Preparing for Election Night at CBS News

By Tim Vercellotti

NEW YORK, NY (November 2, 2008)...While Election Night is not until Tuesday, I have been living and re-living it already for the past couple of days.

I have the privilege of working as a consultant with the decision desk at CBS News on Election Night. I am getting a front row seat on how analysts review exit poll data and early returns to project winners in the races for president, governorships in 11 states, and seats in the Senate and the House of Representatives.

Getting ready to project winners involves months of preparation, including compiling and absorbing hundreds of pages of background material on registration and voting trends in the states, and careful calibrations of statistical models that will be in use on Election Night.

Preparing also consists of conducting Election Night simulations, in which exit poll and voting data arrive in real time, allowing the decision desk to test its models and anticipate conditions on Election Night.

My role is pretty straightforward. Once the decision desk makes projections, I communicate them via e-mail to the rest of the staff at CBS News. In between projections I examine the exit poll data for insights into how a candidate fared among specific groups or in a particular state, then send those findings to the rest of the staff for possible use in the broadcast.

The rehearsals we have had so far reveal just how much care goes into interpreting the data and making projections. The analysts on the decision desk are a seasoned team of CBS News elections and survey experts, as well as some of the top political scientists and political consultants in the survey research field. Getting it right is the top priority while working under deadline pressure that feels real even in the rehearsals. Team members deliberate, re-examine data and assumptions, and check each other before agreeing on a projection.

It is what we political scientists set out to do with our research, although our work unfolds at a much more leisurely pace. You begin with a research question, develop hypotheses based on certain assumptions, test the data against those hypotheses, then examine your results. But you question the results, challenge them, hold them up to experience, and subject them to the review of your peers before reporting what you have found.

On the decision desk, that can happen in minutes. I will be mindful of that the next time I am rushing to finish a conference paper or journal manuscript. Pressure? We academics don’t know the half of it.