Steve Hill, the grand protector of IRV, responds to criticism from CAVEC:

1. CAVEC’s study suffers from serious methodological flaws. The biggest methodological flaw is that CAVEC inexplicably lumps together voters who “dislike” ranked choice voting (RCV) with voters who have “no opinion” about whether they like RCV (see pages 5 and 6 of the handouts from the press conference, particularly the bar graphs on those pages). For example, on page six (bar graph chart entitled: General Attitudes about RCV — Like/Dislike by Ethnic Group), look under the “Asian” bar graph. David lists 46% of Asian voters in his sample “like” RCV, and 54% “dislike/no opinion.” How in the world does David Lee decide that “no opinion” means “dislike”? This incorrectly gives the false assumption that more Asian voters seem to not like RCV than like it, with a larger bar graph for that combined category than the ‘like’ bar graph. One could just as easily say that “no opinion” means that voters “like” RCV, but that would be as disingenuous as what David Lee has done here.

David does this again on page seven of his handouts, where he asks the question “Was Ranked Choice Voting easy or difficult to understand and use?” Here, he lumps together voters who found RCV “difficult” to understand/use with voters who replied “neither difficult or easy” (see page 7, where 67% of voters report RCV was easy, 18% difficult, and 15% “neither,” and note how David lumps together the “difficult” and “neither” categories to come up with a total of 33%!).

This is an indefensible methodological practice. In fact, it’s completely unethical. David Lee has misrepresented his numbers and conclusions, he has twisted the numbers to try and lead the media to conclusions that are completely false and unsubstantiated.

Since the CAVEC press conference on Wednesday morning, one journalist asked David to clarify why he lumped together the “dislike” voters with the “no opinion” voters into the same bar graph/category. This journalist asked him to break down the numbers on page 6 which show 46% of Asian voters “like” RCV and 54% “dislike/no opinion” on RCV, since it is indefensible to misrepresent the numbers this way. David declined to clarify his reasons, but his colleague Ben Tulchin sent out a subsequent email at 4:33 pm breaking down the numbers. Here’s what they are:

26 percent of Asians ‘like’ RCV;

20 percent ‘like RCV a little’;

18 percent ‘dislike’ RCV;

36 percent ‘no opinion.’

From these numbers it’s clear that 2.5 times as many Asian voters ‘like’ RCV at least a little as ‘dislike’ it. In addition, it’s perfectly clear that David Lee already had this information that he could have shared with you in the media. Instead, he lumped his categories and drew his bar graphs to misrepresent to you, the media, what his own data says.

2. Not only is CAVEC’s methodology false, but the conclusions based on that methodology also are false. The conclusions in CAVEC’s study are not supported by CAVEC’s data. On page four of CAVEC’s handouts, David Lee has concluded that “some voters like it [RCV] (whites), others hate it (minorities).” But there is no data in this study whatsoever to support the conclusion that minorities “hate” RCV. Quite the opposite. As I wrote above, CAVEC’s numbers show that that 2.5 times as many Asian voters ‘like’ RCV at least a little as ‘dislike’ it. For Latino voters (according to the CAVEC poll’s addendum released at 4:33 pm), 25% ‘like’ RCV, 23% ‘like it a little’ and only 10% ‘dislike.’ So over four times as many Latino voters like RCV at least a little more than dislike it. How can David Lee possibly conclude that minorities “hate” RCV?

Properly interpreted, CAVEC’s study shows that many more voters like RCV and found it easy to use in all ethnic groups than did voters who dislike it and found it difficult to use.

Furthermore, if you analyze the results by ethnic group, you find only slight differences from ethnic group to ethnic group in the percentage of people who like RCV (46% for Asian, 48% Latino, 47% Black, 45% non-white Democrats and 52% for whites). But because of the larger margin of error for smaller groups, even these minor differences might not be statistically significant. In fact, on page 2 of the CAVEC handouts (titled “Survey Methodology”), it reads at the bottom of the page “The margin of error for the CITYWIDE SAMPLE is +/- 2.2%. The margin of error for SUBGROUPS will be higher.” By subgroups they mean Asian voters, Latino voters, Black voters, etc. They did not bother to tell you what that margin of error is for these subgroups, or even how large their sample was for the subgroups. Again, this is highly questionable methodology, and places the breakdown by ethnic groups all within the margin of error. In other words, you can’t tell from their study were there is ANY difference between ethnic groups.

