Abstract |
A frequent problem in the statistical analysis of data in environmental sciences is the synthesis of results obtained independently from various sets of data such as from different measuring points or from replicated experiments, etc. Unlike the analysis of the separate sets themselves, their ultimate combined evaluation has been mostly given in descriptive terms only. It is that critical final stage of data synthesis where meta-analysis comes in. Some principles of the method are presented and their usefulness is discussed. The approach is illustrated using previously published results on the hydrological effect of rainfall enhancement in Israel. It is shown that while the resulting increase in discharge from separate hydrologic zones is statistically insignificant, the integrated totality of results (assuming independence) is significant at less than 5 percent. |