simple high school economics supply demand curve would lead you to believe that.
But it is just not true.
What leads to a growing economy and growing employment and growing real incomes is rising wages, not cheap labor.
Don's argument is a pure Marxist theory of capitalism that ignores the last 200 years of experience.
Spencer, Don is not a Marxist. Nor is his argument. His argument is not to keep wages low.
Read his article again. No where, NO WHERE, does he suggest that wages should be "kept low." Read the other stuff he has written and you'll understand.
You can will wages to be higher. You can't say "high wages are the basis for a good economy, so everyone should have high wages."
If you do that, what will happen is what has happend in France. If you dictate high wages, then only those who can produce enough to earn high wages will be employed. Everyone else will be out of luck. Which was the point of the original article. A point you did not, will not and maybe can not, get.
I believe this has elegantly proven Adam's statement that "You can use words to argue anything." Moreover, it further evinces how quickly many users of statistics jump to policy prescriptions without reflecting on what the statistics actually indicate. QED
Spencer
You are missing many points, I will focus on one. I think you are implicitly assuming that Don is suggesting it is beneficial to lower wages for all to raise employment. That is not what he is saying. He is saying permitting low wages to be paid gets the unskilled into the workforce, and his point is that although this lowers average wages it does not affect higher paid workers and gets people working who would otherwise be unemployed. This is almost certainly a superior outcome even though a reduction in average wage might suggest otherwise.
Spencer
I also think you misunderstand what it means to say the demand for labour slopes down. I think you are rejecting this because there are highly paid workers. "Doctors are paid plenty, but there is still demand for them, so demand cannot slope downwards". Is this your reasoning?
If so, what you are missing is productivity. High productivity shifts the demand for labour to the right. Doctors and lawyers are each sitting on different demand curves than shelf fillers at Wal Mart.
That is why minimum wages lower employment yet doctors can still command six figures.
Spencer has posted here before with similar statements. He doesn't respond to criticisms in kind, and will change the conversation to his way of looking at things. Previous attemtps at logical converstaion have bore little fruit.
Whether it's immigrants or youths, the stats will show the same picture. It's difficult to argue that the professor is suggesting one is better than the other.
As I've said before, longitudinal studies that show individuals as they progress through time is the only way to show if life is getting better.
It is interesting that Real Estate people are willing to break data into New Homes and Existing Homes, but we don't see this in salary data – Newly Employed, Experienced Employee, for example.
So nobody is willing to address the displacement issue?
Martin,
In the United States what is the unemployment rate of teenagers? How has the rate changed through time? Is there a historical correlation to immigration rates?
Are you suggesting that immigrants enter the country and new jobs are not created? It seems that the more people in the country, the more services, food, clothing, etc. these people will need. How do immigrants affect the natural employment rate?
Your link doesn't actually say that wages are falling in Ireland. I believe it says that wages are falling relative to the averages of the EU. I believe these are different things. If Ireland and the EU both average $18 per hour labor one year, and the next Ireland goes to $18.50 but the EU goes to $19 did Ireland's labor rate drop? I think I am reading your linked article correctly.
And even if labor rates are dropping, it doesn't mean individual's labor rates are dropping. New laborers with lower skills will lower averages. A real study would look at individuals and how their pay has changed through time. Averages tell us almost nothing.
As far as employee displacement, show the data. Your Irish articles don't say the jobs the "school leavers" used to get are filled by immigrants does it?
It would be irrational for a country to allow immigration if the only effect was causing citizens to become unemployed. Do you think that you have figured something out that the Irish government has not? Ireland is thriving now – some would say more so than it ever has.
A recent wave of immigration from Eastern Europe into Ireland will certainly cause the market to adjust in many ways. But only looking at the negative side of the immigration is problematic to say the least.
Don, did you send this article or one of your pithy editor-letters to the offending magazine? You should of course.
Recently there was a U of Nebraska study that said the Netherlands are now the tallest people by average, displacing the hale and hearty U.S.
I just heard it mentioned on some idiotic radio show at work, but the first thing I thought of was the average height of the thousands of Mexican immigrants streaming into my home state everyday. They are much shorter on average. Perhaps a state by state comparison would be in line. AZ would be falling quicker than say, Wyoming.
Point being is that simply being well read, and cognizant of statistics in general goes a long way to combating their misuse. So using stats in an argument is fair and expected.
Spencer says "What leads to a growing economy and growing employment and growing real incomes is rising wages, not cheap labor.
Don's argument is a pure Marxist theory of capitalism that ignores the last 200 years of experience."
Wages are prices, too, Spencer. What you want is high prices. It doesn't matter what your income level is, I'll always prefer a $2 Big Mac to an $11 Big Mac, the same goes for any other good or service, c/p. Do you think the thousands of mortgage borrowers who are defaulting under their ARMs find the higher payments to be a mark of prosperity?
No. High prices are indicative of scarcity. Scarcity is the opposite of abundance, which is the mark of prosperity.
What you're saying is that everyone would prefer to make more money, than less money. This is certainly true. But they would be just as happy if they could get by spending less, on the same income. Your "rising wages" argument is only remotely tenable under the assumption of some sort of constant price level, which is metaphysically impossible.
Furthermore, you've mentioned Marx here now, and I believe you've done it before. I do not think it means what you think it means.
Ray G – I don't think Don said that you can't use statistics in an argument; I did – in the context of why some wonder why the oft used response is always "you can use statistics to argue anything". If your opponent is quick to whip out that line about stats, you probably shouldn't use them because they fall on deaf ears. Then again, I just make up my own stats to throw back because there is no way to fact check it in a verbal conversation
j/k of course….
btw Martin, re-read your cited articles. The stats presented were presented covering the same series of data and one cited article goes on to say that immigration, with exception to the food service sector, isn't displacing Irish workers.
colson:
"I don't think Don said that you can't use statistics in an argument; I did"
I know, you're the one I was pointing to.
But I'm still not going to desist from using stats just because someone else is likely to misuse them. As a matter of fact, I particularly enjoy catching someone using statistics incorrectly. I'm not naturally argumentative, but I do really enjoy a good "gotcha" moment.
My favorite example while among those who may not normally think of such things is the stat about what dog was responsible for the most reported cases of biting people a few years ago. The lab, one of the most friendly dogs around. Of course, that's because they are so ubiquitous, but the statistically ignorant won't think of ratios and such. It's great to throw such a thing out while some hack is holding court on some subject at a dinner party or wherever.
Real wages were extraordinarily high during the Great Depression. What a wonderful time that was.
By the way, I'm almost positive that Spencer is just trolling. Still, if someone wants to volunteer to be the punching bag by making provocatively incorrect statements, I guess I'm fine with that. It gives the rest of us the opportunity to make our case.
Ray G – I know what you mean – it can be hard to resist. One of the funniest is studies that were showing more people dying of second hand smoke "related" ills that were greater than the number of people who die by first-hand smoking.
My favorite has always been when we're told that familiarity with roads leads to accidents, since most accidents occur within 20 miles of your home (or 10 miles, or whatever the latest number is). *Of course* most accidents happen close to your home, since we do most of our driving close to home.