Introduction
One of the most challenging activities in the conducting of surveys is the determination of the sample size to use. This is normally the case since, for surveys to be useful, their target population is usually very large. For instance, for a corruption survey to be useful, it must cover an administrative unit like a province, or a state, or even a country. This paper examines the effect that a large sample has on data, and investigates whether having a larger sample will lead to a more accurate representation of the population being surveyed.
Effects of larger sample sizes on data
There are a number of data characteristics that are determined by the size of the sample. One of these characteristics is statistical power. In cases where more statistical power is needed, it is of essence to use a large sample. Large samples are associable with increased precision of data. For instance, in a situation where a researcher intends to calculate the value of an unknown parameter, a larger sample will give more accurate results (Hernandez, 2006). Thus in a study where the researcher intends to know the proportion of urban dwellers with a certain medical condition, examining 200 urban dwellers will produce more accurate results than a situation in which the researcher examines 100 urban dwellers. There are however situations in which collecting a large sample may not give more accurate results than a smaller sample. This situation is normally observed in situations where a distribution is highly skewed, or due to data dependence. From this discussion, it is evident that sample size has a direct effect on the quality of estimates made using the sample. The quality is normally determined using confidence levels and the power of tests of hypotheses.
Sample size and representation of population
As it can be expected, the use of a large sample size in a survey will lead to a more accurate representation of the population. This is because a sample is chosen so that its characteristics can be extrapolated to give the characteristics of the population (Miles, 2004). The best sample size would, therefore, be a sample size that is equal to the population size. This is practical in cases where the population size is small. In this case, the sample is equal to the population, and thus it is an accurate representation of the population. In cases where the population is very large, a bigger population size will capture more data, and thus it is more likely to represent the population in a more accurate way since the sample will cover both erroneous and accurate points in the dataset. On the other hand, a smaller sample may select data that contains a significant number of erroneous points, which will make the characteristics of the sample deviate significantly from those of the population.
Conclusion
Sample size is a very important factor in surveys. The determination of what size of sample to use is influenced by many factors. Regardless of the factors considered during the choice of sample size, the latter has great implications for the results of the survey. A large sample size has been credited with increasing the accuracy of the estimates that are made using the sample. Thus a large sample is expected to have parameters that are closer to population parameters than their counterparts derived from small samples. With this information, it is apparent that a large sample size is a more accurate representation of the population.
Reference List
Hernandez, P. (2006). The effect of sample size and species characteristics on performance of different species distribution modeling methods. Web.
Miles, J. (2004). Getting the Sample Size Right. Web.