A View of Survey Results: the 2016 Philippine Elections
Assuming the survey responses were gathered appropriately (methodologically compliant), then the estimates of prevalence can be interpreted meaningfully in terms of an electoral race, provided the percent standard error is considered. Bobbi tiglao also posits that another crucial parameter needs to be estimated and included in the calculation of voting preference: that of machinery generated votes, which should be added to the survey responses. Now since the two quantities are additive, we can analyze the survey preferences before or after the machinery generated votes are included, and the results will be the same. It is more intuitive, though, to consider it at the end when the quantities are compounded into a single measure. Let us now go to the final estimates (survey responses plus machinery generated). Source: Tiglao, see FB posts If we use the SWS declared standard error */- 6 percent, then the following conclusions are warranted: 1. there...