Forecast of the Dynamics of the Undecided Population in a Public Opinion Poll by a Neural Network

Abstract

Excerpt

One of the unsettled issues in any survey concerns the handling of the undecided respondents (UR). This problem can be safely ignored if UR constitute a small percentage of the population, and if their number is not enough to substantially sway the outcome of the survey. However, if the UR for example is about 30% and the decided respondents (DR) splitting 40% : 30%, then the results of the survey have to be taken as inconclusive.

The conventional way of dealing with this problem, as shown in print and television allover the Philippines, is simply to split the UR in proportion to the DR. Basically such undertaking merely 'erases' the possible significance of the UR. Another approach is to allocate the UR among the different responses, based on such criteria as geographical distributions (e.g., residents of a given area have traditionally gone with candidate X), or socioeconomic distributions (because, perhaps, affluent voters tend to go Republican in the U .S.)(Visser, 1996). Such may be termed an exogenous method, inasmuch as it relies on considerations other than the data at hand.

Still another method is to use statistical analysis tools, such as discriminant analysis (DA), which tries to look at patterns in statistical distributions to forecast the most probable separation of the undecided respondents. This has been used in the past, with claims of up to 86% success rate (Fenwick, 1982). Such a method we would term endogenous since it looks only at the given data and infer its conclusions only therefrom, with minimal or no heuristics based on outside considerations, beyond what may be found in the data itself.

In this study commissioned by Pulse Asia, Inc. (PA), a professional public opinion poll organization, we demonstrate that an artificial neural network (ANN) can forecast up to 94% accuracy the most apparent sentiment of the UR when he chose to decide, or when he is forced to decide. Specifically, an ANN is tasked to determine how the electorate rates the performance of Joseph Estrada, who was the incumbent president of the Philippines. The poll questionnaire consists of one direct question and 291 (pro-rated) indirect ones that probe into the opinions of the respondent on specific socioeconomic and political issues, and ratings of other government executives, legislators, and institutions. The answers to the direct question can be divided in three parts: one, those who approved of the Estrada administration, second those who disapproved, and third, those who are undecided. A total of 1200 respondents were chosen randomly allover the Philippines with a claim that the sampling procedure constitutes an uncertainty of ± 6% and ± 3% in the national and regional level, respectively (Pulse Asia, 2000).