Elections have become a battle of the “polls.” A recent poll released by the Mike Downs Center for Indiana Politics would lead a reader to surmise that the election for the Third District Congressional race is all but over with Stutzman leading by a whopping 25%.
However, that supposition would be entirely inaccurate – just as the Downs Center poll is inaccurate. The following is the methodology that was used:
Statement of Methodology for U.S. House 3rd District of Indiana: SurveyUSA interviewed 565 registered voters from Indiana’s 3rd Congressional District 10/21/10 through 10/25/10, using Registration Based Sample (RBS) from Aristotle in Washington DC. Of the registered voters, 400 were determined by SurveyUSA to have already voted, or to be likely to vote in the 11/02/10 midterm election. Where necessary, responses were weighted according to the voter registration database. In theory, with the stated sample size, one can say with 95% certainty that the results would not vary by more than the stated margin of sampling error, in one direction or the other, had the entire universe of respondents been interviewed with complete accuracy. There are other possible sources of error in all surveys that may be more serious than theoretical calculations of sampling error. These include refusals to be interviewed, question wording and question order, weighting by demographic control data and the manner in which respondents are filtered (such as, determining who is a likely voter). It is difficult to quantify the errors that may result from these factors. Fieldwork for this survey was done by SurveyUSA of Clifton, NJ.
Note the bold and underlined portion of the statement. In actuality, the responses were incorrectly weighted with Allen County receiving an inaccurate percentage. The following are the Third District registered voter totals from the Indiana Government website:
Allen – 230,789
DeKalb – 28,999
Elkhart – 58,944
Kosciusko – 52, 328
LaGrange – 15,319
Steuben – 23,125
Whitley – 20,518
Total Third District voters – 456,677
I don’t claim to be a statistician; however, if Allen County has over 50% of the Third District voters, then a sampling should include at least 50% from Allen County. This was not the case with the recent poll. My understanding it that Elkhart County was attributed a 25% share of the total even though it is just shy of 13% of total Third District voter registration while Allen County’s percentage was nowhere near the actual 50% + of the Third District sample.
This inaccurate percentage could have been the result of looking at the entire county of Elkhart, not just the portion located within the Third District. The bump to 25% is twice what should have been attributed to Elkhart County and skews the poll results.
In defending its poll, the Downs Center through Andy Downs stated that he believes his data is more accurate because he surveyed those likely to vote, not just those who are registered to vote. So, just what is the likely voter vs. the registered voter distinction upon which the Downs Center considers its poll to be more accurate?
A registered voter is just that – someone who is registered but may or may not vote. A registered voter could skip any number of elections and vote sporadically depending on issues or candidates that are important to that voter. On the other hand, a likely voter is one who has “more than likely” voted in the last two or three elections and may very well continue that trend. A series of questions asked by the polling firm is used to establish whether or not the contacted, registered voter will vote in the current election and what that voter’s preference is at that point in time.
But how large a gap exists between the actual percentages between registered voters and likely voters? In a recent article, the author noted that from 1950 to 2006 – where relevant Gallup poll data was available for 13 midterm elections – the average gap between the preferences of registered and likely voters was only 5 points. Only once, in 2002, did the gap reach double digits.
The newly released poll by the Hayhurst campaign more accurately reflects the actual differences noted by the author of the above article. The Hayhurst poll shows a 4% lead by Dr. Hayhurst with a plus or minus deviation which indicates a possible dead heat in the race for the Third District congressional seat.
The Downs poll is flawed for two reasons: 1) the polling data was not accurately distributed within the district; and, 2) the research does not support the proposition that likely voters vs. registered voters provides a more accurate picture of just who will win this election. If history serves, the difference is a 5-point margin – but that margin in this race leads not to the conclusion that the race is over but that it will be a tight race on election night.
Flawed polls do a great disservice to the voting public. And, is so often stated, the only poll that matters is the one on election day.