Joris Kok, International Institute of Social History
Despite advancements in the access and availability of data for nineteenth-century the Netherlands, research on subgroups of the population, such as Jews, remain limited. In order to study such subgroups, a source needs to report on the religious affiliation or ethno-religious background of an individual. As Jews made up a mere 2 percent of the Dutch population and 10 percent of the capital’s inhabitants, random samples of such sources often lead to Jewish samples too small to use. An alternative has been to use Distinctively Jewish Names to oversample likely Jews. Despite the low levels of miscategorization using this approach, it also leads to small and biased samples. To get larger and less biased samples, I propose using more information in the identification process, in particular more names when available. In this paper, I approach this problem for nineteenth and early twentieth century Amsterdam in two stages. First, following Levitt and Fryer (2004), I construct a Jewish Name Index using full-count information from mid-nineteenth century Amsterdam. As religious disaffiliation and conversion were still extremely rare at this time, information on religious affiliation for this time is considered highly accurate. Second, I apply this index to all names on the marriage certificates of Amsterdam from 1811 up to 1932. Dutch marriage certificates included both the names of the groom and bride as well as their parents’ names. Using the combined information from the groom’s and bride’s families separately, I am able to identify approximately 75 percent of all marrying individuals as either Jewish or non-Jewish. Robustness checks show virtually no miscategorization of individuals.
No extended abstract or paper available
Presented in Session 144. New Historical Data Infrastructure I