Question

You are watching TV with a friend of yours, who was not lucky enough to go to business school and take a course on marketing research. On TV, a journalist explains that, according to the latest poll, “pollsters are confident at 95% that 64% (+ or – 3%) of the French voters would vote for Mr. Biden and 36% (+ or – 3%) for Mr. Trump, if the presidential elections were held today”. The journalist adds that those results are based on a poll conducted among 1,054 registered voters.

Your friend tells you that “It’s ridiculous to interview only 1,054 people and believe you can generalize your findings to a population of millions of people… You have to interview ALL the U.S. registered voters if you want your findings to be accurate! A census is always better that a poll based on a sample!” Based on what you have learned in this course,

(1) explain to your poor friend why she is not right.

Also, (2) tell her what pollsters mean by “margin of error” and “confidence interval”.

Answer 1

Answer 1= Sampling is a method that is used by the experts and the managers who want to estimate something or any scenario on the basis of the entire population. In sampling, the managers take the samples from the entire population that represent the characteristics and traits of the population. In sampling, it is not necessary that we consider the entire population as it is not possible and moreover it will need a lot of time, energy, resources, and effort to analyze the data. On the other hand, by having representative sampling that represents each section of the population, we can find out the same pattern that is related to the entire population. So this is the reason why sampling is used to find out the information or trend in the entire population.

Answer 2= The Margin of error can be defined as a smaller extent or amount which is taken into consideration to accommodate any mistake or error in the calculation or to accommodate any possible change in the scenario when the sampling was conducted than the scenario at present.

Confidence interval indicates that is a similar population is used to take the samples on different occasions and an internal estimation is made every time, then the resulting interval will include the actual population parameter as per the confidence interval.  

A confidence interval provides us a range of the given data that is most probably will include the population parameters of the entire population under analysis