Looking at DCCC Poll Claiming Rep. Issa at Threat

By Adam Probolsky

I don’t enjoy picking apart the work of other professional researchers, regardless of their political philosophies.

Yet in the case of the Democratic Congressional Campaign Committee’s (DCCC) polling memo published yesterday, which suggested that Rep. Issa could lose to a no-name challenger in November, the memo was issued without the benefit of any credible pollster presenting the research. I am not assailing the credibility of the their pollster (whoever it was). My problem is just that no pollster at all was named…hardly a flag of credibility. In fact, I take this as a sign that the pollster had so little confidence in the methodological practices used in the poll that he or she simply would not put their name on the release.

Let’s break things down:

• The memo does not indicate whether they polled mobile phones or whether they placed an artificial ‘cap’ on completed surveys on mobile phones. Limiting completes on mobile phones can fatally flaw research, especially in this district which has small unique pockets of ethic groups including Vietnamese, Chinese and Filipino voters as well has a large Latino voting population which are very difficult to reach without an focused effort concentrating on mobiles.
• Speaking of the 44,590 Latino voters in the district, the memo lacks any mention of Spanish-language interviewing or whether if Spanish was offered to those that prefer to speak Spanish in their home, a key factor which of course greatly improves the accuracy of research.
• There is no indication of what the actual poll questions were. This is more important than you might think. In California there is more to the ballot than just the candidates’ name, we have an amazing tool called the “ballot designation” that allows candidates to describe themselves. Incumbents have wide latitude insofar as what they can choose, such as “Incumbent”, “Congressman”, “U.S. Representative”, etc. Challengers are more limited with the three words that are available to them and it must simply describe what they do for a living with some variation. Applegate for instance, a long-serving Marine Colonel, cannot use his veteran status on the ballot because past positions  are not allowed – just ones you are currently occupying, unless you’ve done nothing since the last job. My point: If the ballot question put to respondents to this survey was, “Who are you voting voting for, Congressman Issa or Marine Colonel Applegate?” then Applegate will of course come out looking good in this Marine and veteran heavy district. But the results are meaningless because the ballot itself will look nothing like that.
• While the memo says they polled likely voters using the voter file, the memo seems to contradict itself by inferring it instead used census data when referring to the phrase “citizen age voting population”. “Citizen age voting population” has absolutely nothing to do with who is actually listed in the voter file, nor is it relevant to elections to talk about “citizen age voting population…” This calls into question whether they really used voter file data at all, or just succumbed to employing long discredited methodologies that rely on consumer data files.

The bottom line: Rep. Issa (just like many of his other Republican counterparts in the House and Senate) may have a more competitive race on his hands than he may have thought he would this year, but probably not as close that this polling memo suggests. To be sure, if you care about this race you will need to do your own poll for any accurate numbers, because you cannot rely on this one.

DCCC Issa-Applegate Poll Memo

Adam Probolsky is president of Probolsky Research