To this point, we have been providing consensus descriptions of schizophrenia, what it is and what it is not, and describing the means by which it is detected and diagnosed. In this and later chapters, we present the evidence about the causal factors, treatments, and outcomes of schizophrenia from scientific studies. Such studies sometimes find results that differ from each other due to differences in the methods used or the types of patients that are studied. Random differences in measurement between studies also leads to discrepancies, which is perfectly normal.How, then, can we come to firm conclusions in the presence of variable results from different studies? Our approach as scientists, and as authors trying to distil the facts, is to always rely on the preponderance of evidence, or the best estimate that can be made when putting all the evidence together. Thus, as we present the facts moving forward, we will base our claims on the largest studies avail­able, since these usually give more reliable results than small studies. Whenever possible, we will present the results of analyses that put the results of otherstudies together using a formal statistical method called ‘meta- analysis’. Thus, instead of comparing and contrasting the results from two or more studies, we will let the reader know the overall result found when all studies were pooled together. In some instances, however, it is instructive to compare and contrast studies because each study tells us something different and uniquely important, and we will point this out when doing so.In this chapter, we describe the epidemiology of schizophrenia. Epidemiology is a branch of science concerned with the distribution and determinants of illness in the population, and the transmission of illness within families. Two important epidemiologic measures of disease burden in society are prevalence and incidence. The prevalence of schizophrenia (i.e., the number of affected individuals in the population) has been estimated at least 60 times in 30 dif­ferent countries. The prevalence estimates seen in these studies are very consistent, despite cultural differences between samples and the dif­ferent methods used and timeframes sampled in the studies.