Jackson Jambalaya

The Clarion-Ledger began today a series of stories that will run between now and primary election day based on a poll recently completed by the Southern Research Group. SRG is a good general purpose marketing research and political consulting firm based here in Jackson on Briarwood Drive. They are not what I would consider to be a pollster along the lines of Zogby, Gallup, Mason-Dixon, Field, et al. But I have had the opportunity over the years to review the guts of other public opinion polls they conducted and have found their work generally good and practices sound.

Blair Goldstein reports that SRG's sample size was 472 people and the poll MOE is +/- 4.5%. (Sample size calculator. Jackson pop = 175,710) Respondents were contacted between Monday and Wednesday of this week and represented all Jackson council wards.

The Clarion-Ledger hasn't provided the guts to the SRG work. I wish that they would and urge them to do so. Because as I was reviewing the small amount of data that was included with Blair's story I couldn't help but notice the problem with the Ward 5 results. Take a look at the graphic below. The table is directly from the story and is the basis for the graph. Can you see the problem?

I don't know where the problem is without reviewing the data behind the poll. But I do know that there is a problem with the Ward 5 results. Even with the most generous interpretation of the results and margins the data for that ward falls far outside any range that could be considered reasonable.

Frankly I expect this basic level of analysis to get missed over at the JFP. But surely there is someone at the Clarion-Ledger who should have caught this. If not in their newsroom then there must be someone in their finance department with enough knowledge of statistics to see the problem.

As the C-L plans to use this same data on Sunday to update their readers on the actual race for Jackson mayor I'd hope they'd jump in with SRG, find out what happened in Ward 5 and fully disclose the nature of the problem.

AND, OR, they could release the detailed poll data to the public and let us take a look. Despite the flaws with his effort, kudos to Chism for doing so.