Understanding opinion surveys
POLL resultsarenot to betaken lightly
or readily accepted at face value,
‘This is a reminder on considering
results of opinion polls, said Carole
Sarthou, managing director of Syno-
vate Inc., inthe recently held Market-
ing and Opinion Research Society of
the Philippines” (Moves) Fourth Gen-
eral Membership Meeting,
Sarthou joined Butch Africa, vice
president of the Philippine Statistical
Association (PSA); and Terry Solivas,
amember of the board of directors of
PSA, in a discussion on what to look
for when presented with and report-
ing the result of opinion polls
The discussion is in line with a
project currently being conducted by
PSA-Mores, titled "Evaluating Opinion
Polls.” The objective of PSA-Mores is to
raise the consciousness of pollsters, the
mediaand the publicaboutresponsible
polling and reporting of results.
“People should ask relevant ques
tions pertaining to the gathering of
data presented in each opinion poll,”
added Sarthou.
Solivas recommended that for ev-
ery opinion survey that publishers
comeacross, theyshould firstlook for
potential biases in the poll results.
Thiscan begained, hesaid, through
careful scrutiny of the data and ask-
ing relevant questions such as: Who
conducted thepoll and whe sponsored
it? [tisalsoimportantto knowdetails
about the data collection, the ques-
tions that were asked, and how the
data were analyzed.
Inaddition, itis also significant to
know the profile of the sample that
was used in the opinion poll. This is
to ensure that it is aligned with the
profile of the target population that
it was meant to represent, inorder to
avoid any skews or aver- or underrep-
resentationof particular subsegments
of the population.
Knowing the types of questions
asked would put things in perspective
foranyonewhowants tointerpret the
¥esults, The timeframe when the poll
was conducted also helps contextual-
ize the results.
Moreover, according to Africa, itis
imperative that, other than checking,
for the details concerning the gather-
ing of data for the opinion poll, pub-
lishers and readers shouldcheckif the
statistical organization that released
the results is a trustworthy one,
“if the pollster is not willing to
give all of the data that has been pre-
sented here, then maybe you should
just forget aboutit,” said Africa, “It is
nice to know, but you shouldn't lose
sleep about it.”
UP profs question veracity
of poll surveys
‘TWO professors of the University of
the Philippines, meanwhile, assailed
as “full of inconsistencies” news re-
portson recentelectoralsurveys inan
apparent attempt to unduly influence
public perceptions in favor of certain
presidential candidates.
“The professors noted that news re-
portsonwhat wasflauntedasthelatest
poll showing the sudden surge of ppro-
administration presidential candidate
Gilbert Teodoro were “disturbingly in-
consistent with previous surveys.”
Prospero de Vera Ill, professor of
UP’s College of Public Administration
and Governance, said the wide dis-
crepancies in the findings of recent
Social Weather Stations, Pulse Asia
and the Advocacy Center need to be
publicly explained, otherwise it will
put in question all surveys leading to
the May 2010 election.
A published report on a survey
purportedly conducted by a so-called
Advocacy Centerconducted the“poll”
among 1,200 respondents showing a
decrease in the lead of presidential
candidate Sen. Benigno Aquino II!
over Sen, Manny Villar and a surge
in the support for Teodoro.
Forher part, Roxanne Girlie Cipria-
no of the Department of Communica-
tion Research of the UP College of Mass
Communication noted that the survey
reportis“full ofinconsistencies"—and
thus “the survey is not credible.”
jpriano noted that even the ter-
minologies used by theallegedsurvey
group “do not fit the time-honored
lingo and practice of the science of
research.” She said there is no such
thing as a “stationary dive,” “signifi-
cantly dropped,” or “pro-rated.”
“The wrong.use of such terminolo-
gies raises serious questions about
the professional competence of the
so-called research group.”
DeVera, whohasconsistentlyealled
for public presentations of survey
groups, said: “Itistimethatwedemand
more accountability from the groups
behind these surveys, including their
owners, directors and analysts, and
who pays for their surveys.”