Thus, for the question of whether or not various ethnic groups “like” or “dislike” RCV, the data does not support a conclusion that minorities hate the system, or even that there are real differences between ethnic groups. Rather, it appears that in all ethnic groups, more voters like the system than dislike it. This would explain why on page 8 of the handouts, all ethnic groups report strong majorities that found RCV “easy” to use, including 59% of Asians, 74% of Latinos, 57% of Blacks, and 71% of Whites (there is another category on this page showing 49% of “Chinese TV/News” voters who found RCV easy to use, and 39% difficult, but I’m not sure what this category means). Overall, a huge 67% of all voters found RCV ‘easy’ to use.

3. How many rankings did voters cast? More poor methodology. David Lee claims there are “serious questions of disenfranchisement” because Chinese voters “have much higher rates of only voting for one candidate and their ballot not counting in the final tally.” Again, David’s methodology and conclusions are completely flawed. Just because a voter only ranks one candidate does not mean that voter’s ballot did not count in the final tally. For example, if a voter ranked only Lillian Sing in District 1 or only Myrna Lim in District 11, that voter’s ballot did indeed count in the final tally since both of those candidates finished in the top two in their respective districts.

In addition, because a voter ranked only one candidate is not sufficient proof of “disenfranchisement.” The voter may have consciously chosen to rank only one candidate because that voter did not support any other candidates in the race. On election day, I was poll monitoring in District 3/Chinatown, and I observed numerous voters who were instructed by poll workers that they could rank a second or third choice but declined to do so and stated “I only like one candidate.” This included Chinese voters and non-Chinese voters. I know for a fact that supporters of Aaron Peskin were ranking only Peskin on their ballot, and were doing so purposefully.

So CAVEC’s data is not revealing of much, since he does not have the numbers to figure out if voters ranked only one candidate out of choice or confusion. And his data certainly does not support the conclusion of “disenfranchisement.”

Because all the absentee and provisional ballots have not yet been processed, it’s premature to reach conclusions about number of rankings and rates of exhausted ballots. Preliminary analysis of District 1, however, shows that many voters effectively used 2nd and 3rd choices. The question to be determined is not whether a person ranked only one candidate; the real question is whether their vote counted in the final round, meaning their vote was “effective.” Here are preliminary numbers based on the ballot image data set released by the Department of Elections calculated by our Center for Voting and Democracy for the rates of effective voting in District 1:

Table 1. Preliminary analysis of District 1 votes and rankings, based on data released by Dept of Elections on November 9, 2004

1st Pass







































What this table means is that 100% of McGoldrick and Sing voters had an effective vote since those two candidates finished in the top two, so all their voters had their votes counted in the final tally. 72.7 percent of the supporters of Matt Tuchow gave their 2nd or 3rd ranking to either McGoldrick or Sing when Tuchow was eliminated, and so their ballots were “effective” and counted in the final round. 66.1% of the supporters of Rose Tsai gave their 2nd or 3rd ranking to either McGoldrick or Sing when Tsai was eliminated, and so their ballots were “effective” and counted in the final round. As you can see from the table, of the four strongest eliminated candidates, supporters of Rose Tsai had a high level of effective votes, meaning they effectively used 2nd and 3rd choices more often than supporters of Dawydiak and Heller. If in fact Chinese voters had a higher level of “bullet voting” than other voters, it’s not apparent from this data, or it means those bullet voters selected strategically for Sing or McGoldrick and didn’t have a reason to list a 2nd or 3rd choice.

More of this kind of rigorous analysis will be forthcoming over the next several weeks from reputable organizations like the Public Research Institute at San Francisco State University, which has been contracted by the City and County of San Francisco to do an extensive study, including an exit poll. I would encourage those who are interested in these matters to be patient until the data and analysis can be made available by a reputable organization.

Conclusion: The exit poll and analysis by David Lee/CAVEC suffers from numerous methodological flaws, insupportable conclusions, and unsubstantiated margins of error that make their interpretation of this data wildly off the mark. In addition, one cannot help but conclude that David purposefully attempted to spin his data so as to mislead the media into making false conclusions.

