Ohio had their primary in early March, 2004 and at that point, there were four candidates left in the race - Dean, Edwards, Kerry and Kucinich. Kerry won the state, with Edwards coming in second. As everyone is painfully aware, Bush won the state in the GE. How do the various demographic groups compare between the two elections?

Kerry carried white males in the primary 53-34 and white women 51-37. (Unless specifically mentioned, Edwards is always the second place finisher.) In the GE, Bush carried white males 56-43 and white women 55-45. Oops, doesn't look like the primary was a good predictor for the GE.

Non-white males went for Kerry 61-22 in the primary and 73-27 in the GE. Non-white women were similar, 49-31 in the primary and 82-18 in the GE. Within these two groups, Kerry did significantly better in the GE than he did in the primary.

Let's look at age now. I'll look at how people younger or older than 65 voted since that shows a clearer picture. In the primary, Kerry carried both groups, under 65 50-36 and the 65 and older group 60-28. However, in the GE, he split the younger group 50-50 but lost the older group 42-58.

Let's look at income. I'll use $50K as the dividing line. In the primary, Kerry carried both groups, 52-35 for people making less than $50K and 53-33 for people making more than $50K. In the GE, he won the lower income group 58-42 but lost the higher income group 42-58.

Finally, let's look at education - those with a college degree and those without. In the primary, as expected, Kerry carried both those with a college degree 50-34 and those without 53-35. In the GE, he lost those with a college degree 46-53 and split those without 50-50.

In summary, how one does in a Democratic primary appears to have little bearing on how one does in the GE. Obama supporters can point to the fact that Kerry did significantly worse with the over 65 demographic in the GE and postulate that Clinton will do likewise. On the flip side, Clinton supporters could point to her primary weakness with non-white voters and make the same arguments. (Although that demographic isn't as large and doesn't have the same influence on the GE result.)

At this point in the process, you're probably better off looking at the crosstabs for those state polls doing GE matchups between McCain and either Clinton or Obama to see if there are differences in how various demographic groups respond than to make wild guesses based on primary results. And even then, who knows what will happen between now and November.