SWS’ mahar mangahas answers richard gordon
it’s hard to believe richard gordon do not have an appreciation of surveys and he is that ignorant of it. he has been a politician for a long time and his previous job at the private sector (procter & gamble) must have given him a good dose of knowledge on it.
but he has taken SWS and Pulse Asia to court, so we can assume he does not know a lot of things. mahar mangahas of SWS gives him an answer, for his knowledge:
Ignorant columnists may not be TRO’d
by Mahar Mangahas
Philippine Daily Inquirer
First Posted 23:34:00 04/30/2010
ABOUT COLUMNISTS. EVERY THREE YEARS, IN the Philippines, comes a silly season when some columnists excel in misinforming the public about survey science. These are the ones who assert, for instance, that “a survey of only one or two thousand respondents cannot possibly represent many millions of voters.” Despite repeated demonstrations that a properly conducted sample survey is indeed representative of the whole population, they will not accept it, and would rather bask in their ignorance.
Now, is there a way for a columnist who propagates falsehoods to be legally suppressed? For instance, may professional statisticians petition a court for a Temporary Restraining Order (TRO) to stop a columnist from vilifying statistical research about voters’ preferences? The answer to this is NO.
A columnist has a constitutional right to display his ignorance, without prior restraint. This is because the right of free expression is a preferred right—“prior restraint” and “preferred right” being legal phrases.
The statisticians’ legal remedy for a columnist’s abuse of free speech would be to sue for damages afterwards. But they don’t bother, because a statistically-challenged columnist doesn’t fool the people who really count.
About Gordon’s complaint. Last week, Sen. Richard “Dick” Gordon asked a Regional Trial Court to issue a TRO against Social Weather Stations and another survey entity, to desist from conducting and publishing their election surveys, which he called “false, inaccurate and flawed,” causing him “grave and irreparable injury.”
Gordon’s suit is ridiculously sloppy. Above all, it is ignorant of the Supreme Court’s affirmation that election surveys are constitutionally protected (see my April 17 column). In SWS v. Comelec (G.R. 147571, May 5, 2001), the Court nullified the section of the 2001 Fair Election Act that attempted to ban publication of election surveys. It ruled that such a ban “imposes a prior restraint on the freedom of expression” and forms “a direct and total suppression of a category of expression” during the elections.
Gordon claims that “surveys issued by the defendants … showed him only at the 29th spot,” and yet he won as senator in 2004. Actually, the SWS surveys of the 2004 senatorial race had him as 14th in Jan. 18-22, tied for 16th on Feb. 17-25, 14th on March 21-29, tied for 8th on April 10-17, and tied for 9th (with 29 percent of the vote) on May 1-4. It looks like 29 percent was misread as 29th place. Thus he was already in the winning circle in the last two SWS pre-election surveys.
Gordon’s complaints about methodology are false. (1) My column of March 6, 2010 reported that SWS received two awards from the Gallup World Poll for excellence in field methodology, among all of Gallup’s Asian field providers. (2) Face-to-face interviewing, which we always do, and which Gordon thinks “outmoded,” is part of Gallup’s job order to SWS. (3) We agree with Gordon that sampling should be done by probability, and not by quota. Apparently he doesn’t know that SWS always does the former, and never does the latter.
Gordon calls it “highly improbable” that SWS did two national surveys over as short a period as March 19-30, 2010. Actually, SWS did eight national surveys, not all about elections, over January-April 2010, plus several local surveys.
Gordon alleges that, last April 14, an unidentified SWS pollster in Cebu asked a respondent to choose between only two presidential candidates, instead of among 10. Comments: (1) SWS had no election survey in Cebu on that date; (2) all SWS interviewers have ID cards—tell us her name so that we can check; (3) the published SWS election surveys always feature the 10 candidates; (4) in any case, it is legitimate for anyone to inquire how a voter would choose between two candidates.
Gordon’s claim that SWS fails to disclose its sponsors is false. Check the website, http://www.sws. org.ph. The SWS Survey Data Library is open to the public. Its staff helps visitors, short of serving as research assistants. The library fee is affordable even to students. Users should come personally, and not expect their technical questions to be answered by mail.
Gordon’s citations of survey errors in past elections are very few; they are the exceptions that prove the rule, like the failure of US pollsters to predict Truman’s win over Dewey in 1948, which he cites as though it was SWS’ fault too. My 2009 paper, “The challenge of election surveys in the Philippines,” summarizes our election survey record; see our website. The error of the 2004 exit poll in Metro Manila was investigated by an independent group of scientists, and no fraudulence was found; see their report on the website.
Gordon’s claim that “there are no associations of professional pollsters and polling firms which regulate, control, and sanction defendants … for their violation of the code of professional ethics …” is false. Seems he hasn’t heard of the Marketing and Opinion Research Society of the Philippines (MORES), founded in 1977. Both MORES and the World Association for Public Opinion Research (WAPOR), to which key SWS staff members belong, have Codes of Ethics. Last Wednesday, the MORES board of directors issued a press statement denouncing Gordon’s petition for striking at the heart of our democratic process.
If the SWS election surveys were not true, accurate, and best-quality, I wonder if Gordon would still be interested in a TRO. Maybe he would just grant us the same freedom of speech that we allow to ignorant columnists.