SWS President Mahar Mangahas sounded peeved (or impatient or irritated) when he said that the latest SWS survey showing the margin between Aquino and Villar at 2% should not be reported as a statistical tie. This was in a phonepatched interview by ANC (ABS-CBN News Channel).
Since this is their survey, maybe they can decide how it should be reported. Note however that it is the media partner of SWS, Businessworld, that first reported it as a statistical tie, and it is linked in the SWS website, as follows:
From the SWS website:
A very tight race
The national status of the three races for President, Vice-President, and the Senate from the BusinessWorld-SWS February 24-28, 2010 Pre-Election Survey is in the March 9, 2010 issue of BusinessWorld, and at the BusinessWorld website.
BusinessWorld (BW) is the media partner of SWS in polling, for exclusive publication by BW, the public’s voting preferences for candidates for national positions (President, Vice-President and Senators) in the 2010 election. The original SWS report will be posted in the SWS website two days following the BW publication.
And from the linked news report, from its media partner Businessworld:
Top ‘presidentiables’ Aquino, Villar in statistical tie
THE RACE to Malacañang has become very tight with the top two “presidentiables” locked in a statistical tie, a Social Weather Stations (SWS) survey commissioned by BusinessWorld showed.
The poll, the third conducted by the survey research institution for this newspaper, found the front-runners — senators Benigno Simeon “Noynoy” C. Aquino III and Manuel “Manny” B. Villar, Jr. — both losing ground among voters with little over two months to go before the May 10 national elections.
The cost, however, was more substantial for the Liberal Party’s Mr. Aquino, who saw his overall score cut by a substantial six points to 36%, compared to just a one percentage-point drop to 34% for the Nacionalista Party’s Mr. Villar.
Mangahas said that any news report should just give the figure, and state: Aquino leads Villar by 2%; he said that this was the more neutral statement and that any report stressing the statistical tie was leaning in favor of one candidate.
Note again that in the past, here was how the SWS reported the one-per cent margin between candidates, (at that time referred to as “recommended successors” in the survey) ( posted it in my blog before ):
7 November 2008. Third Quarter 2008 Social Weather Survey:
Noli De Castro, Manny Villar, and Loren Legarda are people’s top recommendees for 2010
Social Weather Stations
Vice-President Noli De Castro, Senate President Manny Villar, and Senator Loren Legarda are the top recommended successors to President Gloria Macapagal Arroyo in 2010, according to the Third Quarter 2008 Social Weather Survey, fielded over September 24-27, 2008.
xxx
(And in the second to the last paragraph: )
The Third Quarter of 2008 Social Weather Survey was conducted over September 24-27, 2008 using face-to-face interviews of 1,500 adults in Metro Manila, the Balance of Luzon, Visayas, and Mindanao (sampling error margins of ±2.5% for national percentages, ±6% for Metro Manila, Visayas and Mindanao, and ±4% for Balance Luzon). The area estimates were weighted by National Statistics Office medium-population projections for 2008 to obtain the national estimates.
In this report, SWS did not use the term “statistical tie”, but neither did it report the survey as “De Castro leads Villar by one per cent while Villar leads Legarda by one per cent”.
Rather, it enumerated the names of the three, then stated that they were the top recommended successors in the survey. It didn’t say, “tight race” or “narrow margin” or “leading by one per cent”, it said that the three were the top choices – that kind of reporting leans towards saying that they were practically in the same rank – a statistical tie.
Okay. Since the SWS President did not want the term “statistical tie” used in the news reports, maybe we can use as pattern its previous media release, like: enumerate the names then state they “ are the top candidates in the SWS Survey…etc”. That was how the SWS reported the three-way statistical tie in that survey.
Survey results should always be reported with context. Journalism requires context.
You can report “leading by 2 per cent” then give the margin of error, etc., time frame, etc.
Some news reporters and editors however want to spell out the meaning, so they report it as a statistical tie then give the figures and margin of error.
Statistical tie simply means there is a narrow lead that is within the margin of error (usually between 2.2% and 2.5%). It doesn’t mean they are in an absolute tie, it just means that, because of the margin of error, there is a probability that they could in reality be tied but it is just a probability.
So, the news reporters and the editors just made a short-cut of it and reported that it was a statistical tie.
Well, you can say one leads the other and the lead is within the margin of error.
Happy?
ANC later reported it as “Aquino and Villar in neck-and-neck race.” And later reported it as “Aquino keeps top spot” (guess towards whom this headline is leaning), with the lead paragraph stating that they were in a neck-and-neck race; this kind of report avoids mentioning Villar’s name in the head.
Just say it’s a tight race. You can use an active verb if you want, Aquino and Villar running in neck-and-neck race, you can either mention both in the head or not mention either, and all because the news source doesn’t want you to use the term “statistical tie”.
