Callahan’s Cleveland Diary

The U.S. Census released its initial American Community Survey data for 2006 this morning.

The Plain Dealer’s first online take is all about not being the “poorest city” again. (It seems we were only the fourth poorest this time.)

But by tomorrow morning, expect the headline to be: Cleveland falls below 400,000.

The new ACS estimates Cleveland’s 2006 population at 396,910… plus or minus 10,705.

How seriously should anyone take this? Well, just two months ago another Census release put the number of Clevelanders almost 50,000 people higher than today’s ACS, and the PD gave it most of page 1A.

I wonder if they’ve hit their limit.

Update 8/28: Well, apparently they have. This morning’s PD features the usual big-front-page coverage of the ACS release, but the word “population” doesn’t appear in it. “Poverty” is the word of the day.

Good for the editors.

And compliments on a generally informative first look at the 2006 income numbers. Though unfortunately they couldn’t resist this:

Along with shedding its unwanted mantle as America’s poorest big city — dropping to fourth behind Detroit, Buffalo and Cincinnati — Cleveland saw median household income rise more than $1,700…. In both Lorain and Youngstown, the percentage of poor jumped dramatically… In Lorain, poverty shot from 17.6 percent in 2005 to 26 percent — the biggest leap measured in Ohio.

No, guys and gals. The Census people are not asserting that these things actually happened between 2005 and 2006.

What they’re saying is that they did a survey of a sample of households each year, and in 2005 the results from this sample indicated that (for example) the median household income for the community was 90% likely to be somewhere between a and b, whereas in 2006 it was 90% likely to be somewhere between x and y.

This is not the same thing as the Census saying: “We know the median household income in this community was (a+b)/2 in 2005, but changed to (x+y)/2 in 2006.” But unfortunately, that’s what the PD story is reporting.

The ACS for 2005 found, with 90% confidence, that Cleveland median household income was somewhere between $22,750 and $25,460. The new ACS reports, with the same 90% confidence, that the city’s median household income in 2006 was somewhere between $25,415 and $27,655. Note that the bottom end of the range for 2006 is $45 less than the top end for 2005. Note, also, that the ACS estimate for 2004 (same 90% level of confidence) was between $24,579 and $30,797.

So did the average income of Cleveland households actually drop by $3,500 the year before last, then jump back $1,700 in 2006? It’s very unlikely. And it’s very unlikely that 8.5% of Lorain residents suddenly fell into poverty in one year. The annual ACS numbers are survey research artifacts, not hard measurements. Like political tracking poll numbers, they’re very useful to show trajectories — the direction of change over time — but highly unreliable (or maybe I should say, reliably approximate) as single snapshots.

That’s why the shift of a few places in year-to-year poverty “rankings” of Cleveland and other cities — last place, fourth-last place, tenth-last place, etc. — mean just about nothing. (On the other hand, there’s little question that both Cleveland and Cincinnati are really among the poorest 10% of big U.S. cities, because we’ve shown up there consistently over years of surveying.)

While I’ve given the PD a lot of grief over the years for their Census Front-Page Extravaganzas, it isn’t because I think there’s something wrong with featured coverage of data about important issues. On the contrary, I respect the paper’s commitment to capturing its readers’ attention for what little real empirical information we have — especially about incomes and poverty.

But I really think that if you’re going to make a big front-page deal about survey research results — especially on politically loaded topics like local population and poverty — you owe it to your readers to explain what the researchers are actually saying, how they got their results, and what degree of certainty (if any) the results justify.

I know, eyes are glazing over. Oh well. That’s the price of serious journalism.

This entry was posted on Tuesday, August 28th, 2007 at 7:24 am and is filed under Media, Poor Cleveland, Shrinking city. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.