FLHIG

Family and Local History
Indexing Group

Selected articles from the Newsletter

This article first appeared in SIGGNL 7  (November 1994)

Notes on Census Indexing
by  Andrew Warren

Editor's comment

The following article is by one of the newer members of the SI, who by his own admission is a tyro at the game, but in the editor's view with more sound ideas on indexing than many of those who submit indexes on family history.  Once again I urge those who read this to submit their comments —  so many have recently had experience in the countrywide project of indexing the 1881 census, and must surely have opinions or experiences which they can share.

Probably census returns are amongst the most-indexed of all genealogical records.  Census indexes vary greatly in scope and format, from the very simple to the highly elaborate.  They vary in quality too.

Could the SIGG help to advise indexers, and planners of indexing projects, on how to get the best results with the expertise available?  My own experience of indexing, limited and amateur as it is, has been mainly of indexing census returns; so may I try and set the ball rolling?

The census indexer works unpaid, and works mainly for enjoyment.  The desire to help other census users is there, but I think it is secondary: people won't normally commit their free time to copying, filing and typing unless they're enjoying themselves doing it; the indexer whose motives are merely public-spirited tends to fall by the wayside.

This tends to influence the planning of census indexes.  A census index is commonly the product of teamwork. If the team is unpaid, their enthusiasm and their experience have to be taken into account.  In practice, this tends to produce simpler, quicker and cheaper indexes than experts in indexing or in genealogy might ideally like.  Nevertheless, a simple index that is complete and accurate so far as it goes may perhaps be better than no index at all.

The simplest census index is the surname index.  A chosen section of the census is searched.  Every surname found in the chosen section is listed alphabetically.  By each surname is a complete list of the places within the chosen section where that surname occurs.  Quick, cheap, easy to compile, type and store.

At the other extreme is the very detailed census index.  This has an entry for every individual within the chosen section of the census.  The entry tells you where that individual can be found in the census, plus all the other information that the census gives about the person.  Very helpful to the user.

Between these two extremes there are numerous gradations.  But in one respect, it doesn't matter so much what information the index includes or omits; what is perhaps more important is the accuracy of what is in the index.  A simple surname index that is complete, clear and reliable may be more use than an elaborate index marred by omissions and errors.  It is true that surname indexes provide strings of undifferentiated location references, which have a bad name among professional indexers.  At the risk of heresy, I would say it is quicker to check 20 or 30 known locations for a particular surname, than search a totally unindexed census for yourself.  So, from my admittedly amateur standpoint and limited experience, I offer some suggestions that might help a surname index to meet the basic requirements of Completeness, Clarity and Accuracy.

I look forward to reading much better suggestions from other SIGG members!

The golden rule is, "Index what the census says, not what it ought to say."  No doubt enumerators made mistakes.  Never mind: copy them.

Census indexers often find this hard to swallow.  Why perpetuate an obvious mistake, they ask.  In this article, I'm sure I need  not preach to the converted!  But just for interest I've picked out some surname variants found in my own corner of West Yorkshire.  Which are spelling mistakes, which are significant? Ambler / Hambler; Brook / Brooke; Cawdray / Cawthra; Davey / Davy; Ebding / Hebden; Farnish / Furnish; Goldsborough / Goldsbro; Hammon / Ormondroyd.

Years ago, before I'd heard of SI, I did a surname index for my local library containing this note: "Since the spellings in the census are often highly irregular it is wise to check possible variants, for example Bredrick/Bretherick."  Now, with hindsight, I wish I had used frequent cross-references ('Fawcett see also Fawsett'/ 'Fawsett see also Fawcett') as they do in the phone book.

In my bit of Yorkshire, they often used maternal surnames as Christian names.  A man listed in one census as 'Pickles Howker' (poor fellow) appeared in another as 'Howker Pickles'.  Which is the surname?  I've no idea.  In such cases we need an entry under each name, just in case.  Or three entries, in the case of 'Lupton Wright Dawson', I suppose.

A list at the end of the index of location references for illegible surnames is helpful.  Perhaps the user can read the name if the indexer can't - but not if the user doesn't know to look for it.

If the first letter or more is legible, an entry at the correct place in the sequence is helpful: 'RO...' or suchlike.

This topic is, of course, fully covered in BS 1749.  Alas, till I joined SI I'd never heard of it, and I bet not one census indexer in a thousand has read it.  Personally, I used to work letter by letter, ignoring hyphens, apostrophes and spaces.  This is easy - but is it acceptable?

Some indexes, including some of mine, give references by enumeration district and schedule.  This is a mistake.  The preferred system is to quote the Piece Number and Folio Number.  This is just as easy for the indexer, and far easier for the user.

The user may never have seen a census before, let alone a census index.  An explanation should, I think, cover these: the precise section of the census covered; the indexer's practice or all the points listed above; and the meaning of any conventional symbols used (such as dots for missing letters, questions marks to indicate doubt).

My only criticism is to regret the pronouncement of a "golden rule".  In the particular context of the spelling of personal names I fully support Andrew's argument, but there is a tendency to confuse  rules for transcription with those for indexing, and this seems to apply particularly to genealogy.   While transcripts should precisely follow the original, the entry term in an index may have to be modified so that all entries for the same term should come together, with of course the necessary references from alternative spellings.

Perhaps the final paragraph headed "Explanation" is the most important, since it is frequently omitted by inexperienced indexers or maybe deleted by unhelpful editors.  So often one has to hunt for the meaning of an unexplained symbol or abbreviation, that it should be mandatory for an explanation to be provided, preferably at the head of the index.  While it would be less needed if there were a standard set of symbols and abbreviations, or that those which exist were generally accepted, we must always remember that indexes are for the benefit of users who may have no special subject knowledge.

Ç TOP