Certainly this does not mean there was not some confusion among voters, including Chinese and Asian voters, on election day, or room for improvement. Credible studies and analyses will be helpful in pinpointing where future education efforts should be directed. Unsurprisingly, the very first election using ranked choice voting in San Francisco leaves room for improvement and refinement. Hopefully people like David Lee and his organization, the Chinese American Voter Education Committee, will choose to assist with voter education and outreach to his community. I know from personal experience that, in the past, David rejected repeated requests by me and others for his assistance in this area. From what anybody can tell, David Lee and his organization did not make much effort to educate the voters from his community about RCV.

Nevertheless, by any objective measurement, the first RCV election in San Francisco was a success. Voters ranked their candidates, winners in all seven races were declared within 72 hours after the polls closed (in three races winners were declared on election night), winning candidates won with more votes than the winners in the December 2000 runoff election, and the results and margins of victory have been recognized by nearly all observers as legitimate and substantial. San Francisco voters, poll workers, and especially the Department of Elections deserve a big congratulations.

And more:

Poll Shows San Francisco Voters Like Ranked Choice Voting

and Find It Easy to Use

CAVEC Exit Poll Survey Shows Positive Views that

Cross All Racial and Ethnic Lines

The Center for Voting and Democracy,

For Immediate Release: November 11, 2004

San Francisco voters had a very positive introduction to their new ranked-choice voting (RCV) system, according to an exit poll survey of 2,108 San Francisco voters released this week by the Chinese American Voter Education Committee (CAVEC). Those respondents expressing an opinion about the system overwhelmingly expressed support for it. Even larger majorities found that the system was not difficult to use.

69% of those surveyed in CAVEC’s survey expressed an opinion about RCV. Of these voters, fully 71% indicated they liked RCV, with most indicating they liked RCV “a lot.” This support crossed all racial and ethnic lines:

· 83% of Latinos liked RCV

· 70% of whites liked RCV

· 72% of Asians liked RCV

· 63% of blacks liked RCV

Despite the first RCV taking place in a year with high voter turnout where most media attention was focused on the federal elections, only 18% of voters found the new system difficult to use. In every racial and ethnic group a majority of voters indicated the system was easy for them.

· Overall, 67% of voters found it easy to use, compared to only 18% who found it difficult. (The rest did not express an opinion.)

· 74% of Latinos found it easy to use, compared to only 14% who found it difficult.

· 71% of whites found it was easy to use, compared to 13% who found it difficult.

· 57% of blacks found it easy to use, compared to 35% who found it difficult.

· 59% of Asian Americans found it easy to use, compared to 27% who found it difficult.

Actual voting results based on ballot data released by the San Francisco Department of Elections and analyzed by the Center for Voting and Democracy indicate that most voters in the supervisor races made good use of their rankings. In the hotly contested District 1 race, for example, voters on average cast 2.52 rankings. The number of rankings cast by voters did not vary appreciably among supporters of different candidates, ranging from a low of 2.41 for McGoldrick to a high of 2.69 for Tuchow. Supporters of Asian candidate Lillian Sing ranked an average of 2.56 candidates, showing her supporters made good use of their rankings.

Researchers at the Public Research Institute at San Francisco State University are currently analyzing results from their exit poll about attitudes toward ranked choice voting. This study, along with final election results, will shed more light on the effectiveness of voter education efforts and how to most effectively direct future educational efforts.

RCV Opinions from CAVEC survey, as published in the SF Examiner, November 11, 2004

Opinions on ranked-choice voting, by ethnic breakdown

All voters:

Like a lot — 30 percent

Like a little — 19 percent

Dislike — 20 percent

No opinion — 31 percent

‘Easy’ to use — 67 percent


Like a lot — 33 percent

Like a little — 19 percent

Dislike — 22 percent

No opinion — 26 percent

‘Easy’ to use –71 percent


Like a lot — 25 percent

Like a little — 12 percent

Dislike — 22 percent

No opinion — 40 percent

‘Easy’ to use — 57 percent


Like a lot — 25 percent

Like a little — 23 percent

Dislike — 10 percent

No opinion — 42 percent

‘Easy’ to use — 74 percent


Like a lot — 26 percent

Like a little — 20 percent

Dislike — 18 percent

No opinion — 36 percent

‘Easy’ to use — 59 percent


‘Easy’ to use — 49 percent

Source: Exit poll of 2,108 voters citywide by Fairbank, Maslin, Maullin

& Associates conducted for CAVEC.

Isn’t it wonderful how one can use numbers?

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s