Tips on Doing Online Census Searches

Following are tips on making use of the available online census pages (I use ancestory.com, but there are others).   The 1880 census is generally available for free on the internet, but the others you'll probably have to pay for (see my other link, which is a little outdated).

Intro:  I found through a lot of trial and error that using the main thing in finding someone in the census records is persistence.  While I've come to the conclusion that some of my ancestors were "missed",  you'd be surprized how many I've found of *hours* of searching, where the census taker has made a mistake, or the person indexing the page made a mistake.   The trick is to find the ancestor without scouring all the census records.   Some of the indexed information is probably cothererrect, so make use of that to make your job easier.  Trust me, some of your "missing" ancestors are out , they're just indexed wrong, or you've been looking in the wrong place.  Go find them!


1) Make use of wild cards in every search.  If the person's name is Catherine, search for "Cath*".  The search star ("*") will match in the search against any set of characters.   Searching for "Cath*" means you'll find someone indexed as "Catherine", "Catharine", or even "Catharin".   The indexer is just putting down what's on the paper, not the common form of the name.   And the more time you spend searching census records online, the more you'll see the census takers couldn't spell, especially in the older ones.

Also check for variations.   Not everyone is in the census with their formal name.  Catherine might also be Kate, Katie, etc (so search for "Kat*".   Ancestory requires at least three characters before a wild card, so you might need to search for "Tom*" and "Thom*" separately.  "Elizabeth" could be "Eliza" or even "Lizzie", so try all possibilies.

Always pick a range on any year you that you put in.   The errors on children tend to be smaller, so someone less than 10 years old at the time of the census probably only needs +/- 2 years.  Older people just didn't care and you might need +/- 10 years or more to find them.

You can also use wild cards in the middle of names to reduce the number of names returned.   For example, if you're looking for Hartlein/Hartline/Hartlin/Hartlien/etc, you could search for "Hart*n", but there are a lot of Hartman's and you don't want those.  So pick "Hartl*n" or even "Hartl*n*".  And also try "Hertl*n*" since the first vowel might be recorded different (think phonetic - many people were not literate and the spellings are very random.   Don't think that your ancestor changed his/her spelling every 10 years, it's just how the census records work).


2)  I recommend using a mix of populating the search fields.   Don't try to fill in more fields that necessary because  errors in transcribing are more likely to filter out your ancestor.   Try to get a combination of fields that results in about 1 page of results (if you find >200, you'll spend a *lot* of time hitting the "Next" button and looking at people that aren't your ancestor.

Some typical combinations that I use would be First Name/Last Name/State, or Last Name/State/County, or Last Name/Age/State.  But rarely would it make sense to populate all the fields that you can.


3) Make use of the soundex, unicode, or exact switches.   The first two use phoenetic matching techniques.   Be carefore because on ancestory, I've seen that using wild cards (eg, the "*" character), together with phoenetic matching does not work.   If in doubt, search for someone you know exists.  If you find that adding the soundex button results in thousands of uninteresting results, stick with exact spelling, but you will need to manually create the possible spellings.


4) Make use of neighbors.   You may find that you found your ancestor in 1860 and 1880, but they're not listed in 1870.   On a good day, you'll find that they are still at the same address.   So make a list of the adjacent names from 1860 and 1880, and search for those in 1870.  Then search several pages before and after them.   I've found a lot of ancestors by doing this.

Another way to use neighbors (in a big city) is to use a city directory.  Many of the old city directories are online.  If you now the address, try to find someone that lived nearby in the directory (search for the street address).   Note that in many cases, the even numbers are on one side of the street and odds on another.   The census taker may or may not have gone up one side first.   So find close numbers, both odd and even, then find them in the census of interest (using a close year directory if the exact year isn't found).


5) Skip the last name.   In many cases, I've found ancestors by searching only on the first name with a combination of other fields.   If you search for everyone named "Rob*" born in 1850 +/- 2, in Schuylkill County, PA in 1870, you'll only find 57 of them, which is a manageable number to scan, and menas the surname could be totally messed up (or absent) and you'll find them.


6) Search for the children instead.   If the first *and* last names were mangled in the census or indexing, try the same techniques on each child they had in a census before or after the one of interest.   Many times you can find the rest of the family by doing this.  In one case, I found that the person "Nathan Knarr" was indexed as "Nathan Hammel".  Hammel was the previous family listed on the page.   But by checking for the right combination of names and ages, and then checking the original page, I found it.  Realize that the children's ages are probably more reliable, so they're a good way to search.   If you found the family in the following census, and they had children aged > 10, you can search for those children (being 10 years younger) in the problem census year.

Also find the ancestor's older married children.   Many times, when the children moved west, the parents tagged along and lived out their last years with a child.  While your ancestor's siblings may not be your main interest, experienced genealogists rely on researching siblings to find information on the direct ancestors.  In one case, I found my ancestor living with his married daughter, and he was listed under the surname of his daughter's husband (the census taker apparently thought he was the man's father).


7) Skip the place (country/state) of birth.    I've found many cases where the place of birth changes in just one census year.   The census taker got carried away with assuming that everyone was born in the state.  Sometimes they didn't understand the person's country of origin (if they're foreign born), or the name of the home country was changing (you'll see that the Ukraine, Austria, Hungary, Poland might all be used by someone in different years.  You'd need to study your history to understand why, but I'd recommend using birthplace sparingly to filter results.

In one case, I found that the children were listed with the step-father's place of birth by mistake, the census taker not realizing they were step-chlildren.


8) Make use of the spouse field.   Try looking for everyone named "Robert" married to "Cath*" in your area of interest so you can avoid the issues with troublesome surnames.


9) Anticipate spelling mistakes.   This is a last-ditch desparation technique and in general there are better ways to do searches.  But I know it's worked at least once in my searches.  The initial "M" might be confused with an "N" for example in rough cursive writing.


10) Try adding spaces.   If you're searching for a name like "McDevitt", also try "Mc Devitt".   Ancestory.com does not skip the space - you must put it in if it was indexed that way.   Better is to search for "McDev*" of course.   My one McDevitt ancestor was indexed as "Mc Derett", living across the river in another state, and that one took a long time to find!   Names from Holland have the same problem.   It might be "Vanhorn", "Van Horn" (or "Van Horne").


11) Keep notes on how you found the ancestor.   Write down all the information you see (you'll want to come back and double check hard-to-read children's names or dates later, or the entries on occupation or war service later!).   If the name was indexed wrong, try to find how it was indexed so you can find it easier later.  To do that, if the surname was difficult to read, go back to the search page and try to do a search with other information on your ancestor - use just the first name, age and township/city for example.  When you find the wrong name in the search returns, make a note of it for future reference.   You might need to make a note to search for a certain neighbor's name, and then go down 4 lines.  But it saves doing all the work all over again.  Another idea is to save the images to your computer for future reference.


12) Beware of ethnic spelling differences.   For example, the German names starting with "Sch" are typically spelled with "Sh" in English, so the initial letters may change from census to census.   Smith might be Schmidt, then Shmitt, then Smith again, depending on the census taker.   East European names (Poland, Russian, etc) can end up with just about any combination of consants and vowels!