Hodess, RobinBanfield, JessieWolfe, Toby2019-09-252019-09-252011-04-012001393571100Xhttp://hdl.handle.net/20.500.12424/177332Corruption hardly lends itself to measurement. It tends to be hidden from view, and the parties to a successful corrupt transaction seldom have an incentive to be open about their dealings. Until the mid-1990s, most empirical findings on corruption in the academic literature were of an incidental or anecdotal nature. Aggregated analyses, whether across time for a given business sector, or in crosscountry comparisons, were speculative and theoretical, often citing ‘impressionistic evidence’ as their basis.1 In a methodological essay, entitled ‘What Cannot be Analysed in Statistical Terms,’ corruption was cited as the classic example of an observable phenomenon that was not quantifiable since ‘there cannot be statistics on a phenomenon which by its very nature is concealed’.2 Transparency International’s (TI) Corruption Perceptions Index (CPI), first published in 1995, changed these assumptions. In subsequent years, there has been a remarkable growth in empirical research on corruption, fuelled to a great extent by growing international interest in finding the means to curb it. It has been bolstered by support and interest from multilateral organisations, foundations and researchers at universities in a host of countries. This section of the Global Corruption Report 2001 reviews some of the current comparative empirical studies of corruption undertaken by international organisations, aid agencies, research centres, NGOs and private companies. The studies range from opinion polls and composite indices to regression analyses, focus groups and diagnostic studies. They can be divided into three sub-sections: surveys and polls on a variety of aspects of corruption; recent secondary analysis of corruption data; and studies of public integrity and institutionsPages: 92engWith permission of the license/copyright holderresearch ethicsPolitical ethicsCommunity ethicsGovernance and ethicsBook chapter