Emma Diduch
In the autumn of 1896, the employees of the Strutt cotton spinning mills in Belper lined up wearing their Sunday best for a series of photographs marking the firm’s upcoming merger into the English Sewing Cotton Company. The images which survive in the Derbyshire Record Office offer a fascinating glimpse into the lives of ordinary working people – there are friends holding hands, children scowling into the camera, a row of three sisters in matching dresses – and they also spark questions about work in the Strutt Mills and the community surrounding the factory. Did these workers make good wages and have long careers in the mill? Were these children sent to work at a young age to help support their families? Would these young women soon leave the factory to get married?

Derbyshire Record Office, D6948/S/12.
Work and wages in the factory
More detail can be found in the surviving Strutt Mills weekly wage books for the 1890s. There were 657 individuals receiving wages in December 1896, including 499 women and girls, and 158 men and boys. The factory workforce included 146 families who paid rent for factory-built cottages, which was deducted from their wages, with a median rent of 2s 6d a week. Mill workers also routinely had their wages docked for tea, small loans, and other goods provided by the business.
Average wages were 11s 9d a week for men and 7s 4d for women. However, these averages obscure a significant amount of variation, because earnings were determined by age and task as well as by gender. In the daily time books, some children may be identified by their hours – working half time under the age of 13 – but otherwise the wage books tell us little about the demographic characteristics of the workforce and its connections to the wider community.
Links to census records
Given the limited information available in the factory wage books, how can we find out more about the people who worked there? One way to do this is by individual-level record linking: identifying the same people in several different sources, based on the limited information available from each snapshot.
The wage books make this process easier, because employees from the same family were grouped together and paid on one ticket, and these groups can be compared to households in the 1891 and 1901 Censuses. These census records include age and birthplace as well as occupation and relationship to head of household. Similarly, registers of children who were certified as healthy enough to begin work at 10 years old from the 1870s onwards include information about their parents and place of residence.

Lewis Hine, Children changing bobbins in textile mill.
With the additional information obtained from census data, we can see that wages increased with age, but boys were much less likely than girls to continue working in the mill as there was more demand for their labour in other local industries. Those boys who stayed at the mill into adulthood gradually rose to supervisory roles where they made the highest wages. Nevertheless, there were four women listed as ‘overlookers’ in the time books for 1881 (although they earned a few pence less than their male counterparts).
After accounting for age, the remaining gender wage gap in the mill was driven by occupational segregation, as higher paid preparatory and finishing tasks were performed mainly by men. Where men and women performed the same tasks, their wages were more equal. Despite these limitations, adult women in the spinning mill were taking home almost double what they could earn in the other main occupation open to them as domestic servants (although unlike live-in servants, their employer did not provide room and board).

Derbyshire Record Office, D6948/E/9/2.
Demographic consequences
These patterns of wages, gender, and work have additional significance to historical demographers as variables influencing demographic change. For example, we know that when couples reported marriage and fertility information in the 1911 Census, the wives of textile factory workers had some of the lowest fertility rates among the working classes.
An important question, which is difficult to answer based on the 1911 Census alone, is the extent to which wives’ occupations before and after marriage contributed to these fertility levels. Relatively high wages and demand for women’s labour like those found in textile districts could motivate delayed marriage or limited births within marriage to maximize married women’s earning potential. Factories could also represent social networks through which knowledge about birth control spread.
Automated record linking
To explore these questions, we could continue tracing our factory population through the censuses. This kind of painstaking, manual searching for individuals in multiple primary sources has always been a major part of historical and family research – but with the marriage of data science and history, new approaches to record linking can generate millions of pairs of records and identify thousands of likely matches in the time it takes a human researcher to hunt down one footnote.
Record linking at this scale often focuses heavily on men, or considers links for women only as the wives, mothers, and daughters of men. Linking across long time gaps based on the limited information available at each census (e.g. sex, name, age, birthplace) creates challenges for identifying the correct links, especially for women who may have married and changed surnames. However, to address the research questions above, we ideally want to observe women both before and after marriage. This requires combining the census data with marriage records.

James Charles, “Signing the Marriage Register” (1896). Bradford Museums and Galleries.
Full marriage certificates are not available in this kind of large data format, but transcribed indexes of the marriage registers can be used to identify couples and retrieve the brides’ maiden names. Women who did not marry between censuses can be linked directly using both their individual census records and information about their households (spouses, parents, etc.).
Starting from the 1881 Census, when the population of Derbyshire was just under 400,000, and working our way decade by decade up to 1911, when the population had grown to almost 550,000, it is possible to link over 40 percent of people in the county from one census to the next. Linking women from ‘single’ to ‘married’ in the census using indexes of marriages almost doubles the proportion of successful links for women in their twenties. When links for each decade are combined, over 50,000 individuals can be traced all the way from 1881 to 1911.
From individuals to populations
Linking censuses and marriage indexes for the whole county, rather than just for the sample of women and men observed in the Belper mill records, allows this small community to be considered in the context of other environments and employment opportunities. In addition to the small textile mills scattered along the Derwent River, Derbyshire also had districts defined by mining, agriculture, or urban commerce.
Initial exploration of the county census and marriage records suggests that women who were recorded as working in the textile industry before marriage were less likely to have high fertility than their peers in domestic service. The couples in Derbyshire most likely to have high fertility included wives who had been servants and husbands who worked in industry or agriculture. How these differences were achieved, and why, and when couples of all occupations in Derbyshire started limiting the size of their families, are more complicated questions.
Of the three sisters in the photograph, the eldest was married in 1898 at the age of 24 and left the mill, but did not give birth to the first of her three children until 1903. The youngest sister continued working and did not marry until 1910 at the age of 31. The middle sister cannot be linked past 1901 – the record linking process is not perfect, and she may have moved outside the county or died between censuses. Two brothers and a widowed mother also worked in the mill, and the family rented their cottage from the factory estate.
It is compelling to consider these individual narratives, but it remains to be seen how these reconstructed life courses fit into larger population patterns during a period of profound demographic change.
Further reading
- Diduch, E., ‘The Record Linking Glass Ceiling. Applying Automated Methods to the Census and Women’s Marriage Records, 1881–1911’. Historical Life Course Studies 14 (2024), 126-143. https://doi.org/10.51964/hlcs19189
- Garrett, E., and Reid, A., ‘Satanic Mills, Pleasant Lands: Spatial Variation in Women’s Work, Fertility and Infant Mortality as viewed from the 1911 Census’. Historical Research 67:163 (1994), 156–177. https://doi.org/10.1111/j.1468-2281.1994.tb01822.x
Tags: family history, fertility, occupations, women's employment, women's history, women's wages, women's work