More Detail on Why The Times Called the Presidential Election Later

In a post Wednesday, I noted that The Times took longer than most other major news outlets to declare the winner of the presidential election.

Rich Meislin, who is the other half of the “decision desk,” along with the associate managing editor Janet Elder, wrote to me Thursday providing this more detailed explanation of the Times's process.

Mr. Meislin, a consultant to The Times, was editor of news surveys and election analysis from 2003 to 2005, and he has been part of the decision team for the 2004, 2008 and 2012 presidential elections.

He wrote:

A bit of additional insight might be useful about our decision to call the race for President Obama later than the networks and The Associated Press. It was rooted in our lack of certainty about the result in Ohio, which pushed President Obama over the 270 electoral vote mark for the television networks beginning around 11:15 p.m.

First, some background: T he major television networks and The Associated Press are members of the National Election Pool, a consortium that shares the very high cost of conducting exit polls and collecting real vote counts from polling places across the country. The work of gathering and managing this information has been coordinated since 2003 by Edison Research, a survey and analytics firm.

News organizations other than the networks and A.P. subscribe to Edison's work, which provides us with exit poll data as well as guidance, based on Edison's own analysis, on when the winner in a state is ready to be called. The network members of the consortium and The A.P. have an information advantage over subscribers: they receive early access to exit poll results, faster real vote tallies and additional information on voting in key precincts that subscribers do not. This allows them to develop their own models to predict the outcomes in the states that are more sophisticated than subscribers like The Times can create with their more limited data.

We are cautious in our calls. We make our decisions based on a combination of sources: the exit polls, when they are decisive; the actual votes that are being reported compared with the historical performance in counties; the experience of our political staff; and the judgments of Edison's analysts and The Associated Press, which also has access to the expanded data. In the case of Ohio, the exit polls showed a very close race, and so did the real vote counts. The patterns we saw led us to believe Mr. Obama would win Ohio, but not with enough certainty that we felt comfortable to make a call when the networks did. (Indeed, the real vote total in Ohio tightened quite a bit for a while after the network calls before loosening again.) Edison also held back, even with the advantage of its data models, and that increased our caution; Edison ultimately called Ohio at 12:55 a.m.

In the meantime, we began seeing enough resul ts from Virginia, Colorado, Wisconsin and Nevada to allow us to call an Obama victory. (An advantage on Virginia came from Michael Shear, who was able to read the county results with the background of having covered the state for many years as a Washington Post reporter; we called Virginia well before the networks and A.P.)

And until we were able to make our own call, we told our readers that the networks had called the race, which we thought struck the right balance between keeping readers informed and being sure we had it right.