User:Expclss

Sampling For Human Exposure Assessment Studies

One of the first steps in human exposure studies is to determine the target population or the individuals whose exposure is to be assessed. If the target population is small, measurements can be made on each and every individual in the population. For large populations this is time consuming and may not be economically feasible therefore a sample of people is selected and the findings from the sample are extrapolated to the population. It is important to ensure that the sample is representative of the universe or target population from which it is selected.

Probability-based sampling is used in exposure assessment studies. A probability-based sample is one that is selected from the population of interest in a manner such that, the probability of getting selected in the sample is known for each and every individual selected into the sample. Samples selected in this manner allow for defensible inferences to be made from the sample to the target population. This type of sampling ensures representativeness of the target population from which they were selected and also allows for characterization of the uncertainty of inferences from the sample.

Probability-Based Sample And Advantages

In order to reduce bias and to ensure the representativeness of the sample, probability-based samples use randomization to ensure that the sample is representative of the units on the sampling frame. The sampling frame can be thought of as a list of all elements in the population of interest (e.g. names of people, home addresses, telephone numbers). The simplest and most familiar type of probability sample is the simple random sample. A simple random sample is a sample selected from a population in a way that ensures that all possible samples of the same size have the same chance of being selected. In exposure assessment studies it is more common to use stratified sampling instead of simple random samples alone. In this the entire population is divided into a set of non overlapping subgroups or strata and a simple random sample is drawn from each strata. Stratified sampling provides greater precision than a simple random sample of the same size which in turn also helps to control study costs. Stratification improves sample representativeness and ensures that sufficient sample points are obtained to support separate analysis for different portions of the population. Example of a probability-based sample used in human exposure studies, is the National Human Exposure Assessment Survey (NHEXAS) which was developed by the U.S. Environmental Protection Agency (USEPA) in the mid-1990s. It provides critical information about multimedia and multipathway human exposures to environmental toxicants. Members of the sample in this study were selected using a probability-based sampling design.

When To Use A Probability-Based Sample

A probability-based sample should be used when it is possible to collect data for more than a minimum number of units in the target population of interest, i.e. more than 20 or 30 and you want to make defensible inferences of the target population and also quantify those inferences while making a few assumptions about the distribution of the population exposures.

Steps In Selecting A Probability Based Sample for Human Exposure Assessment Studies

The scientific process for developing probability-based sampling as outlined by the USEPA’s Data Quality Objectives is recommended for designing and implementing human exposure assessment studies. This process involves several steps. Firstly, it is necessary to specify the goals of the study. Then the population of interest must be specified both spatially and temporally. The population parameters that need to be estimated must be specified along with the level of uncertainty that is acceptable for these estimates or for the inferences. Based on the earlier specifications the appropriate statistical sampling designs must be identified. The costs must be estimated for the statistical sampling designs. Using all of this information, the optimum sampling design which fulfills most of the design requirements is determined.

References

Ott, W.R., Steinemann, A.C., Wallace, L.A.. Exposure Analysis. CRC Press (2007)

Peck, R., Devore, J., Statistics: The Exploration and Analysis of Data, 6th edition, Thomson Brooks/Cole (2008)