Skip to content
post thumbnail

Stats experts caution public on election surveys

THREE of the country’s leading statistics experts have urged institutions conducting electoral surveys to open their methods to public scrutiny and peer review. Former National Statistics Office Administrator Tomas P. Africa, former Philippine Statistical Association president Jose Ramon G. Albert and PSA life member Corinne B. Burgos also cautioned the public against results from surveys

By verafiles

Mar 27, 2010

-minute read

Share This Article

:

THREE of the country’s leading statistics experts have urged institutions conducting electoral surveys to open their methods to public scrutiny and peer review.

Former National Statistics Office Administrator Tomas P. Africa, former Philippine Statistical Association president Jose Ramon G. Albert and PSA life member Corinne B. Burgos also cautioned the public against results from surveys that are done for self-serving interests or disregard sound statistical methodologies.

They also called on the media to exercise care in reporting survey results, including changes in the results over time.

Africa and Albert and Burgos issued the statement in reaction to VERA Files‘ report, “Gibo surges in survey by poll firm owned by Arroyo allies.”  Their statement:

A Call for Credible Sources and for Proper Analyses of Data

The citizenry in a democracy depends on reliable sources and sound analyses of data to guide them in making decisions.  While more and more data are being produced, archived, and made available for analysis, we, the undersigned, are concerned that our society has become data rich but information poor.

During election periods, especially, results of sample surveys conducted by various poll outfits are reported out to get a snapshot of people’s voting preferences.  But some of these firms may not be managed by professionally trained statisticians, and yet, media provides them space and people tend to treat their survey data as gospel truth.

  • We call on persons and institutions engaged in the production, analysis and dissemination of data to make their practices and instruments available to public scrutiny in order to ensure that their methods and practices are and remain ethical and produce sound and valid results.   They should remain open to peer review by a neutral body of statistics experts, such as the Philippine Statistical Association.
  • We caution the public about results from poll surveys, especially those that may be done for self-serving interests, and/or with disregard for sound statistical methodologies.  Even with sound survey practices, the analyses of data, especially those on changes of survey results across time, should take into account error margins which embody the limitations of precision and accuracy generated from these data sources.  Comparisons of surveys[1] conducted by different institutions are not as straightforward as the public may be led to believe.
  • We also urge the media to be more circumspect in reporting poll survey results, including changes in the results over time.  The quality of data series is reflected in the reliability of the institutions generating these data.

H. G. Wells (1866-1946) asserted that, “Statistical thinking will one day be as necessary for efficient citizenship as the ability to read and write.” We are concerned that such a day has not yet come in the Philippines, but we know that this can still be remedied with, if the citizenry becomes more statistically literate.

TOMAS P. AFRICA[2]

JOSE RAMON G. ALBERT[3]

CORINNE B. BURGOS[4]

[1] One example would be adding up the poll results of each presidential candidate obtained by four different polling firms and then getting the average to mislead the public, irrespective of widely differing sampling designs, sample sizes, questions, and manner of data capture.

[2] Former Director (2001-2006), UN Statistical Institute for Asia and the Pacific; Former Administrator (1989-2001), National Statistics Office; M.S. in Social and Economic Statistics, George Washington University; Former President (1988-1989) Philippine Statistical Association. Email: tomas.p.africa@gmail.com

[3] Senior Research Fellow, Philippine Institute for Development Studies; Ph.D. in Statistics, State University of New York @ Stony Brook; Former President (2004-2005) Philippine Statistical Association. Email: jrgalbert@gmail.com

[4] Statistical Modeling Head, Globe Telecom Inc; PhD. In Statistics, University of the Philippine at Los Banos; Recipient of 2003 Jan Timbergen Award, International Statistical Institute; Life Member, Philippine Statistical Association.  Email: silver2tone@gmail.com, ducati.mamaw@gmail.com

Get VERAfied

Receive fresh perspectives and explainers in your inbox every Tuesday and Friday.