The role of the wedding place. Community context and marital timing in nineteenth and early twentieth century Netherlands

Abstract

This study investigates how community characteristics influenced the timing of marriage of men and women in nineteenth and early twentieth century Netherlands on the basis of a large scale database consisting of marriage certificates covering five provinces of the Netherlands between 1840 and 1922. The results show the significance of religious context for understanding marriage timing in the nineteenth century. Living in a predominantly Catholic community resulted in a later marriage for both men and women, while living in a community that was dominated by Orthodox Protestants resulted in an earlier marriage, particularly for men. In addition, residence in a municipality with a high mobility, a large population size and a high birth rate speeded up marriage timing among both men and women. The results indicate that religious restraint and the urbanization and openness of places are, next to parental social class, of vast importance for understanding marriage timing. As our study only addressed those who married, future research will have to show whether the same mechanisms were at work for those that experienced permanent celibacy.

Notes

1 In this study we focus on only one of the two characteristics of the ‘Western European Marriage pattern’, namely the age at marriage, while excluding in large part the second characteristic of this system, namely the high incidence of permanent celibacy. The relationship between age at marriage and the occurrence of permanent celibacy is contested. Dixon (1971) studied this interrelation on the basis of an analysis of 57 countries around 1960 and found that although these two are interrelated at the aggregate level, it is not pre-determined that in countries where the average age at marriage is high, more individuals remain unmarried (and vice versa). In a follow-up study of Dixon (1978) on the interrelation between timing and prevalence of marriage involving a case study of Japan and Ireland that do not satisfy the expected positive relation between age at marriage and the occurrence of permanent celibacy, she found that marriage delays can be explained mainly by economic factors, whereas permanent celibacy also reflects desirability of marriage in terms of norms concerning spinsters and bachelors as well as employment and career opportunities. Furthermore, other research has also suggested that age at marriage and permanent celibacy are partially subject to different mechanisms (Engelen & Kok, 2003; Gjerde & McCants, 1999). In the Netherlands, the link between age at marriage and permanent celibacy differed between subgroups (Engelen & Kok, 2003). First, for some individuals, permanent celibacy was the result of a delay of marriage, most notably seen among younger siblings who took care of parents until a relatively late age. Secondly, a low age at marriage and a high level of permanent celibacy was found in relatively wealthy groups in which a deliberate choice between marriage and non-marriage could be made on the basis of life plans and style. In the Eastern part of the country, almost all individuals married late, not only younger siblings whom had to wait very long, but proportions single were small as well. Finally, in the labor classes where incomes rose and parental control weakened in the nineteenth century, both the age at marriage and permanent celibacy were low. We can therefore conclude that although age at marriage and the occurrence of permanent celibacy are interrelated, they are not completely subject to the same mechanisms and thus results might change if never-married were included in the sample.

2 In order to test whether endogeneity might disturb the results of our model due to the incorporation of birth rate at the community-level, we have done the analysis without the birth rate variable. We found that none of the coefficients on the individual, family and community level (for men and women) changed substantially as a result of the exclusion of the birth rate variable from the model and that there were no substantial differences in terms of effect size, direction of the effect and significance levels of other variables. The incorporation of the community level birth rate does thus not seriously disturb the effects of the other coefficients in the model. This seems to suggest that birth rate at the community level does have an independent effect that is not incorporated in other variables in the model, giving some evidence for our idea that people anticipate on dominant fertility strategies in their communities.

3 Due to the interrelation between sibship size and parity, we estimated separate models including only each of the two variables in order to determine whether results would change when compared to models in which they were both included. Results did not change substantially due to including either sibship size or parity and given their moderate correlation, we decided to keep both sibship size and parity in the model.

4 We tested whether the assumption of proportionality holds for our data. In order to do so, we plotted survival curves for each predictor in the model and determined whether curves were proportionally diverging. We found no indications of serious violations of the assumption of proportional hazards.