One of the disciplines that have helped Market Research the most is Statistics. Knowing the characteristics of a population always represents a challenge since it is usually impossible for all companies to carry out a census that manages to know all the individuals of a universe at a given time. Sometimes, not even the governments themselves can afford such a company and that is why the Population and Housing Censuses are usually carried out every ten years and Surveys are used to answer questions on more specific topics.
Designing a survey that represents a given population in the best possible way is one of the jobs that requires the most care within Market Research. If it is not done properly, there will be no adjustment or calculation worth correcting it after the field work is finished. And it is that in addition to guaranteeing that all the participants have the same probability of being chosen to answer our questions, a proportional reflection of the total number of people we seek to know in detail must also be achieved.
When talking about this type of study we are talking about quantitative research, since in its qualitative counterpart what matters most to us is the individual or a small group of them. Assuming that we are talking about populations that tend to infinity for statistical purposes (of the order of more than a million people in general) we can say that a random sample should not have less than 1,500 participants if we want our data to have 2.5 more minus percent margin of error with a 95% confidence interval.
The margin of error indicates the degree of confidence that we can have in the data of a study. The smaller it is, the more reliable the results of an investigation will be, since they refer to very small errors. In general, the results of a sample with percentages equal to or less than the margin of error should be rejected due to the uncertainty that they are part or not of the characterization of the population to which they refer.
Regarding the confidence interval, which we have talked about in previous posts on this blog, we can indicate that it is the probability that our sample coincides or does not coincide with reality at a given moment. In other words, if we talk about a 95% confidence interval, we can say that out of 20 times we design or take a sample from a population, only once will it not correspond to the universe we seek to know.
It is easy to see that the smaller our sample, the greater the margin of error and therefore our results will be less reliable for understanding the universe and making decisions based on it. Therefore, when seeking to optimize time and costs, it is always worth taking into account the almost direct proportional relationship between the size of our sample and the margin of error to consider other alternatives such as increasing the number of interviewers or the use of technologies that reduce capture and analysis times rather than reducing sample sizes.
At Acertiva we have been conducting quantitative studies in Latin America for almost two decades. You can write to us today to tell us about your Market Research needs in our region. We are sure that we can be your regional allies in the knowledge of the Latin American consumer.