DANSEY.net

In my opinion we should be grateful for the errors that are made by transcribers. Why? Because it teaches us to be better researchers!
 
The fact is, many of the errors on the census are nothing to do with modern-day transcribers - often they are the fault of the enumerator, and sometimes they are the fault of the head of household (as the 1911 Census schedules confirm).
 
The same search techniques that enable us to overcome transcription errors also help us to overcome other errors - and yet I'm frequently surprised by how few people use wildcards, sometimes because they aren't aware that they can be used. A member who had switched from Ancestry to findmypast wrote to me recently to complain that findmypast didn't allow the use of wildcards - yet findmypast are actually more flexible in allowing wildcards than Ancestry have ever been!
 
Of course, if you use wildcards incorrectly you can end up with far too many results - and whilst more results is usually a good thing, trying to find a needle in a haystack isn't much fun.
 
When I'm using wildcards I try to use them as intelligently as possible - this means considering what might have gone wrong. A very common problem is the misinterpretation of lower-case letters that look similar in cursive handwriting, for example the letters 'm', 'n', 'u', 'v', and 'w' can look very similar. The letters 'r', v' and 'n' are also easy to confuse, as are 'a', 'e', and 'o'.
 
Capital letters can also be misread - 'P' and 'T' are commonly confused, and I saw an example recently where the enumerator's capital 'F' looked just like an 'H'. Had it not been obvious that the family were more likely to be living in a 'Farm house' than a 'Harm house' I'd never have realised my mistake.
 
If I'm searching for someone with an unusual surname I know there's a very high chance that it has been misspelled or mistranscribed. What I try to do is pick out a few letters from the surname that are likely to have been transcribed correctly, for example if the surname was Hazlington I'd probably search for h*l*g* or - if that produced too many results - h*zl*g* (though I might hedge my bets by also searching for h*sl*g*).
 
Most sites will offer to include near matches in the search results, but in my experience the judicious use of wildcards is usually preferable (although when searching the GRO indexes I find that findmypast are very good at delivering up alternative spellings).
 
But wildcards aren't the solution to every problem - you can't use wildcards in the age, or year of birth box. What you can do, however, is provide some leeway for error: I usually allow +/- 2 years, but when you're searching the 1841 Census you most allow more (because most ages were rounded down to the nearest 5 years). Remember too, that woman tended to lose years as they get older - and men often do the same if they marry a younger woman.
 
However the best advice I can give you is to leave as many boxes empty as you can. I enter the absolute minimum of information - sometimes a surname and nothing else. I hardly ever specify a birthplace because even if it's shown correctly (and it often won't be), it won't necessarily be shown in the same way on every census.
 
Finally, transcription errors can be amusing. Jenny sent me a wonderful example recently - she discovered somebody on the 1861 Census who Ancestry had transcribed as Dodo Hackett. Take a look at piece 2453, folio 77, page 21 and you'll see exactly where they went wrong.
 
Tip: it's easy to look up census pages using the references at either findmypast or Ancestry, but not at The Genealogist or Genes Reunited.