The composition of households has changed drastically in many western societies in recent decades. In particular, there has been a significant increase in the numbers of people living on their own and in the numbers of children being brought up by a single parent. Nevertheless, a comparison of present-day households with those of pre-industrial communities, as set out in the first part of the paper, reveals some unexpected similarities. Then, as now, households included few relatives other than parents with their unmarried children. It has often been assumed that families in the past were more caring, more prepared to take needy relatives into their homes. Careful comparisons suggest that this view is simplistic. Nor is there any evidence that the present-day incidence of one-parent families (20% of households with children) is in any way exceptional. The increased incidence of divorce has in the twentieth century increased the proportion of one-parent families, as did early widowhood in the seventeenth century. The intervening centuries were if anything the exceptions: improving mortality and earlier marriage combined to ensure that fewer children would be brought up in a lone-parent family. On the other hand present-day households do differ from those of the past in that they are smaller, and now rarely contain servants and lodgers. In addition, families now contain fewer children, although these children represent a similar proportion of the total membership of the household. There has also been a dramatic increase in the number of people living alone, particularly elderly women. Some of these changes are of fairly recent origin. Table 4, for example, suggests that the composition of UK households containing elderly persons changed as much between the 1960s and the 1980s as between the pre-industrial era and the 1960s.
In the second part of the paper we focus on the variation across countries in household composition in present-day Europe. Southern and Eastern European countries are more likely than those in Scandinavia or Western Europe to have households containing distant relatives (relatives other than spouses and unmarried children). Southern and Eastern Europe also have much lower proportions of one-person households: one in ten households in Cyprus, for example, and one in five in Italy contain but one person, compared with one in three households in Sweden. The latest round of censuses conducted in the early 1980s has produced no evidence of a reduction in cross-country variations in the frequency of persons living alone. A more detailed analysis also reveals that the effects of age and gender on the numbers living alone vary across countries. In general women heavily outnumber men as solitary householders but there are, for example, relatively more men living on their own in Scandinavia and Switzerland than in much of Western Europe. It is far from easy to explain these patterns, although demographic, economic and cultural factors all merit consideration. Demographic changes have modified the age structure of the population in favour of those age groups where people would be most prone to live alone, while economic changes have made living alone more practical and cultural changes have made it more acceptable. Table 9 indicates, however, that there is little evidence in the national data that a higher proportion of people living alone is associated with high levels of consumer expenditure.
European countries also differ according to the age at which children leave the parental home. Figure 9 indicates that daughters generally leave the parental home at a younger age than do sons, but that this tendency is much more marked in Sweden and Finland than in France and Ireland. More detailed work on the life cycle is, however, hampered by the failure to agree on standard definitions of such basic concepts as family< and child and by inconsistencies among countries in defining the role of individuals within the household and in the format and level of detail of their tabulations of household type. The third section of the paper suggests that research efforts should now be directed towards the production of a standard set of tabulations that could be applied to any population. The fundamental features of household and family forms in any one national population can only be fully understood in the context of a perspective that is both historical and international. This will, however, require international cooperation in order to refine the necessary standard tables on the classification of households.