The comparison of allocated treatments is a foundationfor many studies in soil science. We often want todetermine if a treatment applied to a soil has asignificant effect compared with not applying thattreatment (control). Often, a statistical test is performedto establish whether there is a true difference in themeans of the treatment and control. The null hypothesisstates that there is no difference in the means. There aretwo types of error that can be made when a statisticaltest is conducted. These are presented in Table 1.In soil science research, there is an overwhelmingfocus on controlling the Type I error (i.e., the value ofa). When a comparison test yields a non-significantresult, the usual reaction is to assume that there was notrue affect of applying the treatment to the soil.
展开▼
