Finding the variance of a population gives the complete information regarding how the population varies among individuals. Many times, this is exactly what a study hopes to find or infer. By finding the variance, this question is no longer unanswered, and the goal of the study, which is often how to apply the findings, can be expedited. In addition, finding the variance of a population means that there will be no error in predictions using formulas that include standard deviation statistics.
Future studies can benefit greatly from previous studies that analyzed the same population and simultaneously found that population's variance. This is because even if the population has changed since the first study, as in the case of many types of data (e.g., organisms that die and are born), the variance of a population rarely changes to a large degree. This means that the original variance is an accurate estimate of the current variance. This estimate is likely a better estimate than the standard deviation that corresponds to a new sample. Thus, the finding of variance allows future researchers to conserve their resources.
Finding the variance of a population is no easy task. If the population of interest is large, then finding the variance will require much time and, as is often the case, money. Because most research is funded by an academic institution or external organization, the researchers who intend to find the variance of a population must justify the use of resources to the sponsor of the study. This is not an easy task, because the larger a population is, the more infeasible such a study will be. Often, the resources that can be spent finding a population's variance can be put to more efficient use.
Most applied statistical studies do not actually find the variance of a population. Instead, they often use standard deviation statistics as replacements. Because of this convention, most of the statistical world does not expect variance to play a role in statistical analyses. One outcome related to this fact is that most statistical computer software does not take in variance as input.