Copyright Sociological Research Online, 2001

 

Wendy Bottero and Kenneth Prandy (2001) 'Women's Occupations and the Social Order in Nineteenth Century Britain'
Sociological Research Online, vol. 6, no. 2, <http://www.socresonline.org.uk/6/2/bottero.html>

To cite articles published in Sociological Research Online, please reference the above information and include paragraph numbers if necessary

Received: 2001/7/05      Accepted: 2001/8/29      Published: 2001/8/31

Abstract

This paper examines the hierarchy amongst female occupations in Britain in the nineteenth century, using information on marriage and family patterns to generate a measure of distance within a social space. This social interaction approach to stratification uses the patterning of close relationships, in this case between women and men, to build up a picture of the social ordering within which such relationships take place. The method presented here starts, not with the assumption of a set of broad social groups that may interact to a greater or lesser extent, but from the opposite direction, from the patterns of social interaction among detailed occupational groupings. Instead of reading off social hierarchy from the labour market, we use relations of social closeness and similarity (here marriage) to build a picture of the occupational ordering from patterns of relative social distance. Such an approach is possible because of the way in which social relations are constrained by (and constrain) hierarchy.

Keywords:
Assortative Marriage; Differential Association; Hierarchy; Homogamy; Social Distance; Stratification.

Introduction

1.1
Differential association has long been seen as an essential element of stratification, and the impact of hierarchical position on our choice of friends and partners, and on our cultural tastes and activities is well known. Spouses, friends and acquaintances tend to be chosen from amongst those who share a similar lifestyle, which is in turn dependent on the resources and rewards that are available to groups of people. As Bourdieu puts it, agents occupying neighbouring positions in social space develop a 'sense of one's place', which is 'at the root of all processes of cooptation, friendship, love, association etc.' (Bourdieu 1987:5). There is, for example, a wealth of empirical evidence on patterns of endogamy and homogamy: by education, by class or socio-economic status position, by religion and ethnicity (see Kalmijn, 1991; 1998 for reviews), however, the theoretical implications of such research have been less well developed.

1.2
Whilst research on marital selection has been used to look for signs of social distance, this is usually an investigation of barriers or limited interaction within a pre-ordered occupational or educational hierarchy. Indeed, despite a long theoretical tradition seeing stratification in associational terms, information on differential association has rarely been used to generate statements of the ordering or hierarchy (although see Laumann and Guttmann, 1966; Laumann, 1966; Stewart et al. 1980; Levine, 1990; Rytina, 1992; Bakker, 1993). However, relational - or social interaction - approaches allow us to map the social hierarchy with no a priori assumptions about the nature of that structure, simply using patterns of association as the basis of the technique. Social hierarchy acts as a constraint on all close social relationships and, in turn, the patterning of such relationships helps to transmit and reproduce hierarchy itself. Through processes of contiguity and social similarity we tend to form close relationships with individuals near to us in the social order, and thus help to reproduce that social order. Relational approaches indicate that the patterning of a variety of different relationships - such as friendship, marriage and family connection - indicates a very similar social hierarchy (Mitchell & Critchley, 1985; Hout, 1982; Prandy & Bottero, 1998; Prandy & Jones, 2000)[1]. By exploring the pattern of occupations held by those in close social relationships – and therefore the relative distances between groups – we can map the social ordering in which such relationships occur (Prandy, 1999).

1.3
The relational approach to measuring stratification arrangements has been most fully developed in the Cambridge Scale, which used friendship and marriage choices amongst modern-day respondents to order occupations (Stewart et al, 1980; Prandy, 1990). The historical record is more limited, but allows a detailed examination of the relationships that surround marriage (Kendall, 1971). This paper uses information on nineteenth century marriage patterns to build up a picture of the social order in that period. The starting assumption is that close social relationships (through friendship, family or marriage) occur in situations of social similarity. We might expect the families of two people entering into a marriage partnership to be roughly similar in their social position. The idea, put crudely, is that 'like marries like'. This does not assume a simple model of endogamy since marriage inevitably occurs across social divides, between individuals with a gulf between their backgrounds. However, such marriages are relatively infrequent and the overall pattern of marriage choices is more likely to occur between persons sharing a similar social position.

1.4
A cross-tabulation of the occupations of the respective parties to a marriage (brides and grooms and their parents) provides an indication of which occupational groups are more likely to inter-marry and which less likely. We can regard the extent of inter-marriage as indicating the degree of social 'closeness' and 'distance' and from such a cross- tabulation it is possible to develop a hierarchical ordering of the occupations, given solely by the patterns of marriage between the occupational groups. Although such an analysis is relatively unproblematic when we link the occupations of men related by marriage (Prandy & Bottero, 1998), additional problems arise when we consider the links between the occupations of husbands and wives, both in assessing the relative significance of occupation in women's lives, and also in making comparisons between the male and female occupational structures.

1.5
The particularity of women's occupational experience has always created problems for mobility analysis, and researchers have struggled to develop a more inclusive approach. Historical mobility research has tended to concentrate on the experience of men alone, typically exploring the extent to which the social position of fathers is inherited by their sons. Women's mobility experience has been, at best, represented by proxy, through the occupations of their fathers and husbands, with little consideration of the influence of women's occupations on either their own situation or that of their families. This partly reflects the different significance of paid employment in women's lives in the nineteenth century. The under-recording of women's occupations in official records creates difficulties for those attempting to explore the employment experience of women, as does the 'crowding' in the female occupational structure that is recorded. The difficulty of the exercise, however, does not limit its consequence. Social reproduction involves not merely the passing on of family position, but also – through marriage – the creation of a new household, in which the representatives of two families are joined. In this process, it is as much the social position of the bride (and her family) that is reproduced as the social position of the groom (and his family).

1.6
The role of women in the social reproduction of advantage needs to be more fully explored, not through an examination of female occupations or wealth alone, but through a fuller consideration of how women are involved in the social relationships which are both the product and reproducer of hierarchy. The method advanced here attempts to include women more directly into mobility analysis, and on the same terms as men. Although, as we shall see, the nature of women's occupational experience in the nineteenth century still creates difficulties for analysis, a social interaction approach to occupational hierarchy allows for a more integrated treatment of women's situation.

Placing Women's Occupations

2.1
The more intermittent nature of female employment has created difficulties in 'placing' women's jobs in class schemes, since it is not clear whether to take a woman's own labour market situation, or her wider social relationships, as determining her class position (Stanworth, 1984; Heath & Britten, 1984). The 'conventional' solution to the problem takes the household as the unit of class stratification, and uses the occupation of a single individual within that household (typically, the male head)[2]. For individuals with no current paid occupation, position in the social hierarchy will depend on relationships with individuals who do have a direct link to the labour market. The trouble with this solution is that it is incomplete. If we accept that this influence operates for individuals without an occupation it must also operate for those individuals with an occupation.

