The core differences between adaptive sampling and network sampling rest on their applications. Adaptive sampling in an entirely new sampling design in which regions, defined in its context as “units” are selected based on the values of variables that are of interest during the observation process in the sampling survey while network sampling takes the approach of duplicating the counting of population elements by the application of multiplicity counting rules. Such rules include friendship and kinship rules that involve surveys related to households and that which links one person to a number of households of their friends and relatives. Secondly, as opposed to network sampling, in adaptive sampling, the selection of the sampling unit does not depend on previous observations made during a previous survey. This differentiates from network sampling in that to avoid biases in the calculated statistics, different estimators must be implemented. Such a process is not necessary for network sampling that guarantees the fact that calculated statistics remain unbiased.
In addition to their definitions, the core differences between the two methods of sampling also come from the advantages derived from their applications. According to Thomson, “The primary advantage of adaptive sampling is the ability to incorporate population characteristics to obtain more precise estimates of population density in that for a given sample size and cost, more valuable information can be obtained than is possible with conventional designs.” This is because, for populations that include plants and animals, fossils and mineral, this sampling method has the capacity to provide a unique way that improves the effectiveness of the sampling project. The second advantage as clearly stated by Thompson is that “adaptive sampling increases the yield of important observations (e.g. the number of endangered species observed), which can result in higher quality estimates of parameters such as the mean and variance”
The adaptive cluster sampling introduces the biases into conventional estimators so new unbiased estimators are needed. According to Thompson “if additional units are added to the sampling design wherever high positive identifications are observed, the sample mean will over-estimate the population mean”. “A method of obtaining unbiased estimators is to make use of new observations in addition to the observations initially selected”.
Network sampling on the other hand has effectively improved design efficiencies especially in areas where classical sampling methods are infeasible or inefficient. Furthermore, network sampling forms an interdisciplinary survey method of research in that it intersects the cognitive, behavioral and statistical sciences. This demonstrates that it can effectively be applied in across range of research works in the process of survey sampling. According to Thompson and Seber “fundamental knowledge about information networks linking relatives and friends are critical in designing surveys based on network sampling and knowledge gained about the robustness of these information networks from survey applications of network sampling is potentially valuable in sociological research”. This forms one of the core advantages of this sampling method and makes it superior in comparison to adaptive sampling method.
The application of adaptive sampling method in educational system cuts across a wide number of fields. This is because of its adaptability in the sampling of clustered and rare populations. In educational system, it takes the form in which the “conditions may be a given set of criteria, or a ranking system using the largest, second largest, and/or third largest, etc., order observation and taking two examples that involves a given one-dimensional case scenarios, one with a specified criteria and the other with a ranking criteria”.
Network sampling on the other hand has been in use within the educational research endeavors in surveys that presented difficulties in the definitions and execution of urinary counting rules. In addition to the above, it is in the sampling of house hold surveys and rare events. This is because; according to Thompson and Seber (1994) it is a composition of network estimation procedures that effectively reduces the levels of biasness in the population samples. In close relation, this sample method is more applicable in the sampling of elusive and sensitive populations.
The central role of statistics in decision making processes in management circles cannot be underestimated. This is because statistical data form the foundations in which critical management decisions are made in line with the aims, objectives and priorities of an organization. National Science Foundation demonstrates this fact by stating that “collecting relevant facts, figures and statistics to facilitate and support decision making process”. This fact is further buttressed by National Science Foundation in succinctly stating that “Today’s good decisions are driven by data in that all aspects of our lives, and importantly in the business context, an amazing diversity of data is available for inspection and analytical insight”. The functional area in business management where statistics play an essential role is not particularly hard to discern. This is because managers need to demonstrate strong evidence in support of their decisions. Such evidence can only be proven by statistical data. In summary, the functional area within management that statistics play a critical role is decision making.
References
National Science Foundation , Statistical Thinking for Managerial Decisions. 1994. Web.
Thompson SK & Seber AF. Adaptive Sampling. John Wiley and Sons, Inc., New York, 1996, p. 71.
Thompson, SK, Sampling. John Wiley and Sons, Inc., New York, 1992, p. 40.
Thompson WL Sampling rare or elusive species: concepts, designs, and techniques for estimating population parameters. Island Press, New York, 2004, p. 14.