Statistical inference is used to summarize data from the sample for the entire population. In fact, the sample is never a 100% population model, but only its more or less distorted version. Statistical inference is used to evaluate such distortions and, therefore, to draw more accurate conclusions about the general population. It is important to remember that achieving a 100% guarantee that the study results are characteristic of the general population is possible only when continuous research is conducted. For example, the survey of all representatives of the general population. In this case, the electorate is not limited only to students, so Sara’s conclusions regarding the election results cannot be considered correct.
Most statisticians claim the minimum size of samples for obtaining meaningful results should be at least 100. Nevertheless, the surveyed population group may be less than 100. In such cases, it is appropriate to study the total size of them. Statisticians use the term “good maximum”: it means that the surveyed population sample size is about 10% of the overall available quantity (“How to choose,” n.d.). For instance, for the population of 3000, 10% would be 300, whereas, in a population of 10,000, it would be 1,000.
The number of samples closer to the minimum is chosen in specific conditions. It is appropriate when the result’s rough estimate is needed. It is also applied when the sample is not planned to be divided into diverse groups during the survey. Additionally, a “good maximum” is relevant when it is supposed that most people are likely to answer similarly, or the conclusions of the study will not have any significant consequences. In the situation with the president’s elections on the state level, the potential electorate is far broader than the students.
Sara’s simple explanation might be as follows: she might be proposing to imagine that she wants to survey her friends at a high school. The chosen group enrolls fifty students, so she needs to scan all of them. Still, the obtained results would give her a rough idea of potential election results (“Statistical inference,” n.d.). If she wants to find valid perspectives, the surveyed group should be at least 100 people, including not only her friends but also the representatives of the whole high school’s staff. Hence, it would be appropriate to engage the students, teachers, ordinary school workers, parents, friends, and other available electorate groups.
References
How to choose a sample size (for the statistically challenged). (n.d.). Tools4dev. 2020, Web.
Statistical inference and estimation. (n.d.). Stat 504. 2020, Web.