2.2
Each household member in paid work will affect the lifestyle and outlook of the people they live with, so the network of close social relationships in which every individual is located will influence their position in the social hierarchy and mediate the 'direct' effect of their occupation. In which case, social class position is not just a relationship to the labour market, but the sum result of the close social relationships in which individuals are located. The social interaction approach to measuring the social order focuses on the location of occupations within a network of wider social relationships. Here it is the social ordering of occupations which is key, rather than their labour market location alone. Thus, the occupations of both women and men are ordered in terms of their relative social distance given by patterns of social interaction. This technique allows us not only to explore the relationship between the male and female occupational structures, but also to consider how gender inequality is built into relationships of 'social equivalence'[3].

2.3
Within marriage, husband and wife may meet on terms of inequality which are reflected more generally in society, for example, in the different labour market returns and status of 'male' and 'female' occupations. However, notwithstanding this overall gender inequality there is a very strong patterning of similarity in the other relationships around the marriage tie itself. We can see the female deficit in rewards as a reflection of the different status of women and men in society, however, such deficits are routinely bridged in close social relationships, such as those between husbands and wives, or fathers and daughters. The 'in- laws' of a married couple, for example, are generally from around the same social level (Mitchell & Critchley, 1985; Prandy & Bottero, 1998). So we can regard the female relations of the husband as socially equivalent to his wife, whilst the male relations of the wife are socially equivalent to her husband.

2.4
In looking at the occupations of women and men therefore, it is important to try and establish which occupations are socially commensurate in marriage and the family, even though in labour market terms they may be quite unequal[4]. Unlike class analysis, which first establishes economic divisions in the labour market and then looks at the consequences for social relations, the social interaction approach locates the hierarchy of occupations in terms of the way that they map on to patterns of social interaction. Here it is the place of occupations within patterns of social distance – for men and women – which takes priority.

Social Interaction Linking Male and Female Occupations

3.1
Our study uses family histories, drawn from members of family history societies, to generate a national sample over an extended time period (Prandy & Bottero, 1998; Prandy & Bottero, 2000)[5]. Since our analysis of the social order takes as its starting point patterns of social interaction, in order to place those involved in the relationships we need to give them occupational tags. Some caution must be exercised with historical occupational titles, and indeed with the significance of 'occupation' in the past, and further complications arise when we consider the female occupational structure. Inevitably, caution must be exercised in the analysis of women's occupations, both in how we interpret the meaning of those occupations which are recorded and in how we assess the potential bias of those occupations which are not recorded (Higgs, 1987; Snell, 1985; Lindert, 1980). Yet large numbers of women did work in the nineteenth century, and contributed substantially to the financial maintenance of their families. Not all of this work is recorded in the official record, but we can try to explore the significance of what is recorded and whether that significance changed over time.

3.2
Marriage records typically give the occupation of the groom, the father of the bride, the father of the groom, and (less frequently) the bride. All previous studies of endogamy in nineteenth-century Britain have compared the occupations of the male relatives of the union, taking the occupation of the father of the bride as proxy for the bride's occupation (Foster, 1974; Crossick, 1978; Gray, 1976; Penn & Dawkins, 1983; Penn, 1985; Miles, 1993; Mitch, 1993). Our original intention was to analyse all six (potential) pairs of relationships in the marriage union, enabling us to deal with both male and female occupations on exactly the same terms. A technique such as correspondence analysis (Greenacre, 1984; Weller & Romney, 1990) uses the cross-tabulation of pairs of occupations to determine (if there is one) the major dimensions along which the row and column categories can be ordered. So, the occupations of the women are ordered by their relationships to their male relatives, whilst the occupations of the men are ordered by their relationships to both their male and female relatives[6].

3.3
However, we found that using female occupations to order the occupations of their male relations gave rise to a hierarchy considerably at variance with the orderings produced by the links between different groups of male relatives. Occupational groups which appeared at the top of the social hierarchy in the all-male pairings (such as 'large farmers' and 'professionals') turned up much lower in the social hierarchy produced by the male-female pairings. This was because difficulties associated with the female occupational structure in the nineteenth century created problems for the correspondence analysis method. The tendency to under-record women's occupations in official records meant that we often lacked the necessary information about the occupations of the women in our sample at the time of their marriage[7] and, the absence of an occupation for women in the sample was not randomly distributed. The higher the man's occupational position, as determined by the male-male pairings, the less likely we were to find an occupation for his female relative. It was rare to find men in very high level occupations (such as the professions, large farmers, army officers etc.) who could be matched to wives, daughters or daughters-in-law with a listed occupation. This reflects the 'genteel' tendency for women from higher status or more wealthy families to shun paid work. However, it creates problems for a relational approach to measuring the social space.

3.4
To place occupations in a relative social ordering we need to be able to make the necessary pairings, linking the occupations of individuals in close social relationships. The skewed nature of the female occupational structure means that, for higher status male occupations, the female half of the pair is frequently missing. As a result, the female pairs that we do find for such jobs are likely to be highly unrepresentative of the wider social position of that occupational group as a whole.

3.5
In our sample the daughters of 'professionals' rarely have an occupation recorded at their marriage, so those men whose daughters do have an occupation are likely to be atypical of their group. Only a small number of men in the 'Clergy' category had a daughter with a listed occupation at marriage. Some of these daughters' occupations were very clearly matched to the occupation of their father (for example the daughter of a 'missionary' recorded as a 'head mistress, missionary school' at her marriage in 1851) and were part of the very select number of higher social order jobs available to women at the time. Other female jobs seem less well matched to the father's occupational group. Another clergyman in our sample is listed as a 'curate /vicar' on his daughter's marriage certificate in 1845, whilst his daughter is working as a 'seamstress'. However, the father's occupation on the marriage certificate is posthumous (his death recorded in 1840). The apparent gap in the social standing of the two occupations may therefore reflect a decline in the daughter's social circumstances following her father's death, rather than any actual 'social equivalence' between the jobs, a surmise reinforced by the fact that the man she marries in 1845 is a 'tailor'. Of course, all occupational groups contain individuals in a range of social circumstances, and it may be that the category of 'clergymen' includes an extreme of 'lowly' or penurious cases. However, because so few of the men in these categories have daughters in paid employment, the father-daughter pairings become unduly weighted by what appear to be atypical individuals.

