III. >> M  ethodological Approaches to Human Rights Indicators >> Sources and data-generating mechanisms Box 21 Are statistical surveys better than expert judgements? DIAL household survey on corruption Research by Development, Institutions and Long Term Analysis (DIAL), a research organization based in Paris, used household surveys which revealed the weaknesses of expert opinion surveys on corruption. It also showed the limitations of using some global databases based on expert opinions/judgements for cross-sectional comparisons and comparisons over time. Using simultaneously household and expert opinion surveys on corruption in eight sub-Saharan African countries, the research revealed that experts systematically overestimated the extent of corruption compared to household perceptions. It also showed that perceptions on corruption or governance as a whole between vulnerable groups or victims and the influential group, which includes the experts, can vary significantly. Sources: M. Razafindrakoto and F. Roubaud, “Are international databases on corruption reliable? A comparison of expert opinion surveys and household surveys in sub-Saharan Africa” (DIAL, 2006). See also Naval, Walter and Suarez de Miguel, Measuring Human Rights and Democratic Governance, box 6.5, p. 117. 4 Data based on expert judgements Data based on expert judgements covers data generated through combined assessments of a human rights situation with the help of a limited number (or sample) of “informed experts”.35 The information generated is essentially judgement-based or subjective and needs to be translated into quantitative form through coding,36 as with the household perception and opinion surveys. Unlike the latter, it usually involves a more systematic use of diverse sources of information, including the media, government reports and reports from NGOs, by a limited num- ber of experts (e.g., advocacy groups, academic researchers, social scientists, managers) who are asked to evaluate and score the performance of States. Notwithstanding the obvious limitations of this method, data based on experts’ judgements have been frequently used for cross-country ranking and comparisons over time. Initiatives in this category have primarily focused on assessing civil and political rights, though the extent of references to the human rights normative framework varies significantly among them.37 As with household perception and opinion surveys, 35. This category of data-generating mechanism is sometimes referred to as “standards-based data” in the human rights literature. The terminology chosen here seeks to avoid possible confusion around the notion of “standards”, which is also referred to in other categories of initiatives, as in events-based data on human rights violations whose definitions are also based on international or national human rights standards. 36. Coding refers to a procedure for converting verbal or textual information into numbers or other symbols which can be more easily counted and tabulated. 37. On this issue, see, for instance, K.A. Bollen, “Political rights and political liberties in nations: an evaluation of human rights measures, 1950 to 1984”, in Human Rights and Statistics. 66 Human Rights Indicators