Parochial registration began in England in 1538; civil registration, apart from the period 1653–60, in 1837. If all registers had survived, all had been kept to a uniform, high standard throughout, and all had survived the hazards of fire, damp, rodents and insects, theft, malicious damage, and general neglect, it would be possible to select a sample of registers for a particular purpose over a particular span of years in the knowledge that all the parishes in the sample would yield information of value throughout the period in question. It would then also be possible to select a random sample from the totality of parish registers and ensure thereby that the results of the exercise would be representative of the country at large. In practice, the defects and deficiencies of Anglican registration ensure that any study which attempts to cover the bulk of the parish register period and to make use of a significant number of registers must face two difficulties: the implications of being unable to use a random sampling approach; and the need to adopt a strategy for coping with the fact that, even with care in the initial selection of registers, some will ultimately prove to be unusable for a variety of reasons over the whole or a part of the period of interest.

The first of these two questions, the problems associated with the use of information drawn from a non-random sample of registers while attempting to identify national patterns, is discussed in the next chapter.