Addi van Bergen

General introduction 13 1 Measuring social exclusion Not surprisingly, given the information in the previous paragraph, a generally accepted measure of social exclusion is lacking in health research [17, 38, 51, 53, 54]. The most common approach to measuring social exclusion is to use indicator lists with data that are usually drawn from pre-existing datasets [38]. The number, type and dimensions of indicators used to define exclusion vary greatly [17, 38, 55, 56]. Usually, measures focus either on participatory aspects of SE, social relationships and networks or on poverty and labour market participation [17, 38, 56]. The whole construct of SE is rarely represented. A further issue is the general lack of clarity as to whether the items included are risk factors or outcomes of SE, i.e., indirect or direct indicators of SE [17, 38]. Studies rarely attempt to quantify SE using indicators across a number of domains or dimensions [38]. Typically, no composite measure is calculated, or simple sum scores are used with equal weights given to all items or dimensions, which is unlikely to be empirically correct [57]. Few or none of the measures of SE identified were formally validated [38, 54]. Over the past two decades, significant research has been done by the SCP on the measurement of SE in social and economic policy research [29, 33, 41, 42, 58]. In this dissertation, we build on the knowledge and experience gained in this process. In particular, the SE index developed by Hoff and Vrooman [29, 59] has the potential to be developed as a standard in the health domain if adapted to the needs and preferences of users. This index, here referred to as the SCP social exclusion index, consists of 15 items measuring the four dimensions of SE, ‘lack of social participation’, ‘material deprivation’, ‘lack of normative integration’ and ‘inadequate access to basic social rights’. The selection of items was not defined a priori but was determined empirically. Four focus groups were conducted with ‘average citizens’ to test the relationship between the SCP theoretical concept and the everyday meaning of social exclusion in the Netherlands. Persons with low levels of education and/or low income were overrepresented. Some typical answers on what social exclusion in the Netherlands currently means, were: ‘ being very lonely, breaking down a little day by day’, ‘having no contacts’, ‘having no respect for other people’, ‘not holding the door for somebody, not saying ‘thank you’ when receiving change’, ‘being in debt’, ‘not having much money, because that means you have fewer opportunities ’, and ‘ people who are disabled or don’t know the language, or who have no idea where to turn to ’ [29]. In each session, the participants evaluated whether the items of a lengthy master questionnaire gave an accurate and complete picture of the subject matter. Based on their comments, questionnaire items were supplemented, reformulated or removed. The amended version then served as input for the next group session. In the next step, individual cognitive interviews were held with eight ‘average citizens’ to test the interpretation and comprehensibility of the questions and answer options.