Sunday, May 14, 2006

A Few More Ohio Gov Numbers

First off, Bryan -- whom I am embracing as my new best friend solely to piss off his annoying troll -- has a post up at BSB crunching crosstabs numbers from a Zogby pre-primary poll (Evidently he is a Zogby Gold subscriber). It shows that Blackwell's strength is not with African Americans, but with whites and Hispanics. Interesting, happy news. We will need to keep an eye on those numbers.

Second, in following up my research about the previous post, I noticed a curious trend in the SurveyUSA primary numbers. Results among Black Republican likely primary voters shift sharply in the couple weeks leading up to the election:

Date: 4/22/06-4/24/06
Blackwell: 42
Petro: 51
Undecided: 7
% of Likely R Primary Voters: 3%

Date: 4/28/06-4/30/06
Blackwell: 64
Petro: 14
Undecided: 23
% of Likely R Primary Voters: 4%

A couple of factors may come into play here. The first is that the numbers are small, so a shift of only a couple of voters skews the results. Here are the numbers for the two polls where n is the number of likely Republican primary voters and n(aa) is the number of likely Republican African American voters:

Poll1
n=435
n(aa)=13

Poll2
n=428
n(aa)=13

After all, the MOE for the poll as a whole is a mesomorphic 4.8%.

But a possible second factor is the recruitment of African Americans as Republican primary voters. As I noted, the buzz around town is that the BOE noticed a high number of blacks picking up Republican ballots. Only Blackwell was reportedly trying to do that. Because the numbers are so tiny, a few recruited primary voters could throw them off.

Getting the canvass for Summit County is on my to-do list. It will be interesting to see if Republican primary voters spike in the predominantly black wards.

6 comments:

abdirissa said...
This comment has been removed by a blog administrator.
staff said...

yeah you cant rely on a sample of 13 people. it just isnt scientific. It's like using what your friends say at the dinner table.

At least now i hope we can all agree i was right from the start about Blackwell's penetration into the Dem AA base ? And perhps get a little more respect around here.

Anonymous said...

It actually is scientific if the the 13 people are selected randomly. the margin of error will be big, but it's still "scientific".

Anonymous said...

i also don't select random people to eat dinner with me. sayin'.

bryan said...

i'd say it's 0% MoE. can there be more than 13 likely aa republican primary voters in ohio???

ok, yeah, there are, but total population size has to be taken into account when looking at MoE.

be that as it may -- blackwell's race will be an issue. i just agree w/ pounder that the current narrative is the wrong one.

Anonymous said...

Bryan,
Total population size does not have to be taken into account when looking at the margin of error. the margin of error is purely a function of sample size.