3.6
The minority of men in 'elite' occupational groups who had wives, daughters or daughters-in-law in paid employment were often linked to male relatives who were much lower in the occupational scale than was typical for 'elite' occupations. The very fact that 'elite' men had female relatives in paid work indicated that their social circumstances (and wider social relations) were unusual for their group. To order the whole group of 'professionals' by the occupations of that minority of their daughters who were in paid work may build up a false picture of the mens' relative social location, underestimating their position in the hierarchy. The truncated nature of the female occupational structure in the nineteenth century and thus the dearth of paid employment amongst women at higher social levels means that (at least for elite occupational groups) we must regard the male-female pairings as too unreliable an indicator of social position.

3.7
Because the skewed nature of the female-male pairs created problems for a direct correspondence analysis we decided to adopt a different solution to locating the women in terms of their wider social relations. The first stage was to develop historical scales for male occupations, using correspondence analysis, based on the marriage interaction patterns of male respondents (the groom, father of the groom and father of the bride) alone. This gave rise to two nineteenth century scales for male occupations (for the first and second halves of the century) in which occupations were ordered by (male relatives') marriage patterns. The technical aspects of this work have been detailed elsewhere (Prandy & Bottero, 1998).

3.8
Because of the problems that we have described, we adopted a different technique in order to deal with women's occupations. However, it is based on the same principle that social position should be indicated by wider social relations.

  1. Male occupational scores were first established using correspondence analysis of the occupational pairs of male relatives.
  2. These scores were then used these to build up a picture of the social relations of each of the female occupational groups by aggregating the occupational scores of the male relatives linked to the female members of these groups (see Figure 1).

3.9
As the diagram indicates, this means that while male pairings produce scores on a mutual basis, for female scores the direction of measured influence runs in only one direction – from men to women.

3.10
As we have tried to make clear, this is not because we believe that women's social position is determined only by their male relatives or that women's occupations have no influence on close male relatives. Indeed, our concern in constructing a historical scale of female occupations has been to try to determine, at an individual level, what role women's occupations play in their own mobility experience and how it is related to that of their male relatives. To do so, however, we need to get some sense of the overall social ordering of the female occupational structure. It is because of the truncated nature of that structure that a full range of mutual relational pairings with male occupations is not possible. Instead, to indicate each occupation's place in the social hierarchy we have used the aggregate pattern of male relationships of the incumbents of each female occupational group. Despite the difference in technique, the same theoretical principle underlies our approach to both women and men. Occupations are ordered in a hierarchy on the basis of the wider social relations of the incumbents of the occupations. For both men and women it is the social relations of occupational groups that is key.

Calculating Scores For Female Occupational Groups

4.1
In the official record there is a tendency for younger women's work to be more visible than the occupations of older married women. This pattern, found in the census, is repeated in marriage registers, which allow for the recording of the occupation of the bride at the time of the marriage, but not for the occupations of the mother and mother-in-law. Women's death certificates normally record the occupation of the woman's husband rather than indicating any occupation that she might herself have held.

4.2
Our initial intention was to use information from marriage records alone, but this would have meant an almost exclusive emphasis on the occupations of younger women. However, our alternative averaging method is not tied to marriage records. All that is required is that the occupations of women and their male relatives should be matched at about the same point in time (because we are trying to link those in situations of relative social equivalence). So, as well as linking the occupations of women to those of their husbands, fathers and fathers-in-law at the time of their marriage (from marriage records, the point at which two families join), we have also linked women to their husbands at a second and later point in their life (where both have a listed occupation held within five years of each other)[8].

4.3
This draws on a wider variety of official and non-official records (including trade directories and workplace records) and places a greater emphasis on the occupations of older women. To ensure a greater number of occupations held by married and older women, for this second group of matches between husbands and wives we took the last occupation held by the woman. However, we excluded occupations held by widows (even where we could make a match to the husband's occupation held within five years of the date of the widow's job). Typically, a widow's occupation is more likely to reflect a change (usually a drop) in the woman's social position following the death of her husband, so linking her job to the husband's last job before death is not necessarily a link between commensurate positions in the social order.

4.4
In analysing the social relations of the female occupational structure it was necessary to first group the occupations in question. Our principle has been to retain as much information as possible about occupation, distinguishing jobs at the level of occupational title. This level of detail is particularly useful when looking at female occupations, because of the great concentration of women into a limited number of industries. Despite the high concentration of women in domestic services, it was possible to distinguish more detailed groups such as 'housemaid', 'lady's maid', 'cook' and so on. Despite these fine distinctions, the occupational 'crowding' of women meant that we had still had fewer occupational groups for women than for men. Large numbers of women are listed in official records simply as 'servant' or 'maid' and no further differentiation is possible. Within these groupings there are undoubtedly differences of social situation (by precise job task, or status of employer, for example), about which we simply lack the necessary information.

4.5
Some process of aggregation was inevitable, because of the need to ensure that there were a sufficient number of individuals in each occupational group to make the necessary relational pair comparisons. There was also a trade-off between the numbers available and the fineness of the time periods into which our data could be broken down. The compromise was to split the period into two parts with roughly equal numbers, as had been done in the case of the scales for men: using all occupational matches taking place between 1777 and 1866 (the median year being 1851) and those taking place between 1867 and 1913 (median year 1888)1[9].

4.6
The basis for the analysis was to take each female occupational group and to calculate the mean occupational score of all the male relatives of women in that occupation. This gives three sets of scores (and thus three orderings) for the female occupational groups, derived from the relationship of the incumbents to fathers, husbands and fathers-in-law. We can regard each set of scores as a different 'version' of the same underlying structure – the social ordering of female occupations, which is expressed in the pattern of social relationships associated with that hierarchy. If these very different relationships (to father, husband, and father-in-law) do express the same underlying structure then we would expect the three 'versions' to be very similar to each other.

4.7
In fact, the three sets of occupational scores are extremely well correlated with each other[10],with great similarity in the pattern of occupations held by husbands, fathers and fathers-in-law. This indicates that all three sets of relationships are embedded in the same underlying social hierarchy. Factor analysis of the three sets of scores revealed that (for both periods) a single factor accounted for 95% of the variance, whilst the factor score coefficient matrix indicated that in the analysis the three different 'versions' each carried an equal weighting. This confirms the underlying similarity of the three 'versions' of the occupational ordering. The overall score for the female occupations was constructed by simply adding the solutions given by the three sets of relationships (fathers, husbands, and fathers-in-law), giving each an equal weighting (see Figure 2). The ordered occupations were then given a hierarchical scoring.

