With the Iowa caucuses coming in a few days, everyone is watching the polls like a hawk. A few bloggers now have pointed to this latest Zogby poll that shows John Kerry opening a five-point lead. Lemme be the first to call “bullshit” on this one :

Candidate	1/13-15	1/12-14	1/11-13	1/10-12	1/9-11	1/8-10
MA Senator John Kerry	24	22	21	17	16	15
Former VT Governor Howard Dean	19	21	24	28	26	25
MO Congressman Richard Gephardt	19	21	21	23	23	23
NC Senator John Edwards	17	17	15	14	12	14
Retired General Wesley K. Clark	3	3	3	2	2	3
OH Congressman Dennis Kucinich	3	3	2	3	3	2
CT Senator Joseph Lieberman	1	1	1	1	2	3
Civil Rights Activist Rev. Al Sharpton	.1	0.1	0.1	0.1	1	1
Undecided	13	11	13	12	14	14

Something about these weird three-day polls that overlap each other rubs me the wrong way. Like all polls, even if the results can be somewhat trusted, the margin of error should be considered before drawing any conclusions. In this case, the MOE is big, and the additional caveats are just crazy :

Polling results will be released daily through Monday, January 19th, the date of actual caucus voting in Iowa.Zogby International conducted telephone interviews of a random sampling of 503 likely caucus voters statewide over a three-day period. All calls were made from Zogby International headquarters in Utica, N.Y., from Tuesday (1/13/04) thru Thursday (1/15/04). The margin of error is +/- 4.5%. Slight weights were added to party, age, education, union, and gender to more accurately reflect the voting population. Margins of error are higher in sub-groups.

Huh? Not only does the MOE almost completely cover Kerry’s lead, but they’ve been tinkering with the results. If they’re adjusting these findings to “more accurately reflect the voting population”, can I assume that whatever measure they’re adjusting against is based on polling as screwy as this one? And what exactly are these “sub-groups”, what’s the MOE, and why is it higher in some more than others?

With a sample size of only 503 people, here’s all you need to know : Al Sharpton’s 0.1% showing equals 0.503 people. So either only half a person supports Sharpton, or these results are garbage.