Democrats, the majority of whom have favored Hillary in the primary contests held to date,
So let's break this down. Penn can't talk about overall votes in caucuses and primaries, because even without scaling the caucuses up to primary-type numbers, Obama is beating her (see earlier post). Obama is even beating her if we just look at primary votes. So what is Penn talking about? Ah, it's the "Democrats." And maybe even the "favored." Turns out Penn is referencing a little spreadsheet at The Perfect World that took all the exit polls, scaled up the self-identified Democratic voters' preferences and added them up. So if Democratic voters made up 80% of the respondents in the exit poll, and 60% went for Clinton, then to get Clinton's total you'd multiply the total votes cast in the election, say 1,000,000, times .8 times .6 to get 1,000,000*.8*.6=480,000 Democratic voters for Clinton. Doing this, we get a pretty sizable lead for Clinton among these voters. There are 5 reasons why this is bullshit and anyone referencing these numbers should be ashamed. In ascending order of number-crunching geekiness, they are:
1. States determine their primary voting rules. If a state decides that it wants independent (or even Republican) voices in the primary, a campaign doesn't get to ignore them. Would the Clinton campaign care about this number if it was losing among Democrats? Why not restrict just to people who have voted in the last 5 Presidential elections, to make sure they're really Democrats.
2. There's no good way to account for caucuses with these numbers. Because fewer people vote in caucuses, they can't be added in directly, which is what the spreadsheet does. Caucus states count too. Just as silly, the spreadsheet includes Michigan. Michigan!! Obama wasn't even on the exit poll, let alone the ballot, and you want to count those voters? Of course, it includes Florida too, as well as excluding DC because there were no exit polls there. Guess it's Obama's bad luck to win votes where there weren't exit polls. Oh, and by the way, the spreadsheet numbers for New York are totally wrong, and gave Clinton an extra 200,000 votes. No biggie, I guess.
3. Exit polls are unreliable. They're supposed to be used to predict outcomes, not to select the winner. As an example of how unreliable they are, look at the New York exit poll. People who "usually think of [themselves] as " Democrats voted 60/37 for Clinton, while independent or something else went 55/40 for Obama. There's just one problem. The New York primary is a closed primary -- you can't vote in the Democratic primary unless you've been a registered Democrat for at least 25 days. In other words, those "independents" who went for Obama just "usually think" of themselves as independents -- they're actually registered Democrats.
Think about what this means. These numbers are no longer talking about what Democratic voters want -- they're talking about what "people who usually think of themselves as Democratic voters" want. Is that the test now -- the Democratic nomination should be decided by only those people who really feel Democratic, not those people who are registered Democrats, but like to think of themselves as independent-minded?
4. To continue on this line, if you add up the votes from the closed primaries (CT, DC, DE, LA, MD, NM, NY), which is the only reliable way to count the votes of just Democratic voters, Obama is winning (note that this includes New York, which Clinton won by a quarter of a million). Of course, we should really include the closed caucuses as well, but we'll leave the crying about caucuses for another day.
5. The spreadsheet overestimates Clinton's support by assigning all Democratic voters to either Clinton or Obama, when in fact a few percentage points backed other candidates. Take away that and the weirdness in Michigan, and Clinton doesn't have a "majority" of Democratic voters, period. Actually, I'll do you one better. Include Michigan, and don't include Texas/Ohio, which hadn't happened when this memo was written. She still doesn't have 50%. Throw in the caucuses (which are hard-core Democratic voters, you would think), and you can add in all of Clinton's states, she still doesn't have 50%. So Mark Penn's statement was just flat-out wrong. There is no legit way to spin it to be correct.
No comments:
Post a Comment