4.8
In producing these measures of social distance for two time periods, we would expect there to be a good relationship between the measures over time. There will inevitably be some difference, with changes either in the composition of occupational groups or a rise (or fall) in their social position. On the other hand, there should be a substantial amount of continuity in the social order. In fact, the two measures are very well associated. The two scales are highly correlated1[11] and show a considerable degree of continuity in the structure of social distance over time. Table 1 shows the occupational scales for the two time periods.

4.9
Having produced these measures of social distance amongst women in the nineteenth century, we must next ask – what is the nature of the social order that is revealed by this exercise? The nature of the dimension that emerges in each period is, as we have seen, very similar. At one end are property owners and other elite groups; at the other farm workers and labourers. In that respect, the hierarchy revealed by the patterns of social interaction conforms to the usual expectations and demonstrates considerable 'face' validity.


Table 1: Women's Occupational Scale Scores, 1777-1866 & 1867-1913

First Period,
1777- 1866

Score

Second Period,
1867-1913

Score

Independents

99

Professionals

99

** incl.. Professionals


Farmers

98

Farmers (nmi)

97

** incl. Small farmers


Clerks

95

Farmers' wives

92

Farmers' wives

91

** incl. Farmers' daughters


Governesses

87

Independents

86

Farmers' daughters

86

Governesses

85

Small farmers

85

Music teachers

83

Teachers

81

Clerks

82

** incl. Music Teachers


Teachers

81

Barmaids

80

Milliners

75

** incl. Waitresses


** incl. Hat workers


Shops

72

Shops

71

Book binders

70

Other crafts

68

Food shopkeepers

69

Nurses

67

** incl. Innkeepers


Barmaids

67

Non-food shopkeepers

68

Waitresses

65

** incl. Dealers


Innkeepers

65

Milliners

65

Non-food shopkeepers

62

Hat workers

63

Dealers

62

Housekeepers

56

Food shopkeepers

58

Ladies' maids

55

Book binders

56

Bakers

55

Tailoresses

54

Dressmakers

52

Childrens' nurses

52

Miscellaneous crafts

51

Housekeepers

51

Tailoresses

48

Miliners/food workers

51

Textile workers

47

Dressmakers

51

Textile finishers

46

Seamstresses

51

Maids

46

Ladies' maids

48

** incl. Children's nurses


Garment trades

43

** incl. Housemaids


Factory hands (not textile)

40

** incl. Parlourmaids)


Housemaids

39

Shoe/leather workers

46

Cooks

36

Nurses

46

Spinners

36

Seamstresses

43

Maids

33

Cooks

43

Weavers

32

Winders/piecers

38

Textile finishers

30

Spinners

37

Combers

29

Weavers

35

Winders/piecers

28

Garment trades

35

Knitters

28

Laundrywoman

30

Textile workers

27

Knitters

28

Farm workers

27

Servants

26

Servants

26

Combers

25

Parlourmaids

25

Metal trades

25

Shoe/leather workers

24

Factory hands (not textile)

19

Laundrywomen

21

Farm servants

15

Metal trades

21

Miscellaneous unskilled

13

Farm servants

21

Farm workers

13

Miscellaneous unskilled

19

Straw plaiters

11

Labourers

14

Lace workers

7

Lace workers

8

Agricultural labourers

4

Agricultural labourers

5

Labourers

1

Straw plaiters

1

4.10
At the top end of the scale are property owners (independents – comprising employers and rentiers, farmers, and farmer's wives). These titles were predominantly held by married women. Other groups at the top of the hierarchy (held mainly by women on the point of marriage) include governesses, clerks and teachers (including music teachers), as well as, in the second period, professionals (a small group dominated by missionaries). Shop workers are near the top of the hierarchy, in keeping with the early craft history of retailing and the 'respectable' nature of female counter assistants at this time. Following these groups come skilled craft workers (milliners, book binders) and retailers. Here also are innkeepers, shopkeepers and dealers. These groups fall relatively low in the social ranking, reflecting the preponderance of small, or low status, establishments amongst the women retailers. Food shopkeepers (which in the male scale fall significantly below non-food shopkeepers) are, in the female scale, not distinguished in social position from non-food shopkeepers, again probably the result of the relatively small scale of the establishments concerned (slop sellers, pawnbrokers, market stall holders etc.).

4.11
A further interesting feature is the relatively high ranking of waitresses and barmaids, alongside the female inn and shopkeepers. We might expect the assistants in such enterprises to rank well below the owners and managers. However, the family patterns of the waitresses and barmaids in our sample reveal that they are often either the daughters of publicans and restauranteurs, or else subsequently marry such men (and sometimes both), so that their ranking alongside female publicans etc. is less surprising. This feature may be a peculiarity of our sample (our male sample is overweighted to such occupations, so we are more likely to pick up their female relatives). Relatively few women occur in these groups, and it may be the case that lower level workers were given other titles (such as inn servant for example).

4.12
Domestic service occupations fall into the lower half of the scale. However, differences in the social position of the various occupations within the broad realm of domestic service provide further 'face' validity for the ordering of groups given by the social interaction technique. 'Upper servants' (housekeepers, ladies' maids etc.) rank significantly above plain 'servants' or 'maids' in the scale, whilst 'farm servants' are significantly lower, lying at the bottom of the hierarchy. There are some discrepancies. In the second period, for example, 'housemaids' rank above 'parlourmaids', for reasons which may simply be artefactual. 'Cooks' rank around the middle of the domestic servants, below, for example, 'housekeeper'. However, this may reflect the heterogeneity covered by the title 'cook', which includes cooks from outside domestic service (in restaurants, industrial food production etc.). 'Governesses' rank much higher than any of these other domestic service groups, being associated with elite groups such as farmers and farmers' wives. This confirms the idea that 'governess', like 'teacher' and 'milliner', represent those few occupations which could be held by women from 'genteel' families.

4.13
At the lower end of the scale are workers in the garment trades, textile and other industries, with farm workers, farm servants, unskilled workers and labourers bringing up the rear. Lace making and straw plaiting are both at the very bottom of the hierarchy, alongside labourers in industry and agriculture. These trades were often undertaken as outwork in areas with limited employment for women and in agricultural districts, often introduced by Poor Law authorities to relieve pressure on the poor rate (Hill, 1989: 67). Women in the 'lace making' and straw plaiting' groups in our sample were predominantly older women married to farm and other labourers, suggesting that these occupations were trades undertaken in hardship.

