The organisation of any project should be set within realistic research goals. Since oral history recording is dependent for worthwhile results on human memory, this fallible faculty must be accommodated by careful preparation. The planning of the project should, therefore, be based on as thorough understanding of the subject field (and of the availability of informants) as the existing records permit.

It is prudent, first, to fix a research period which is historically identifiable as being self-contained. In the lower deck project, for example, the so-called Fisher Reforms of 1906 altered several important aspects of naval life: the First World War stimulated further changes during the early 1920s; and the Invergordon Mutiny in 1931 was another watershed for Royal Naval seamen. The combination of these three distinct periods in a recording project, would have made it extremely difficult for sailors who served throughout them to avoid confusion on many details of routine life which, for research purposes, might be of critical importance. Three distinct periods of social change within a single career of professional experience are clearly difficult for informants to separate with few points of reference beyond their own memories. By setting the general limits of the lower deck project at 1910 to 1922, a reasonably distinct period of naval life was isolated as appropriate for oral history research.

The research problems which are created by rapid social change can seldom be eliminated entirely from oral history recording. It is for this reason that historically unsophisticated interviewing can result in information of uncertain reliability. Therefore, the project organiser's responsibility is to minimise the dangers implicit in such situations by his own common sense and historical sensitivity, and he should always apply the question 'Is this reasonable?' to the goals which he sets. Some practical examples of the application of this principle in oral history research are given overleaf.

The chronological scope of an oral history project should be fixed before any recording begins, bearing in mind the age of the likely informants as well as the historical character of the subject field. By the time the lower deck project began in 1975, men who saw service in the Navy as early as 1910 were in their eighties, and thus the opportunities for preceding this date were limited. This basic consideration affects all oral history recording. The informants who are actually available to be interviewed, also predetermine many of the topics which may be sensibly raised. Thus, owing to the slowness of promotion in the Royal Navy there was little point in introducing questions about, for example, conditions in petty officers' messes in 1910. Only informants into their nineties would have had the necessary experiences to be able to answer them. The chances of locating a sufficient number of interviewees of this great age, were sufficiently slight to preclude this and many similar topics -from being a practical aim within a systematic research project.

Similarly, the project organiser must take into account the structure of the particular group of people he is concerned with. For example, a battleship of the Dreadnought era - with a complement of some 700 men - might carry one writer (i.e. account's clerk) and one sailmaker. The odds against tracing such rare individuals more than fifty years after the events eliminated some aspects of financial administration and some trade skills aboard ship from the range of what it was likely to be able to achieve.
The selection of and possible bias among informants, are related factors which have to be appreciated. Between 1914 and 1918 the total size of the Navy increased threefold owing to the needs of war. A substantial proportion of those who served for hostilities only may not have accepted the traditional mores of regular lower deck life. At the end of a carefully organised and conducted project, the organiser had no clear idea of whether wartime personnel generally adopted the attitudes of those who had been in the service since they were boys, because the original selection of informants simply did not permit systematic investigation of their particular prejudices. An appropriate selection of sailors to be interviewed would have produced a representative sample of these kinds of informants and thereby provided suitable evidence from which conclusions about this particular question could be drawn. This obviously does not devalue the information for the purposes for which it was recorded, but it does eliminate the range of hypotheses to which this body of data is open. Thus, the project organiser must take into account the relationship between the subject matter of the project and his selection of informants and -at one stage yet farther removed from recording - this involves being clear about the kind of research evidence he is actually seeking to collect.