Readers of The Sun, not to mention this blog, may remember that I had a story last Thursday about a study showing that sales of less expensive homes were actually rising year-over-year despite the steep falloff in sales overall.

Alas -- the study was wrong.

Delta Associates, a real estate research firm that prepared the report for multiple listing service Metropolitan Regional Information Systems, said today that it included condos in its calculation for the first three months of this year but excluded them in the first quarter 2007 calculation. That apples-to-oranges comparison made it look like cheaper homes were selling like hotcakes when in fact they too were seeing sales declines.

Delta says now that single-family home sales -- that is, not condos -- fell 26 percent in the $120,000 to $249,999 category. (Its original calculation was a sales increase of nearly 20 percent.)

Ouch.

So why did no one -- either at Delta, at Wonk Central or in the Realtor ranks I interviewed -- suspect there was an error at first? Because all anecdotal evidence has for months suggested that this price range is doing significantly better than more expensive homes. Over and over, agents have said there are more buyers interested in under-$250 homes.

And they're right, relatively speaking. Cheaper homes haven't seen as steep a sales dropoff as more expensive ones. Sales of homes priced at $250,000 and above declined 37 percent in the first quarter of this year vs. the first quarter of last year, according to Delta's new calculation. (That includes a dropoff of almost 50 percent in the $500,000-plus range.)

It's a useful reminder that data errors can crop up, even at reputable research firms. And it's made me even more paranoid about the accuracy of any analysis that I haven't done personally. (Which is not to say I'm not extremely paranoid about potential data foul-ups in my own analyses, because I am.)

We'll have a correction in tomorrow's Business section.