FLHIG
 

Family and Local History
Indexing Group

Selected articles from the Newsletter

This article first appeared in SIGGNL 20  pages 10 to 13

Census Indexing in North-West Kent FHS
Stephen Archer

The 1851 Census Indexing Project in North West Kent F.H.S. began in 1981, using a set of 22 census microfilms lent by the PRO. Originally the intention was to transcribe all the entries for our area, producing a typed transcript with surname index for each parish. The indexes would be lodged in the Society's library.

The technique was based on one of the suggested methods that had appeared in a booklet issued by the Federation of Family History Societies in 1981. This involved transcribing on to standard A4 (landscape) forms, from which the transcript would be typed. To produce each surname index, the forms would be photocopied and then cut into long narrow strips (one for each entry). These would be manually sorted and typed up. However, as some of our indexes would contain up to 35,000 entries, this method was clearly going to be extremely unwieldy. I was very surprised that the Federation had suggested it. Moreover, the method used page numbers rather than PRO folio numbers.

When I took over the project in 1986 I made an audit of what had been completed, and from that I estimated that we would take another thirty years if we continued to use the Federation method. So I suggested that, henceforth, we should extract only name, age and folio, and use computers to index the data rather than using a manual method. We would not produce a transcript as such, but a name index for each Registration District which would then be printed and made available in booklet form. The Society committee accepted the change of direction and I set to work early in 1987, initially circulating a questionnaire to all those who I knew had been involved with Society projects in the past. I devised a new A4 transcription form that could accommodate two frames from the census film instead of one.

Use of microfilm readers

Shortly afterwards I was lucky enough to obtain some second-hand 35mm microfilm readers, and these were rapidly put to use by Society members working at home. This certainly accelerated progress, although having an individual film tied up with one user for several months meant that work on a given Registration District couldn't be parcelled out as much as I would have liked. I resisted the temptation to physically cut any of the films!

Use of data entry program

When we began, IBM-compatible PCs were relatively expensive and not common in peoples' homes, but several of our members (including myself) owned BBC microcomputers which were well-suited to data entry. So I wrote a simple data entry program in BBC BASIC that could be distributed to all the helpers, of which there were around half a dozen any one time. The data entry program held all its data in memory (RAM), and could accommodate a file of some 800 entries at one time - enough for most enumeration districts. I would post out the transcription forms for two or three enumeration districts at a time, together with a 5¼" disk containing the program. Within a few weeks I would receive the disk back, and in this way we worked through our first Registration District - Bromley - which had 16 parishes and some 18,000 entries.

First stage checking

Unlike the 1881 census project, we couldn't afford the time to do the transcription twice, but we did use two checking stages to try and minimize transcription and data entry errors. I would list out the data for checking, and hand out these listings to the first checker, who would generally be more experienced than the transcriber. They would work through the film again, marking any corrections on the listings.

Second stage checking

The second checking stage was possible only because we all lived relatively close to London, and could actually get access to the original census books at the Public Record Office in Chancery Lane. Although this took some organization, it allowed us to make our indexes that much more accurate than they would otherwise have been. A checking party would normally involve some four or five of us, working through the computer listings and comparing against the original books. The intention was primarily to focus on the problems identified by the first-stage checker, but very soon we came to realize that some of the entries passed as correct by the checker were in fact wrong. Thus we began to check every surname. The difference in clarity between film and original book was alarmingly high for some districts, and many entries obliterated by pencil strokes are that much clearer on the originals.

Amalgamation

Although it would have been possible to do the whole indexing and printing exercise on a BBC microcomputer, by then I had access to PC facilities. A friend helped with the task of transferring data from BBC micro to PC format, using a serial cable and terminal emulator program run from the PC end. After a few glitches this was working fine, and allowed me to continue to use the services of those members with BBC microcomputers. I transferred all the data into dBase III files on the PC, amalgamating all the files for a given Registration District.

Utility programs

Working through the data doing all the corrections was a lengthy exercise, but I developed a series of utility programs that would rapidly read through all the files listing various oddities and inconsistencies. Another program printed the total number of entries for each enumeration district, which I would then compare with the summary page at the start of each district, allowing missing entries etc. to be identified, sometimes throwing up errors in the Victorian arithmetic.

Incomplete and dubious entries

One problem that faces any indexer of original material is how to deal with incomplete and dubious entries that fail to respond to the eyes of experienced transcribers. In the absence of any suggested guidelines on the subject, I decided to allow the published indexes to contain entries spelt in two (or sometimes more) ways. For instance, the entry "HEYSON/KEYSON, Rebecca, 29", from Volume I of our index series also appears as "KEYSON/HEYSON, Rebecca, 29". In order to automate the process of adding such entries, I had a program that would look for the "/" character in the original data file, extracting all matching entries, then swapping the two surnames round and creating a file of additional entries. These I would then append to the original file just prior to indexing the data.

Additional research to ensure accuracy

Shortly into my period of tenure of the project, I adopted other techniques to try and minimize errors. I obtained photocopies of relevant pages from the 1852 Post Office Directory of Kent, which the transcribers and checkers could use to aid them. I also had a utility program that listed out all the problem entries relating to children aged 14 or under. These could then be compared with the G.R.O. births indexes, then located at St. Catherine's House. Sometimes after a little work an individual could be identified in the indexes, thus clarifying a surname that was difficult or impossible to read in the census return. Nonetheless, a name would occasionally be so garbled that only an initial letter followed by a series of dots would suffice.

Final output

For printing the indexes I developed a program using Clipper, a compiler for dBase III, which would snake the output into three columns, generating camera-ready copy directly to a laser printer. In our first index (Bromley Registration District, 1988) I arranged all the entries in strict alphabetical order of surname followed by Christian name. However, starting with Volume II of the series, I decided to preserve the sequence of forenames within each family so as to aid the searcher looking for more than one family member. This change involved indexing on the surname and folio number, rather than surname and forename.

However, after ten years and six published indexes I decided to hand over to another volunteer, who is now working on our seventh and final booklet. This will cover the parish of Greenwich, which in 1851 had a population of some 35,000 people.

Ç TOP