The Pattern of Husbands' and Wives' Occupations

5.1
We can explore the patterning of male and female relationships by looking in more detail at of just one of the relationships – the pattern of husbands' and wives' occupations1[12]. This builds up a picture of who is marrying whom and allows us to ask the question: which male occupations are socially equivalent to female occupations?


Table 2: Pattern of Husbands and Wives Occupations (First Period, 1776-1866)


Female
Occupational
Groups

Husband's Occupational Group
(Occupations around median score for husbands in that group)

Standard Deviation of male scores

Independent

Manufacturer, Professionals, Dealers, Employers, Non-food Shopkeepers

24.8

Farmer (Nmi)

Farmers (83%), Managers, Builders, Food Shopkeepers, Innkeepers

17.44

Clerk

Cash Clerks, Cabinet makers, Printers, Farmers, Non- Food Shopkeepers (31%)

18.31

Farmer's Wife

Farmers (80%), Builder [ & farmer], Innkeeper [ & farmer]

16.31

Governess

Farmers, Builders, Butchers, Tailors, Cabinet Makers

21.43

Farmers' Daughter

Clerks, Farmers (50%), Builders, Hat/glove Maker, Curriers/tanners

21.55

Small Farmer

Farmers & Small Farmers (37%), Fishermen (47%)

12.00

Teacher

Farmers, Butchers, Cabinet Makers, Clockmakers, Transport Owners

22.96

Barmaid

Farmers, Innkeepers (25%), Butchers, Ship's Officers

16.74

Shop Worker

Farmers (nmi & medium), Painters, Mechanics, Plumbers, Brewers, Joiners

20.84

Book Binder

Non-food shopkeepers, Bookbinders (23%), Cordwainers (20%), Mechanics, Joiners, Carpenters, Paper makers

13.11

Inn-/food shopkeeper

Builders, Food Shopkeepers, Innkeepers, Shoemaker, Cabinet Makers, Tailors

17.66

Non-food shopkeeper

Shopkeeper, Millers, Shoemakers, Bakers, Warehousemen, Joiners

22.97

Milliner

Shopkeeper (6%), Tailors (9%), Plumbers, Cordwainers, Mechanics, Butchers

21.62

Hat Worker

Hat maker (14%), Painters, Mechanics, Spinners, Tailors, Millers

21.79

Housekeeper

Cordwainers, Farmers (22%), Mechanics, Painters, Smiths

24.93

Baker

Bakers (21%), Butchers (11%), Cordwainers, Tailors, Cabinet makers, Shoemakers

18.84

Lady's Maid

Butler, Carpenters, Coachmen (12%), Groom, Cordwainers

17.85

Dressmaker

Joiners, Textile finishers, Carpenters, Shipwrights, Spinners, Shoemakers

19.12

Misc Crafts

Carpenters, Spinners, Painters, Shoemakers, Gardeners, Transport Owners

25.53

Tailoress

Carpenters, Tailors (22%), Spinners, Shoemakers, Seamen

17.34

Textile Worker

Spinners (11%), Shoemakers, Misc Non-skilled, Masons, Seamen

18.69

Maid

Seamen, Smiths, Coachmen, Shoemakers, Sawyers, Carpenters

20.77

Shoe/leather Worker

Joiners, Shoemakers (55%), Shipwrights, Sawyers, Seamen

13.58

Textile Finisher

Textile finishers (18%), Spinners, Weavers, Railway Porters, Masons, Smiths

9.37

Nurse

Gardener, Shoemaker, Carpenters, Smiths, Joiners

23.98

Cook

Carpenters, Coachmen, Wheelwrights, Smiths, Shoemakers, Masons, Seamen

17.83

Seamstress

Smiths, Weavers, Sawyers, Personal Service, Seamen, Gardeners, Carpenters

21.47

Winder/piecer

Weavers (12%), Spinners (14%), Smiths, Seamen, Gardeners, Railway Porters

19.06

Spinner

Miners, Colliers (13%), Combers (10%), Weavers (10%), Masons

20.12

Garment Trades

Gardeners, Misc. Non-skilled, Weavers, Smiths, Fishermen, Servants

21.81

Weaver

Weavers (26%),Spinners, Gardeners, Misc. Non- skilled, Smiths, Fishermen

16.87

Laundrywoman

Carpenters, Gardeners, Misc. Non-skilled, Weavers, Smiths

18.87

Knitter

Knitters (47%), Masons, Fishermen, Coal miners

14.38

Servant

Servant, Gardeners, Misc. Non-skilled, Weavers, Smiths, Farm workers

21.74

Metal Trades

Weavers, Smiths, Carters, Coal Miners, Cutlers

19.10

Comber

Combers (16%), Coal Miners, Colliers, Misc. Non- skilled, Weavers (13%)

12.47

Factory Worker

Gardeners, Misc. Non-skilled, Masons, Shoemakers, Paper/chemicals (12%)

18.89

Farm Servant

Farm workers (33%), Coal miners, Miners, Brickmakers

17.29

Farm Worker

Farm workers (33%), Coal miners, Animal Workers, Fishermen (13%)

19.85

Misc. Unskilled

Carters, Colliers, Coal miners (16%), Watermen, Farm workers

23.32

Straw Plaiter

Farm Worker (31%), Labourers (47%), Bricklayers

20.30

Lace Worker

Farm Worker (27%), Labourers (36%), Metal Workers,, Watermen

18.06

Agricultural Labourer

Farm worker (38%), Labourers (31%), Animal Workers

16.96

Labourer

Labourers (65%), Farm Worker, Carters

19.12

5.2
Table 2 shows the pattern of husbands occupations by each female occupational group. For brevity, only the pattern of relationships in the first period is discussed. The table was constructed by taking the median score of husbands' occupations for each female occupational group, and then listing occupations which occurred in the marriage pairs around that median score. The median was chosen to represent the typical or average level of job actually held by the women's husbands in each female occupational group. We look at the typical husband's score in order to assess which occupational groups can be regarded as socially equivalent to each other, as revealed by the patterns of assortative marriage.

5.3
The table indicates that the male and female occupational structures are closely related, but do not map onto each other exactly. Obviously, given the way in which the female scale was constructed, as we go down the female occupational hierarchy there is a decline in the overall ranking of husbands' jobs. Within this overall pattern, there is a great deal of both occupational homogamy and marriage linkage between jobs strongly related by industry. So, for example, women listed as farmers or shopkeepers are typically married to male farmers or shopkeepers (in the sense that these occupations are close to the median or average score for husbands' occupations for that group). As you would expect during the first half of the nineteenth century (when production and commerce were still dominated by family enterprise), we find a good deal of what appear to be 'husband and wife' joint enterprises. Such jobs are, naturally, similarly ranked in the male and female occupational orderings and strongly linked by marriage patterns. Shoemakers, for example, turn up amongst the median scored husbands of female 'shoe/leather workers'. We can regard such jobs as being socially equivalent by virtue of their marriage patterns, although we might expect equivalence simply on the basis of job title alone.

