Study Design Essay

A study design is a design structure planned to conduct a particular study which enables a researcher to solve a research question. The main objective of a study design is to give the most productive and accurate estimates of a population. First, we need to determine the scope of inference of the study design. Scope of inference is the population studied on a particularly interested study area where inference made will be applied.
To evidently define the population, spatial, temporal and ecological dimension are needed to be considered. Spatial is the space where the research is conducted and spatial should also be considered when choosing the quantity of sampling units and sampling units itself. Temporal is the period needed for the research to be conducted. Temporal is important as it helps to estimates the minimal time needed for the inference to be valid. Ecological dimension refers to the interested study area where the research should be focusing which is involved in determining the scope of inference. Therefore, population needs to be precisely defined or else the study design would be ineffective and inefficient. But sometimes, in real practices, exploring more than the intended scope of inference would further obtain more information.
Sampling units plays a vital role in study design as the units sampled will provide answers or solution to the research question. Therefore, sampling units used in a study needs to be well-defined in terms of spatial, temporal and ecological dimension as well. Types of sampling units include natural or artificial units. Sampling units are free to take on any interested independent or dependent variable which will either provide an explanation or a response. Pseudoreplication is a sampling error occurred when the sampling unit is a subsample assumed as an independent replicates. Pseudoreplication mistake are commonly made, either intentionally or unintentionally because it would greatly increase the degree of freedom thus making a hypothesis test looks important. Most of the time, sampling units occurs in multiple levels where the lower level sampling units are branched from higher level of sampling units. There are two ways to handle such sampling units, one is where an analysis for each level is conducted separately and the other is to combine all levels of sampling units into a model.
There are three types of experiments or studies namely manipulative experiment, comparative mensurative experiment and observational studies used in a study design which depends on the sources of variability. Manipulative experiment needs at least two manipulations which are imposed by the researcher and the outcome obtain is being observed. The sources of variability are fully controlled and thus avoiding confounding. Furthermore, comparative mensurative experiment also involves two or more treatments but the researcher do not impose and the attention is on the correlation between the sources of variables. Next, observational study does not involve any treatments therefore control over the sources of variability is low. The attention of an observational study is on association where the output will be associated with the input.