Monday, October 25, 2010

In Maryland: Two Polls, Two Very Different Findings, but the Same Results?

Two polls in as many days suggest that Martin O'Malley has opened a formidable 14 point lead over former governor Bob Ehrlich - the Baltimore Sun reports that O'Malley has a 52% to 38% lead over Bob Ehrlich. The next day the Washington Post reported an O'Malley lead of 54% to 40%. Reporters from both papers seemed less than convinced by the poll results, but took comfort in the seemingly similar results - two polls, two papers, two pollsters - a 14 point lead.

In reality, the polls are quite different. Consider:

The Washington Post sees an likely voter electorate that is 59% Democrat, 30% Republican, and 11% Independent or other. The Sun poll is 57%, 30%, and 13% - not much difference, but both likely overstate Democratic turnout.

Things get a little weird when you dig a bit deeper into the cross tabs- the Post reports that Ehrlich wins Republicans by a 92% to 6% margin and wins Independents 52% -36% - he also receives only 7% of the Democratic vote. In the Sun poll, Ehrlich is doing poorly among Republicans winning 78% to O'Malley's 12% (the rest are undecided). The Sun also reports that Ehrlich is losing among Independents 37% to O'Malley's 52% - a mirror image of the Post poll. Finally, Ehrlich wins 18% of Democrats to O'Malley's 72%. The Sun also finds Ehrlich losing among men, 43% to 48% and women, 34% to 55% - in the Post Ehrlich is winning men 48% to 46%, and losing women by 62% to 33%.

So, compared to the Post, the Sun finds Ehrlich doing far worse among Republicans, losing considerably among Independents, but faring much better among Democrats- yet the topline margins are the same 14 points. So how do they arrive at the 14 point margin?  The Sun assumes a higher turnout among Independents than does the Post, but has O'Malley winning them. The Post has Ehrlich doing well among Independents, but horribly among Democrats and over states Democratic turnout.

When I take the various results and turn them into more likely turnout and support assumptions - 56% Democratic, 31% Republican, 13% Independent - and assume, realistically, that Ehrlich wins Independents and Republicans by the margin in the Post, and Democrats by the margin in the Sun then this becomes a 6 point race - 53% to 47%, advantage O'Malley. I think that's about where this race is - and the Sun and the Post should resist using the other's polls to assuage doubts about each poll based on the significant differences evident in the cross-tabs. There can be no question, however, that Ehrlich is clearly behind and by more than just a few points.