5.4
However, the links between spouses' jobs are not simply the result of husband and wife businesses, but are also the product of family connections, contiguity, workplace and neighbourhood association and so on. When we find male shoemakers married to female shoemakers this need not reflect a joint business, it can also reflect social processes which lead women working as shoemakers to be more likely to meet and marry men working as shoemakers. For example, for females with domestic service occupations ('Maids' and 'Lady's Maids') we find that 'coachmen' crop up very frequently amongst the median scored husbands' occupations. The fact of marriage links, and social equivalence, between women and men in domestic service is not surprising. However, it is interesting to see where female servants have looked for husbands when they have married outside domestic service. 'Maids', for example, typically marry men from around the social level of smiths, carpenters, and shoemakers – artisanal or skilled trades.

5.5
This helps to establish links of equivalence between the male and female occupational orders. It is also interesting to see how social hierarchy within the male and female orders plays out. So whilst we might expect female clerks to marry male clerks, it is interesting to see that they also marry printers, shopkeepers and cabinet makers. The patterns of social equivalence that we can see established here are between female white-collar work and highly skilled artisans and tradesmen.

5.6
At the very bottom of the hierarchy, as at the top, we can see a great similarity in the pattern of husbands' and wives' occupations: just as the propertied marry the propertied, so farm labourers marry farm labourers. This is what we might expect. However, it is in the middle of the social order, the great mass of intervening ranks, that we see much more complex patterns of association. It has always been much more difficult to establish links of equivalence between the 'middling orders' of the male and female occupational structures. By making patterns of social interaction (and distance) central to our analysis of female occupations, we are able to explore these patterns of social equivalence in much greater detail.

Conclusion

'The body-makers are the first on the list; then follow the carriage makers; then the trimmers; then the smiths; then the spring-makers; then the wheelwrights, painters, platers, brace-makers; and so on. . . . . A body-maker is considered a "good catch" as a husband for the daughter of an ordinary mechanic; and the carriage maker excited much anxious feeling on the part of mothers, who consider marrying to a carriage maker as important a matter as vulgar-minded mothers in the classes just above them consider "marrying to a carriage"'. (William Bridges Adams, English Pleasure Carriages (1837); cited in May, 1987: 233.)

6.1
This paper offers a very different approach to the understanding of the social ordering of women's occupations from that provided by the conventional one in terms of social classes. The method outlined here puts social relationships at the heart of stratification analysis. The quotation nicely illustrates two points about the members of British society in the nineteenth century; they were well aware of the subtle gradations that distinguished occupational groups from one another and this showed itself particularly through marriage patterns. However, such concerns not only reflect a view of the social order; since the selection of a spouse is also an aspect of the reproduction of that order. These choices are underpinned by a range of structural constraints on social interaction and lifestyle that make choice of an 'appropriate' marriage partner a normal and natural act.

6.2
We have had to use the relation of women's occupations to those of men in order to develop a better understanding of the social structuring of the former, but we should emphasise that this is very different from allocating a position to individual women on the basis of their husbands' (or fathers') occupation. Rather than give individual women the class of their male kin by proxy, we have instead measured the location of women's occupations by the patterning of the social relations of the women in that occupation. For both women and men (albeit with different techniques) it is the social distance of occupations, as given by interaction patterns, which has been used to order them.

6.3
It should be emphasised that although the ordering of occupations is revealed by social processes that tend towards endogamy (that like tends to marry like), the method discussed here is only reliant on this as an average tendency. Homogamy - inter-marriage within occupational groups - is not a requirement of the technique and is not assumed in the analysis. Indeed, in general terms, social interaction approaches to measuring the occupational order rely on there being a spread in the background of marriage partners, since if marriage occurred entirely within occupational groups we could not establish the relations between groups. It has been established that relatively weak associations between the occupations of husbands and wives can still be used to establish the patterns of social distance, and thus the social ordering, of occupations (Prandy and Jones, 2000). The (adapted) technique in this paper establishes the social position of aggregated female occupational groups by averaging the hierarchical position of the male relatives of women in the group. What is produced is a measure of the social position of the occupation taken as a whole, which inevitably will have a dispersion around the mean. We can thus still explore the extent to which individual women within occupational groups have experienced mobility, by comparing their occupational position to that of their male (and female) relatives. We can also explore the extent to which occupational groups collectively experience shifts in their social position by comparing their hierarchical position across the two time periods.

6.4
The scales for the two periods show considerable consistency. However, there are some changes in the placing of particular occupational groups, which we can link to changes in the social context and meaning of certain jobs with the development of more 'modern' work processes. 'Nurses', for example, rank relatively low in the hierarchy in the first period, but rise considerably in the second, to rank just below professional groups such as teachers. This is as we would expect, given the increasing professionalisation of nursing. Similarly, we can see a drop in the ranking of 'shoe/leather workers' over time. Again, this coincides with a drop in the numbers of family-based independent shoe makers and the rise of industrial factory production. Insofar as labour market shifts have real consequences for the social relationships of the incumbents of jobs (and thus can be regarded as meaningful in social terms) the social interaction approach, which orders occupational groups by social relations alone, thus reveals patterns of collective mobility.

6.5
The scales for the two periods indicate the success of the social interaction approach to stratification. The network of links between the occupations of male and female relatives show a very strong patterning that can be used to reveal the social hierarchy of women's occupations. Ultimately, however, the usefulness of these social interaction scales lies in what they can tell us about women's lives. Whilst there is wealth of detailed empirical studies on particular female occupations, there is comparatively little work on the female occupational ordering considered as a whole. The mapping of the female occupational hierarchy produced here provides a much more detailed account of the social distances and inter-relations between finely demarcated occupational groups. There is a related question, however, which is how does the social ordering of women's occupations in the nineteenth century relate to women's wider social position? We intend to use the information on women occupational scales to explore the influence of women in processes of social mobility and the transmission of advantage and disadvantage. It should also be possible to use female occupational scales to examine patterns of illness and disease. There is not the space to describe such work here, but we have explored what the female scales reveal about the relationship between the male and female occupational orders and the patterns of social linkage and equivalence that underpin this. Being able to place women's occupations in relation to the male occupational hierarchy helps to position women's occupational inequality in a broader social framework.

