Stratigraphic and sedimentologic studies have statistical implications not recognized by many field workers. Only the geologist who has been statistically trained will be cognizant of the statistical aspects of what he sees in the field.
A detailed Mississippi Delta map was divided into three environments (land, fresh water, and marine), and random sample points were used to prepare maps to demonstrate the extent that resolution improves with increased control points. Most modern stratigraphic studies are based on too few data points to resolve complex environments by mapping.
Brunton-compass and tape determinations of 32 duplicates of measured stratigraphic thickness in varied structural situations yield a 95 percent confidence limit of 11 percent for precision of individual thickness measurements. The minimum thickness difference that can be detected with 95 percent certainty is 17 percent, assuming that one point is compared with four others for isopaching (Duncan’s new multiple range test, 1955). Variable isopach spacing is recommended to yield maximum information at a uniform level of reliability.
To extract maximum geologic meaning from field data, there should be feedback between the statistical and field phases of the work. An example is developed for analysis of variance study of grain size in the “Chemung” Formation of the central Appalachians. A preliminary experimental design is modified in response to detailed stratigraphic description, so that a more sophisticated analysis of variance factorial design permits more specific questions to be answered than was previously possible.