Prof. Ana Maria Tabunda PhD, University of the Philippines statistics professors, former dean of the University of the Philippines School of Statistics and Pulse Asia Research Director, in an interview with DZMM yesterday, pointed out that in the most recent Pulse Asia survey, Poe and Binay are statistically tied (Poe at 27% and Binay at 21%). She explained further that in a survey such as this one, with a margin of error of +3/-3, for an advantage or lead to be considered statistically significant, it should be more than 6%.
That can be taken further as you go down the so-called ranking made by the news media of the most recent Pulse Asia survey. Because of the margin of error of +3 /-3, for the survey results of Poe 27%, Binay 21%, Roxas 18%, Duterte 15%, reporting the Pulse Asia survey result using a ranking of 1-2-3-4 is not as close to reality than reporting it as: a statistical tie between Poe and Binay, and a three-way statistical tie among Binay, Roxas, Duterte.
So, the “ranking”, after factoring in the margin of error is:
For 1-2: Poe, Binay in no particular order (statistical tie)
For 3-5: Binay, Roxas, Duterte in no particular order (3-way statistical tie)
i know that many people want to see a definitive ranking rather than statistical ties — but there it is: there isa margin of error estimated for the number of respondents at 1,200. Ok, if you’re really more comfortable with a definitive ranking, here it is: Poe tops Roxas and Duterte; statistically tied with Binay. Happy 🙂 ?