6.6
The particular nature of women's occupations in the nineteenth century raises difficulties, but we believe that our method remains a robust approach to analysing women's occupational experience. Our approach retains, as far as possible, the fine distinctions between women's occupations, distinctions which are particularly important given the crowding of women's work. It is also clear that the intermittent or seasonal nature of women's work is less of a problem for a relational technique. It is not the contribution of a woman's job to her overall social position or its significance in her working life which counts, but rather the overall social location of the occupation itself. So it does not matter that any individual woman works as a 'straw plaiter' only seasonally, or for a short time, since what we are trying to assess is the network of relationships in which 'straw plaiters' as a group are placed. If it is routine and typical for 'straw plaiters' to be closely related to 'farm labourers', then we can begin to place such groups in a social ordering of occupations. Our procedure allocates a woman on the basis of the average position (in terms of relationships to men) of women in her occupation. So, men at a given level will not all have wives who are, by definition, at the same level. Our procedure reflects far more effectively the social reality that there will be range of values for the wives of men at any particular level, a range that reflects both the wives' social backgrounds and their own achievements.

6.7
It is important that women's occupational experience be included in accounts of the historical social hierarchy. By using a social interaction approach, the social relationships of both men and women are given equal attention and are subject to the same theoretical principles. In the process, we can address more directly both the relationship between the male and female occupational orders, and the contribution of women's experience to the reproduction of social hierarchy.

Notes

1Mitchell and Critchley looked at the hierarchical ordering produced by a multi-dimensional scaling of the occupations held in three sets of relations – between male respondents and their fathers, fathers-in-law, and (male) spare- time associates – and found that 'the basic structure is essentially the same in all three cases' (Mitchell & Critchley, 1985: 72). Similarly, Hout examined the relationship between the occupations of husbands and wives in dual earner households and found a very close relationship between the pairs of occupations (Hout, 1982).

2A later amendment refers instead to the materially 'dominant' occupation within a household (Erikson, 1984). This can be held by either gender although, in practice, it is generally held by a man. However, the 'dominance' of an occupation is a question of degree, which raises the question of how 'subsidiary' occupations within a household affect the material conditions and social outlook of all household members.

3In arguing that close social relationships, such as marriage, occur in situations of social similarity and involve shared levels of lifestyle we are not forgetting that marriage may involve considerable disparities in access to resources and in status within the household. The method used here can say nothing directly about gender inequality within households – which requires a much finer level of analysis.

4 The same applies to the social significance of jobs within an individual career. Generally, movement from less skilled and well paid jobs to better paid and skilled jobs is regarded as upward career mobility. However, if movement between two jobs (say 'clerk' and 'manager') is routine and typical (such that 'clerks' can reasonably expect to become 'managers' in due course) then it is not clear in what sense real social 'mobility' is taking place (Stewart et al., 1980). If the jobs are linked by the routine and predictable career transitions of the same individuals, then such jobs are socially commensurate despite the gap in rewards. The labour market deficit does not translate into a straightforward social divide. For example, if a 'manager' father has a son who becomes a 'clerk' this might look like downward mobility. However, if that son then rises up to the position of 'manager' (and this is a typical transition) then the mobility (both downward and upward) is more apparent than real.

5 Our sample consists of 80,000 individuals whose occupations were discovered by family historians researching their family tree, raising issues of representativeness. Assessment of the sample is made difficult by the nature of nineteenth-century official statistics on occupation since the early census (until 1911) categorised occupations by industry. Such schemes did not make distinctions of skill and authority which are central to modern classifications (Szreter, 1996; Higgs, 1991). We have coded the occupations of our sample to the census classification for each decade, permitting a broad comparison by industry, and the figures, for both men and women, correspond reasonably well with the distributions given in the censuses.

6As Diagram 1 indicates, there would be three 'versions' of the male occupational ordering for each of the three men and three 'versions' of the female occupational ordering (given by the pairings with their husband, father and father-in-law at the time of marriage). The nine 'versions' of the male occupational ordering and the three 'versions' of the female one can be regarded as estimates of the same underlying (related) social structures, and can be combined to build up a picture of those structures.

7The analysis on occupational pairings has been done for two time periods: 1776-1866 and 1867-1913. There were far fewer female to male occupational pairs than male to male pairs.

Father of groom to Father of bride pairs: 5208 pairs (1776-1866), 5046 pairs (1867-1913)

Father (of groom) to Groom pairs: 7760 pairs (1776-1866), 6787 pairs (1867- 1913)

Father-in-law to Groom pairs: 6896 pairs (1776-1866), 6668 pairs (1867-1913)

Father-in-law to Bride pairs: 1416 pairs (1776-1866), 2113 pairs (1867-1913)

Father (of bride) to Bride pairs: 1548 pairs (1776-1866), 2077 pairs (1867- 1913)

Groom to Bride pairs: 2944 pairs (1776-1866), 3182 pairs (1867-1913).

8 In a discussion of the methodology of researching assortative marriage, Kalmijn argues that only the occupations of newlyweds should be considered, to avoid capturing the effects that longer-term spouses have on each others careers, and to avoid the biases that may arise from attrition due to separation and divorce (homogamy rates are inversely related to the risk of divorce) (Kalmijn, 1994). Notwithstanding the different rates of separation and divorce between modern and historical samples, we believe that taking the occupations of longer-term spouses is fully justified, since what is being established in our procedure is the social equivalence of aggregate pairs of occupations. If certain pairings of occupations held by spouses are less likely to give rise to divorce or separation than others, then this is a feature of the long-term social equivalence of occupations and, as such, is a feature to be incorporated into analysis rather than a factor to be controlled. Similarly, if, over the long term course of a marriage, spouses affect the social location of their partner (so that occupations of discrepant status or position are matched), this will only show up in the aggregate pattern if this is a large scale social effect, in which occupations have, by virtue of consistent marriage links, become socially commensurate. To the extent that long-term marriages indicate aggregate or routine social links between occupations, they are vital features of the map of social equivalence that is being measured.

9The analyses for the first period were based on 9,340 occupational matches, using 45 female and 83 male occupational groups; for the second period based on 10,818 occupational matches, using 50 female and 91 male occupational groups. Occupational Pairings for Construction of Female Scores

Husband to Wife pairs: 5574 pairs (1776-1866), 5662 pairs (1867-1913)

Father to Daughter pairs: 1848 pairs (1776-1866), 2570 pairs (1867-1913)

Father-in-law to Daughter-in-law pairs: 1918 pairs (1776-1866), 2586 pairs (1867-1913)

5574 pairs (1776-1866), 5662 pairs (1867-1913)

10Product moment correlations of 0.93 and above for the first period analysed, 0.89 and above for the second period; rank order correlations of 0.92 and above for the first period, 0.90 for the second period.

