REVISITING THE U.S. VOTING SYSTEM: A RESEARCH INVENTORY

November 27-28, 2006

Convened by the American Association for the Advancement of Science

Main | Participants

Fritz Scheuren

USE OF EXIT POLLING AND TELEPHONE SURVEYS TO DEEPEN UNDERSTANDING OF ELECTIONS

An election is an exceedingly complex undertaking and the management of elections has arguably grown harder as a result of HAVA, if only because of rising expectations. Others here will focus on various aspects of US elections (See Exhibit A, available separately.) I look forward to learning from them. This short note focuses only on using exit and other polls as measures of the fitness for use of election results. Telephone and exit polls – as do other survey forms -- all have weaknesses that make them hard to employ to check whether an election outcome was fair. Articles, like those in the November 2006 issue of Chance, can help here (See Exhibit B, available separately); but, for many reasons, weaknesses in polls can still lead to controversy. The stolen election headline, even in 2006, is just too attractive, apparently (See Exhibit C, also available separately.) Even so, surveys, whether exit or telephone polls, can inform election officials and the public at large of voting issues, as seen by voters that might bear more attention next time. This is true even in an election, like 2006, which was viewed favorably by most media sources, unlike the 2004 results that were seen as more controversial. To illustrate this point we give an early partial example from the 2006 telephone data collected by NORC in Franklin County. Our focus is below entirely on absentee voters. (Data unweighted and not adjusted for differential nonresponse.)

Franklin County Results
Absentee voters

Table 1. -- Comparison of Characteristics of Households with Absentee Voters who were Respondents versus All Responding Households with Voters.

Gender	All (363)	Absentee (67)
Male	36%	24%
Female	58%	73%
Unknown	6%	3%
Total	100%	100%

Age	All (363)	Absentee (67)
Under 25	5%	1%
25 to 54	45%	19%
55 or over	42%	76%
Unknown	7%	3%
Total	100%	100%

pt'> <

Ethnicity	All (363)	Absentee (67)
Caucasian	70%	78%
African American	13%	10%
Asian/Pacific	2%	1%
Hispanic	1%	0%
All Other	4%	6%
Unknown	9%	4%
Total	100%	100%

There were 67 absentee voters out of the 363 Franklin County respondents. Only 2 questions relate to problems encountered in the process: Q47 – Were the instructions clear? And Q50 – Were you satisfied overall with the process? Of the 67 absentee voters: 61 had no problems indicated, 5 indicated 1 or 2 problems and there was one missing value i.e. one nonresponse to Q50 is included (labeled missing).

Q50 Satisfied overall?	Q47: Were the Instructions clear?
	Yes	No	Total
Yes	61	1	62
Missing	1	-	1
No	2	2	4
Total	64	3	67

One might also want to consider including Q52 = Do you think your vote will be counted accurately? There was only one absentee voter who did not think the vote would be counted correctly and this voter did not indicate dissatisfaction on the other two questions. So this would be a 6th voter with a “concern.” In the comments made on the questionnaire absentee voters surfaced issues that were to be given prominence in the media, notably the ambiguity about what identification number from the driver’s license number to use on the application. Apparently the absentee ballot asks for the number on an Ohio driver's license, but there are actually two numbers on the driver's license, and it was not clear which number was required. (See Exhibit D, available separately, for other absentee voter comments on problems and on the absentee process.) Reasons given for voting absentee included 40% who said they voted absentee to avoid long lines -- remembered apparently from the 2004 election (See the November 2006 issue of Chance for 2004 details. Shown as Exhibit B, mentioned earlier.) There were also some absentee voters (15%) who wanted to avoid electronic voting.