Democrats for Brown in latest PPP poll: 19 percent.

Democrats for Brown in latest ARG poll: 20 percent.

Democrats for Brown in latest Suffolk poll: 17 percent.

Democrats for Brown in latest Daily Kos/Research 2000 poll: 9 percent.

This is where I usually say, “I respectfully dissent…” but when the sample is this bad, I just don’t respect it. I think what jumped out at me the most is that in this poll, 21 percent of Republicans have “no opinion” on Martha Coakley; 16 percent have a favorable opinion of her. Eh, no.

They have her favorable/unfavorable split at 58 percent/31 percent. That seems awfully high for a woman campaigning on keeping nuns out of emergency rooms, the idea that shaking hands at Fenway Park in the cold is silly, and that Curt Schilling is a Yankee fan. With numbers like that, running as a Democrat in Massachusetts, you would expect her to be way ahead.

I realize getting a good sample and projecting the turnout in a special election is hard. And I suppose it’s theoretically possible that everybody else is wrong, and that the Kos/R2000 folks have figured this all out, and the electorate will look like that. But I doubt it. And I recall that this poll had Christie by 1 late in the New Jersey governor’s race (he won by 4.3 percent) and McDonnell by 10 in the Virginia governor’s race (he won by 17.5 percent).