An update on yesterday's discussion of the ABC/WaPo poll on Social Security -- and this one's a pretty big deal.

Last night I wrote that the poll showed that 53% of respondents favored private accounts but that number dropped to 46% (with 47% opposed) when costs of up to $2 trillion were added to the equation.

But it seems that's incorrect and the real finding was far more favorable.

I went back and looked at what appears to be a revised version of the Post story I referenced yesterday to see exactly what they said. But even the revised version states: "The president also has at least general support from 53 percent of the public for the concept of letting people control some of their contributions to invest in the market ... Support dropped to an even split when people were told that the cost of the transition to a new program could reach $2 trillion over time ..."

The original version of the article, I believe, provided the specific numbers, but this reference to an "even split" is clearly a reference to the numbers falling to 46% for, 47% against. Or at least I can't see any other way of interpreting it.

All I can figure is that the authors of the piece weren't clear on what the numbers actually said. (I don't say that lightly; they're both solid reporters. But I don't see any other way to interpret how they characterize the data.) This is clear in ABC's compilation of the data; and it's noted in their write-up of the poll as well. Indeed, if you look at the section of the Post's complete poll data which references this question, you'll see quite clearly that this question was "asked [only] of those who support a stock market option."

That means that it's not 46% of respondents who still support private accounts but only 46% of the originally-private-account-supporting 53%. Or, in other words, just under 25% of respondents support Bush's plan once they know the costs and an overwhelming 69% oppose it.

That's a pretty big difference.

[ed. note: A special thanks to readers JD and SC for flagging my attention to this point.]