On the methodology of electoral vote trackers

Someone has urged me not to pay much attention to the electoral vote trackers

on the front page of MyDD, because in some respects they differ from state polling averages you will find at pollster.com or at Real Clear Politics.

Jerome Armstrong, the founder of MyDD, addressed concerns about the electoral vote tracker in this post:

As it says, when you click on either of the map counters of EV’s:

“This Electoral Vote Map is updated constantly to forecast the 2008 Presidential election based on the latest available state polling.”

The very latest poll in each state, without weighting or averaging.

There isn’t a bias as to the pollster, if you see the poll listed as credible on Pollster.com, or RealClearPolitics.com, it’ll be included. But, if the latest poll is tied, then the result remains the same as the previous latest poll.

This is a forecast made by the very latest poll. If you see a mistake, perhaps a poll was missed that is the latest, then point it out, and one of the admins will make the change.

The forecast isn’t a prediction of the election, but a simple up-to-the-minute poll temperature of the state polling.

(update) And yes, you can edit the map yourself, as one user explains:

1) When you first log onto mydd, it populates the two maps with the most recent single poll for each state.

2) If you then click on the map and change it (for example, you don’t believe that Hillary would lose WA to McCain), the numbers update to your settings – now it becomes like an EV calculator

3) The next time you log on, or refresh the page even, the counters go back to their poll-generated state.

Polling averages (for instance, of the five most recent polls in a state) are great when you have several polls taken within a short period of time, as we’re likely to have this fall.

But I don’t think it makes a lot of sense to average the last five polls in a state if that takes you back a couple of months.

Some people have objected that the MyDD tracker recently showed Iowa in Clinton’s column against McCain. That was based on a Research 2000 poll taken on April 22 and 23, which showed her slightly ahead of him, 43-42. Currently, the most recent poll is from Rasmussen on May 13, which showed McCain beating Clinton in Iowa 45-42. As you can see if you click over to MyDD, Iowa is now red for McCain against Clinton.

You may prefer polling averages to whatever the latest poll says, and I will too, once we start getting more frequent releases of state polls. For now, I think that MyDD’s methodology is sound.

Armstrong is probably the blogosphere’s most prominent Obama critic, and Clinton supporters usually dominate the recommended diary list at MyDD, but I encourage you not to write off everything you see at that site, even if you don’t like Clinton. Anyway, front-pager Jonathan Singer is a big Obama advocate.

About the Author(s)

desmoinesdem

Comments