Stratified randomization

The aims of sampling are to establish whether eontaminants are present, their distribution and eoneentrations. Commonly-used sampling regimes inelude square grid, stratified random or simple random teehniques. Evenly-spaeed sampling points may be appropriate if the eontamination is visible otherwise judgement is required based on whether the land slopes or is flat. Samples are also taken near to the point of release. 

Spatially distributed using site criteria Stratified random... 

Furthermore, the exposure data should be obtained by stratified random sampling of houses based on population registers and data on building stock. 

While randomization eliminates bias (as least in expectation), simple randomization of all animals may not be the optimal technique for producing a sensitive test. If there is another major source of variation (e.g., sex or batch of animals), it will be better to carry out stratified randomization (i.e., carry out separate randomizations within each level of the stratifying variable). 

A nested case-control study within a cohort of rubber workers in the United States was performed to examine the relationship between exposure to solvents and the risk of cancer (Checkoway et al., 1984 Wilcosky et al., 1984). The cohort consisted of 6678 male rubber workers who either were active or retired between 1964 and 1973. The cases comprised all persons with fatal stomach cancer (n = 30), respiratory system cancer (z7 = 101), prostate cancer ( = 33), lymphosarcoma (n = 9) or lymphocytic leukaemia (z7 = 10). These sites were chosen because they were those at which cancers had been found to be in excess in an earlier cohort analysis (McMichael et al., 1976). The controls were a 20% age-stratified random sample of the cohort (z = 1350). Exposure was classified from a detailed work history and production records. An association was observed between exposure for one year or more to carbon tetrachloride and lymphocytic leukaemia (odds ratio (OR), 15.3 / < 0.0001, based on eight exposed cases) and lymphosarcoma (OR, 4.2 p < 0.05, based on six exposed cases) after adjusting for year of birth. The relative risk associated with 24 solvents was examined and levels of exposure were not reported. [The Working Group noted that overlapping exposures limit the ability to draw conclusions regarding carbon tetrachloride.]... 

In a population-based, case-control study in 232 postmenopausal women who had had a non-fatal myocardial infarct during the previous 3 years, a stratified random sample of 723 postmenopausal women without a history of infarction acted as controls (32). Among hypertensive... 

Stratified random sampling. The parent population has first to be subdivided or... 

Stratified random sampling, which is a variation of simple random sampling, is used for media that are stratified with respect to their chemical and physical properties. Each stratum is identified and randomly sampled. The number of grab samples and the sampling point selection depend on the nature of contaminant distribution within each stratum. Stratified random sampling is used for the characterization of multiphase liquid wastes or process waste batches that undergo stratification over time and/or space. 

Kay (2005) highlighted several methods of randomization, including simple randomization, block randomization, and stratified randomization. Simple... 

In stratified random sampling, the total site is broken down into a number of strata or subpopulations and random samples are taken from each stratum. Figure 2.3 provides an example. 

Possible uses for stratified random sampling are for spill sites, waste sites, where site-specific management is needed, for atmospheric deposition, in soil sensitivity studies, for effluent and discharge dilution or extinction analysis, and lastly, for modeling. 

Figure 2.3 Stratified random sampling. This approach divides the site into smaller zones. Samples are collected randomly within each zone, and could even be composited.

The use of a stratified random sampling technique would be useful. 

You decide to compare the records in all of these categories during several periods of disaster, with the records of the same hospital just before the disaster in a stratified random sample of hospitals from regions in which major disasters occurred in the last... 

---