SPADA Meeting Book

Prior to using a soil for detector or assay testing, it is important to consider 339 experimental design and what is required to have defensible data prior to collecting a soil 340 or starting an experiment. 341 Judgement samples are those resulting from an individual deliberately avoiding a 342 particular area, thus exercising judgement in selecting sampling sites, and therefore are 343 highly biased and not recommended for most purposes. Judgement samples should be 344 reserved for instances when the individual collecting the sample is only using the sample 345 as a source of soil microbes. 346 From a statistical perspective, simple random samples are more representative of an 347 area than judgement samples as each sample has an equal opportunity to be selected. A 348 common technique for collecting simple random samples is to establish a grid consisting 349 of two sets of parallel lines at right angles to each other. Each line is assigned a unique 350 number. Pairs of numbers are drawn from a random number table and used to establish 351 intersecting points that denote where a sample will be collected. This process is repeated 352 until the number of required samples is reached. Another simple random sample method 353 is to use a pair of random numbers to designate a distance and angle from a selected 354 starting point. 355 Stratified random sampling is similar to simple random sampling except that the area 356 of interest is divided into smaller sub-areas. These sub-areas are selected based on known 357 variations in the soil or other factors of interest. Samples are collected within each sub- 358 area in a random manner as described above. The advantage of stratified random 359 sampling is that a researcher can compare results between the sub-areas and possibly 360 correlate results to the factor of interest. 361

Draft Version 18

17