Determining the Accuracy of Census and Other Genealogy Records

Mariann asked Olive Tree Genealogy a great question about the accuracy of ages on census records.

Do genealogists often find a lot of  inaccurate ages (and projected birth years) on the US Census? I've been using ancestry.com, and relatives often seem to vary their stated ages (thus birth years) from one census to another. Then when I see a Death Certificate, the stated birth year sometimes doesn't agree with any of the census dates. There can be a variation of as many as 5 or 6 years. How do I know which date is the closest to accurate?

Mariann, 

This is a really important question. As you research ancestors you will often find large discrepancies in ages from one census to the next. There are several reasons for this:

1. People often did not know exactly when they were born. There were no ID cards, credit cards, Driver's Licenses and so on in the 1800s and earlier. These are all items that require an exact date of birth and so it is imperative that we know when we were born. But before the 20th century it was not important. In fact I have a copy of a letter written in 1841 by my 2nd great grandfather Levi Peer asking his mother to check her Family Bible and tell him when he was born.

2. We do not know who provided the answers to the census taker's questions. Was it the head of house? Was it a spouse? Was it an older child or a neighbour? Not knowing if it was the individual giving the answers means we do not know how accurate the information is

3. Different census years had different questions re ages. For example the question on age might be phrased as "Age at last birthday" or "Age at next birthday"  or simply "Age" And of course the census taker might not ask the question as specified.  Of course that won't give a 5 or 6 year variation in answers but it does account for some.

4. Women in particular often lied about their ages. Many individuals didn't trust the census or the census taker and would simply give whatever answer they felt like giving.

So there you have four reasons why census ages can't be trusted. You mention Death Certificates not agreeing with census records and that's another problematic record. A Death Certificate has information that was not provided by the deceased! So depending on who the informant was and how much they knew about the deceased, the information provided can be completely inaccurate.

I have a good example of a Death Certificate that is completely wrong but will, I'm sure, be accepted as accurate by researchers. My grandmother's 17 year old brother was the informant for their mother Mary Peer's death. He has given the wrong mother and father for his mother. Her parents were Isaac Vollick and Lydia Jamieson (this is proven with  documents such as her marriage record and census records). But he has said her parents were Stephen Vollick and Mary. Such a discrepancy took me awhile to solve.

Then I realized that in all likelihood, in his grief and confusion when the clerk asked him "Father's name" he gave his own father's name (Stephen). When asked "Mother's name" he gave his mother's name - Mary. Thus the entirely fictitious couple Stephen and Mary Vollick were created. Researchers finding that death certificate who have not yet found Mary's marriage record will have no idea that they are looking at completely erroneous information.

The bottom line is that it is wise to study all records and determine their validity by looking at who gave the information contained in each record. 

So for example a death record or a gravestone for example may not be accurate because the person themselves did not provide the information.

A birth record is considered quite accurate because the information is usually given by parents. But there's a caveat - if you are not looking at the original record there is a chance that errors crept in. Each generation of a record increases the chance of errors. You may want to read my article I Found My Great-Grandfather Online - Now What?? for more information on how much an original record can change.

Just to make things more confusing, remember that individuals might lie for various reasons. That is why we should always gather as many records as possible for each individual. Calculate the trustworthiness of each one. Then study them as a group. Sometimes you have to be content with an estimated year for an event based on all the records you've found.

But of course you should keep looking for a record that will give a first-person account of the event. In any record you should think about who provided the information. If it was not the individual then the accuracy is immediately in question. And even if it was the individual, did they know the answer? Did they lie?

Genealogy research is never easy but analyzing records is part of the fun!