 |
 |
|||
 |
|
|||
| |
|
|||
|
|
|||
Statistically Speaking
When Polls Rush In, Losers Fear To Read! 
by Dr. Romulo A. Virola 1
Secretary General, NSCB
Should election polls be regulated or banned?
With barely over a month before the May 14 elections, calls for some form of action against election pollsters have resurfaced. They are not transparent; they are commissioned by vested interests; they are used as a trending tool to condition and influence the minds of the electorate; they do not ask the right questions; etc. etc. the critics of election surveys would argue! Mostly, the losing candidates in those surveys, of course! However, one should not discount the possibility that these politicians, actually believe the survey results, except that it would be politically incorrect for them to admit so in public!
In the 2004 elections, the SWS and Pulse Asia did very well in predicting the results of the national elections for president, vice-president and senators. The last survey SWS conducted before the 2004 elections successfully predicted all the winners while Pulse Asia and the SWS exit polls correctly predicted all but one of the winners. Statistically Speaking gave them ratings2 of 90%, 90% and 82%, respectively. However, the SWS exit polls for NCR revealed some weaknesses in the design of the survey which I hope they have corrected. Also, I wish the two firms would hire different teams for their field operations to enhance the independence and credibility of their results!
Are voters influenced by these surveys? Maybe not, maybe yes! But the appropriate question really should be - Is it wrong to allow the conduct of such surveys if they can influence the decisions of voters? The purpose of legitimate surveys is to provide information that can be useful in decision-making. They are not conducted to influence the results of the elections. If voters vote for those who lead in the surveys, is that sufficient reason to ban the surveys? Just as there are those of us of who are risk-takers, while others are risk-averse; just as there are those of us who bet on the dejado while others bet on a llamado like Manny Pacquiao against Solis, shouldn’t the voters be allowed to vote any way they want? After all, isn’t democracy about informed and free choices, even if some of the choices are obviously incorrect?
If indeed, voters are influenced by the results of these polls, one can view them as an application of the Delphi Method, a spin-off of a U.S. Air Force-Rand Corporation defense research in the 1950s called “Project Delphi”. The study sought the most reliable consensus of opinion of a group of experts on the selection, from the point of view of a Soviet strategic planner, of an optimal U.S. strategy that estimates the “number of A-bombs required to reduce the munitions output by a prescribed amount”3. Certainly, no way could we say that all voters are experts, but no one could say that the Delphi Method is inappropriate either!
Three years ago, Statistically Speaking wrote about this issue. Essentially, we said that scientifically-conducted election poll surveys gave us a feel of the pulse of the voters and could and should be used by politicians to inform their decisions. We also averred that in a democratic society like ours, it was wrong to ban these surveys, even as we listed countries where they are banned and countries where such bans had been declared unconstitutional. Earlier, on 5 May 2001, the Supreme Court, bless its statistically-literate judges, declared4 the unconstitutionality of § 5.4 of R.A. No. 9006 and § 24(h) of COMELEC Resolution 3636, dated March 1, 2001 which stated that “ Surveys affecting national candidates shall not be published fifteen (15) days before an election and surveys affecting local candidates shall not be published seven (7) days before an election”. Doesn’t that say a whole lot on the quality of our legislation? And know what? On 30 November 2006, the COMELEC issued en banc Resolution 7767, § 32 of which states the very same provision previously declared by the Supreme Court to be unconstitutional! Ano ba yan?
Honestly, though, I think that all the genuine candidates and their campaign managers actually believe the SWS and Pulse Asia surveys. The published results probably cause them discomfort but it is the inconvenient truth! If they act accordingly, they can use statistics to their advantage. If they refuse to accept the stories told by those statistics, they only have themselves to blame for losses and failures they may have been able to minimize if not outright avoid. Secretly, they are probably doing something to re-shape the stories told by the surveys. Legally, one hopes!
Yes, the science of sample surveys shows the power of statistics! The lies and damned lies referred to by Disraeli are not the statistics done by professional statisticians. They are the ones peddled by vested interests and we have to be wary of them! To help the public discern the difference, therefore, I hope the Philippine Statistical Association (PSA) will heed my previous call for the PSA to take on the responsibility of accrediting the legitimate pollsters. However, noting that SWS is an institutional member of the PSA, this has to be handled most carefully! And after the results of the May 14 elections are out, Statistically Speaking will assess anew the performance of these surveys! Just to keep them on their toes!
Surveys aside, wish you had not been April Fooled; more importantly, wish the Holy Week had cleansed our souls and ennobled the passion of our minds! Politics and politicians, notwithstanding! Happy Easter to all!
Reactions and views are welcome thru email to the author at ra.virola@nscb.gov.ph.
_______________
1 Secretary General of the National Statistical Coordination Board (NSCB) and Chairman of the Statistical Research and Training Center (SRTC). He holds a Ph. D. in Statistics from the University of Michigan in Ann Arbor, U.S.A. and has taught mathematics and statistics at the University of the Philippines. He is also a past president of the Philippine Statistical Association. The author thanks Noel S. Nepomuceno and Candido J. Astrologo for the assistance in the preparation of the article. He also thanks the data sources, particularly the Social Weather Station and Pulse Asia.
2 The ratings took into consideration not only the success in predicting the winning candidates but also the closeness of the predicted versus actual number of votes.
3 The Delphi Method, Techniques and Applications. Edited by Harold A. Linstone and Murray Turrof. http://www.is.njit.edu/pubs/delphibook/
4 The decision was very close: Davide, Jr., Vitug, Gonzaga-Reyes, Melo, Puno and Panganiban concurred while Kapunan, Belosillo, Pardo, Ynares-Santiago and Sandoval-Gutierrez dissented and Quisumbing, Buena and De Leon, Jr. were on leave.
Posted 10 April 2007.
 |
  |
 |
||||||
  E-mail the webmaster |
Terms of Use |
Home • Top of Page | ||||||
1997-2008,
National Statistical Coordination Board |
||||||||
On Election Polls: Part