Thursday, October 29, 2009

Misleading percentages

When is a fact less than a fact?

When the fact is a percentage of a very small absolute number.

For example, I can state with near certainty that I gave more to philanthropy last year than most other Americans. Indeed, while individuals in general decreased their giving by 2.7 percent (even steeper after inflation), according to the Giving USA Foundation, I increased my giving by 100 percent!

What I'm not revealing, of course, is that my altruism amounted to precisely one dollar in 2007, and to two dollars the next year--a one hundred percent growth rate.

Okay, this is an extreme hypothetical example, but real life examples abound, especially in political campaign literature. A critique of the Michael Bloomberg campaign for Mayor in New York City, for example, cites US Department of Education figures showing declines in violent crime at certain Impact Schools during Bloomberg's reign "as large as 59 percent for major crime...and 33 percent for all crime...." But it adds immediately that "the numbers on which these percentages are based are so low that even very small numerical deceases create large percentage changes."

That's the key point. In one NYC high school, the critique notes, violent crime dropped forty-one percent between the 2004-5 to 2005-6 school years. A phenomenal decline? Not when you realize that the number of incidents in the earlier period was only seventeen, and in the later period ten. To be sure, any decline in violence is good news, but the percentages cited present an unwarranted picture of astounding success by the Administration.

As noted in one of my favorite books, "How to Lie With Statistics" by Darrell Huff (1954), "Percentages offer a fertile field for confusion. And like the ever-impressive decimal they can lend an aura of precision to the inexact."

For journalists covering political campaigns--or anything else--it's good to keep in mind Huff's admonition: "Any percentage based on a small number of cases is likely to be misleading. It is more informative to give the figure itself." (Emphasis added)
Full disclosure: The critique cited above was prepared by the former leader of a tenants organization in which I was an active member.

Tuesday, October 20, 2009

Citing imports to gauge US trends

Unless they work for Goldman Sachs or JP Morgan, even the rich are having problems these days. In fact, that's the theme of the story your editor assigns to you. Give examples, s/he says, of how the rich aren't buying as many luxury products as they once did.

You select jewelry (the non-costume kind).

Where to start? Typically, you call Tiffany's, and ask how business is doing. Or, you go to Tiffany's website, or to the website of the Securities and Exchange Commission, and get the company's annual report (10-K, in governmentese). From that, you compare this year's sales with last year's, and draw the appropriate conclusion.

But Tiffany's, no matter how prominent a luxury retailer, is still only one. There are many others, and most of them are privately held, which means they aren't likely to release their revenue or profit and loss figures.

So here's another approach: find out how the countries that export the most jewelry to the United States have fared in 2009 versus the year before.

Start with one of the government units that tracks imports and exports. It's the International Trade Administration. (There are others, but don't bother about them now.) Then click on "Consumer Goods."

On the left, you'll see a list of industries: click on "Jewelry."

When the new window opens, look for "Current Imports," and under that "Jewelry (except costume)." Clicking there yields a page of how all the major exporters--India, China, Thailand, Italy, Hong Kong, France, etc.--are doing. Glancing at the very last column on the right, you can see that all of them have sharply lowered their exports of jewelry to these shores, at least over the first six months (from January through June). India is down more than twenty-eight percent, China more than thirty-seven percent, Italy almost forty-two percent.

So this column contains the most important figures for your story. If these countries are shipping so much less, it can mean only one thing: the rich (and likely the upper levels of the middle class, which is also a vast market for jewelry) simply are not buying as much as they used to.

That's a good statistical grounding for your report.

Friday, October 16, 2009

Bigotry resurgent

Three news items of recent days belie the oft-touted boast that, in light of Barack Obama's election to the Presidency, the United States is now largely "post-racial." Indeed, Obama himself has said as much. Further, with all the gains made by the gay and lesbian rights movement, we are said to be done with bias in that arena. And for a long time now, we've certainly been religiously tolerant.

Really. On October 15, a white justice of the peace in Louisiana denied a marriage license to an interracial couple, arguing that his sole concern is for the possible offspring of such a union, as he is convinced that neither white nor black society will readily accept them. Justice Keith Bardwell is worried that the hypothetically future kids will be treated, alas, unfairly. Unbelievably, he reportedly told the Associated Press, ''I'm not a racist. I just don't believe in mixing the races that way.'' You can't make this stuff up. (Note: Bardwell resigned in November.)

Also that day, 53 House Republicans have demanded, in a petition to Obama, that he oust Kevin Jennings from his Administration position promoting school safety "because of his career as an advocate of teaching tolerance of homosexuality," according to a report in the New York Times. After all, they continued, the message of tolerance--which they label a "pro-homosexual agenda"--runs counter to the "values that many parents desire to instill in their children." Seems that to some parents, and to lots of Congresspeople, tolerance is simply intolerable.

And finally, another four House Republicans are warning that Muslim "terrorists" are infiltrating U.S. institutions, including Congress, by getting positions as interns. And no doubt by selling Halal food to unsuspecting Americans. And by refusing to drink alcohol.

That sound you hear is coming from the late Wisconsin Senator Joseph McCarthy, cheering and applauding from his seat of honor in hell.

Wednesday, October 7, 2009

Subjectively objective

Your story on how well or poorly American families fared in 2008 compared with 2007 is due shortly. You've completed all your interviews with selected families, and with two respected professors of economics. Income, not surprisingly, is the first issue you tackle. Here's where your hard facts come in. . .or so you think.

For your most important hard fact, you will cite data on income. But reviewing your notes, you see that one professor said families increased their income by nearly a thousand dollars during the year, while the other professor said household income declined during that time. Not only that, but the first professor said that 2008 income was over $67,000 per family, while the second maintained that, no, it was only around $52,000.

Both professors cited the same source: the U.S. Census Bureau.

Neither professor is wrong. The Bureau, rightly regarded as the gold standard in data collection on income and population, is also (in)famous for its blizzard of statistical reports, so many of which appear to present conflicting data, on the same subjects, for the same periods of time.

What does this mean to you, the journalist? It means you have to visit the Bureau to find the "correct" figure yourself. Not a daunting task, but it is essential that you pay close attention to the definitions the Bureau uses. For example, it makes a distinction between families and households. To confuse matters more, it has a separate definition for family group. And another for family households.

So in one report, Median Household Income in the Past 12 Months by State and Puerto Rico: 2007 and 2008, you will see that the estimate of median household income in 2008 was just as the second professor noted: $52,029, around six hundred dollars less than in the year before. But in another report, Median Income for 4-Person Families, by State, you'll find 2008 median income for this particular family composition just as the first professor said: $67,019, or some nine hundred dollars more than the previous year.

Further, one report uses fiscal years, the other calendar years.

There's really no rule of thumb regarding which figures to use. Most likely, your decision will reflect whether you want to show that we're worse off now, or better off. In other words, no matter how "objective" you think you are, to some extent the hard facts you choose to cite will reflect your subjective view. That's inevitable. If you wish to be honest, alert your readers to the differences between households and families (essentially, households may include people who are not related to each other). Full definitions are available at the Census Bureau site.

In the end, to maintain your integrity as a journalist, don't wave the flag of "objectivity." That term is a lot murkier than it appears.