Based on my personal knowledge in statistics, surveys, if done properly, can be a powerful tool in predicting the overall “mood” of the society. How was that possible? We’ll, surveys take a small part or sample of the entire population, examine its properties using statistical techniques and assumes that it represents the entire population.
In case of election surveys, the survey firm just interview few people, say 1000 persons. The specific people to be interviewed were pre- determined using random sampling techniques. The sampling method also considers the distribution of the population. For example, they may take more samples in places with more registered voters than those with little. Then, the survey team goes to the actual houses of these persons and asks them regarding whom they may likely to vote in the coming election.
Does the result of the survey determine the actual result of the election?
It depends on how near the surveys are done with respect to the date of the actual election. Of course, people’s mind is always changing. The society is dynamic; it keeps on changing from time to time depending on many factors. Surveys done a few days before the Election Day though have proved to be more or less than 99% accurate in predicting the actual outcome. But we should always keep in mind that surveys have a margin of error. That’s the nature of survey since it only relies on statistical tools to “infer” the behavior or “vote” of the entire population.
Pre-conditioning the minds of the people?
Many election candidates (especially those who don’t have good survey results accuse survey firms of “conditioning” the mind of the people. It can be true but with regards to Pulse Asia and SWS survey accuracy, I can confidently say that the results can be trusted. That’s the very reason why their surveys are always welcomed with much buzz.