Identifying dissonant and complementary data on women through the triangulation of historical sources

Abstract

If triangulation and its worth have long been contested amongst social scientists, historians have not discussed it. In this paper, a historical demographer practises data triangulation by combining qualitative and quantitative sources. The aim is to explore how these sources identify nineteenth-century women's occupations and thus challenge the gender bias found in population registers as they report incomplete information on women's work. This bias is acknowledged by feminist historians and also evident in quantitative records in developing countries. To explain the outcome of dissonant data that this historical study shows and shares with modern triangulation approaches, women's ability to represent their occupational identities in the different sources is discussed. Some of the epistemological implications that arise from the triangulation of data that subsists under separate paradigms are also reflected upon. Although triangulation is far from infallible, it is argued that it helps to gain, view and question knowledge.

Digital humanities and the history of working women: a cascade

“…Perhaps it is the illusion of the present, where one is often inclined to believe that knowledge is expanding more rapidly than in the past, but it might not be entirely illusory to believe that these components are stronger than ever, and mutually supportive. It is, we believe, no coincidence that a number of fine essays (Goose, 2007;Janssens, 1997;Sharpe, 1996) and works of synthesis based on studies processing vital registers as well as census material have recently appeared (Battagliola, 2000;Burnette, 2008;Goldin, 1990;Vikström, 2010). And we will no doubt be seeing more appear in the near future as the cascade of data, occupation-based measures, and statistical methods we are currently witnessing translates into ever more innovative and wide-ranging studies.…”

“…In fact, as the under-registration of women working on a farm or in a shop owned by their husband is often regarded as a particularly urgent problem (see, for example, the recent special issue of this journal on households, family workshops, and unpaid work, with contributions by Bessière (2014); Cesaroni and Sentuti (2014); Martini and Bellavitis (2014); Micheletto (2014);Schmidt (2014), and Whittle (2014)), the level of under-registration is likely to have been less of a problem where there were few farms or shops. Comparisons of census data on women's work with corresponding data from other sources -sometimes called data triangulation -can be both very time consuming and highly illuminative in that respect (Vikström, 2010). The articles by van Nederveen Meerkerk & Paping (2014) and by Devos et al (2014) in this issue show great ingenuity in triangulation, as well as in finding routines to remedy flaws in the census.…”

“…Many types of such data are now accessible: court records (Erickson, 2010), trade directories (Vikström, 2010, and the article by McGeevor (2014) in this issue), insurance records (Kay, 2004(Kay, , 2009), interviews (Bras, 2004(Bras, , 1999, probate inventories (Whittle, 2005), poor relief records (Van Leeuwen, 1993, household budgets (Horrell & Humphries, 1995), autobiographies (Humphries, 2010;Maynes, 1995), photos (Jacquet-Francillon, Kerroubi, & Cacoualt, 1994), newspaper advertisements for jobs (Schulz, Maas, & van Leeuwen, 2014a; see also Franzosi, 1987), and, last but not least, historical censuses and vital registers. These types of data are not new to historians, of course, but due to advances in digitization methods their scope in terms of geographical and temporal coverage as well as their sheer volume have increased enormously in the recent past (Hall, McCaa, & Thorvaldsen, 2000).…”

Lost in registration? Missing occupations of single women in the Bruges countryside, c.1814

“…Since it is highly unlikely that an elderly woman would start such a career so late in life, we can assume that the census-takers, for whatever reason, chose not to register Rosa's job. However, the amount of time required to trace a single individual through multiple sources makes this kind of triangulation prohibitive when dealing with large sample sizes (Whittle, 2005;Vikström, 2010;Erickson, 2012).…”

“…The second trend in the historiography uses alternative sources to find data on women's labour: probate inventories (Whittle, 2005), court records (Erickson, 2010), business and trade directories (Vikström, 2010), local population listings (Grantham, 2012), etc. Other authors rather seek to exploit census data in new ways.…”