11Product moment and Spearman rank order correlation coefficients of 0.92 and 0.91 respectively.

12The three sets of scores given by the women's relations to their husbands, fathers and fathers-in-law were highly correlated, so any examination of the commensurability of the husbands' occupations should also apply to male occupational order generally.

Acknowledgements

The work reported in this paper was undertaken with funding for two projects from the Economic and Social Research Council: The Family Occupation and Social Stratification: 1840 to the Present (R000235147) and The Family, Occupation and Social Stratification in Scotland and Ireland (R000221729). The support of the Council is gratefully acknowledged. The authors also wish to thank the three anonymous referees for their helpful comments.

References

BAKKER, B. F. M. (1993) 'A new measure of social status for men and women: the social distance scale' in Netherlands Journal of Social Sciences, 29: 113- 129.

BOURDIEU, P. (1987) 'What makes a social class? On the theoretical and practical existence of groups', Berkeley Journal of Sociology, 32:1-17.

CROSSICK, G. (1978) An Artisan Elite in Victorian Society: Kentish London 1840-1880, London: Croom Helm.

ERIKSON, R. (1984) 'Social class of men, women and families', Sociology, 18 (4): 500-515.

FOSTER, J. (1974) Class Struggle and the Industrial Revolution, London: Weidenfield and Nicolson.

GRAY, R. Q. (1976) The Labour Aristocracy in Victorian Edinburgh, Oxford: Oxford University Press.

GREENACRE, M.J. (1984) Theory and Applications of Correspondence Analysis. London: Academic Press.

HEATH, A. & BRITTEN, N. (1984) 'Women's jobs do make a difference', Sociology, 18 (4), 475-90.

HIGGS, E. (1987) 'Women, occupations and work in the nineteenth century censuses', History Workshop Journal, 23: 59-80.

HIGGS, E. (1991) 'Disease, febrile poisons and statistics: the census as medical survey, 1841-1911', Social History of Medicine, 4: 465-78.

HILL, B. (1989) Women, Work and Sexual Politics in Eighteenth- century England, Oxford: Basil Blackwell.

HOUT, M. (1982) 'The association between husbands' and wives' occupations in two-earner families' in American Journal of Sociology, 88 (2) 397-409.

KALMIJN, M. (1991) 'Status homogamy in the United States' in American Journal of Sociology, 97 (3) 496-523.

KALMIJN, M. (1994) 'Assortative mating by cultural and economic occupational status' in American Journal of Sociology, 100 (2) 422- 452.

KALMIJN, M. (1998) 'Intermarriage and homogamy: causes, patterns, trends' in American Review of Sociology (24) 395-421.

KENDALL, D. G. (1971) 'Maps from marriages: an application of non-metric multi-dimensional scaling to parish register data' in F.R. Hodson, D. G Kendall and P. Tautu (eds.) Mathematics in the Archaeological and Historical Sciences, Edinburgh: Edinburgh University Press.

LAUMANN, E. O. (1966) Prestige and Association in an Urban Community: An analysis of an urban stratification system. Indianopolis: Bobbs-Merrill Inc.

LAUMANN, E. O. and GUTTMANN, L. (1966) 'The relative associational contiguity of occupations in an urban setting' in American Sociological Review, 31 (2)169-178.

LEVINE, J. H. (1990) 'Measuring occupational stratification using log-linear distance models' in R. L. Breiger (ed.) Social Mobility and Stratification, Cambridge: Cambridge University Press.

LINDERT, P. (1980) 'English occupations, 1670-1811', Journal of Economic History, 40 (4): 685-712.

MAY, T. (1987) An Economic and Social History of Britain 1760- 1970, Harlow: Longman.

MILES, A. (1993) 'How open was nineteenth-century British society? Social mobility and equality of opportunity, 1839-1914', in A. Miles and D. Vincent, Building European Society: Occupational change and social mobility in Europe', Manchester: Manchester University Press.

MITCH, D. (1993) '"Inequalities which everyone may remove": occupational recruitment, endogamy, and the homogeneity of social origins in Victorian England', in A. Miles and D. Vincent, Building European Society: Occupational change and social mobility in Europe', Manchester: Manchester University Press.

MITCHELL, J.C. and CRITCHLEY, F. (1985) 'Configurational similarity in three class contexts in British society, Sociology, 19 (1) 72-92.

PENN, R. (1985) Skilled Workers in the Class Structure, Cambridge: Cambridge University Press.

PENN, R. and DAWKINS, D. (1983) 'Structural transformations in the British Class Structure: a log linear analysis of marital endogamy in Rochdale 1856-1964' in Sociology, 17 (4) 506-523.

PRANDY, K., (1990), 'The revised Cambridge scale of occupations', Sociology, 24 (4), 629-55.

PRANDY, K. (1999) 'The social interaction approach to the measurement and analysis of social stratification' in International Journal of Sociology and Social Policy, 19 (9/10/11) 215-247.

PRANDY, K. and BOTTERO, W. (1998) 'The use of marriage data to measure the social order in nineteenth century Britain', Sociological Research Online, 3. <http://www.socreso nline.org. uk/socresonline/3/1/6.html>

PRANDY, K. and BOTTERO, W. (2000) 'Social reproduction and mobility in Britain and Ireland in the Nineteenth and Early Twentieth Centuries', Sociology, 34, 2: 265-281.

PRANDY, K. and JONES, F. J. (2000) 'An international comparative analysis of marriage patterns and social stratification' paper presented at the Conference of ISA Research Committee 28 on Social Stratification and Mobility, Libourne, May 11-14 2000.

RYTINA, S. (1992) 'Scaling the intergenerational continuity of occupation: is occupational inheritance ascriptive after all?' in American Journal of Sociology, 97 (6) 1658-1688.

SNELL, K.D.M. (1985) Annals of the Labouring Poor: social change and agrarian England, 1660-1900, Cambridge: Cambridge University Press.

STANWORTH, M. (1984) 'Women and class analysis: a reply to John Goldthorpe', Sociology, 18 (2): 159-170.

STEWART, A., PRANDY, K. and BLACKBURN, R.M., (1980), Social Stratification and Occupations, London: Macmillan.

SZRETER, S. (1996) Fertility, Class and Gender in Britain 1860-1940, Cambridge, Cambridge University Press.

WELLER, S.C. and ROMNEY, A.K. (1990) Metric Scaling. Newbury Park, CA: Sage.

Copyright Sociological Research Online, 2001