This is an interesting panel for me as I tend to get tons of calls from pollsters. In fact I may be on their speed dial. Because of this I am supremely interested in attending this panel and hearing the very nuts and bolts information that is promised by the panelists who are both experts in this field.
This was an informal panel with the moderator asking questions and allowing the panelists to lead the discussion. As such, I’m providing the meat of what each expert shared.
Ed Morrissey: You have to be skeptical of Democratic Pollsters. There is an art to polling likely voters; so weight your results accordingly. Headlines are a result of that weighting. Top line results hinge on doing this properly.
Analysis on current polling numbers: If you are leaning to the right in the presidential race the current polls are very encouraging. An incumbent president has a lot of advantages, because historically very few incumbents run and lose the presidency. The incumbent can change the narrative, through policy changes during the race. The immigration policy change was sorely needed by Obama to end a horrible run for him in the campaign. This is a move that can move polling in Obama’s favor because the change makes him look like a leader. The bump may be small, but it is a calculated move in the hopes that it will benefit him.
Scott Rasmussen: People assume that the poll that shows the better numbers is the better poll. Don’t do that. Look at tons of different polls. Good polling is a result of a good mix of people in the sample and proper methodology. Rasmussen Reports does online polling now, because many young folks don’t have land lines. New tools to reach different people in different ways. Speaking to the topic of likely voters: it’s fairly easy to get a sample of the population. Even if you do the process exactly right, you aren’t guaranteed a proper result. Question wording is very important, although there is no right question. There is value in looking at question wording when comparing polling data.
Daily Tracking Poll mechanics. Asks a series of questions of 500 people, which are combined with two previous days. The next days data is added and oldest days data is dropped. This reduces statistical noise and gives more accurate data. National tracking polls are the best indicator of how the presidential race will turn out. Three points wins the electoral college for either presidential candidate.
Intensity Indicators are moving in the favor of the Republicans who are motivated by the desire to remove Obama. The job approval number is the single most important indicator to follow and is related to voters personal finances.
Conclusion: look at tons of polls. Don’t just find the one that you like and quote that. Excellent panel. Huge thank you to both Mr. Rasmussen and Mr. Morrissey!