This post is a follow-up to my post about using enumerator instructions for census research. Check it out to learn about all the great information you might be missing in census records.
A great thing for Occasional Genealogists to do when they don't have a lot of time is correlate or compare census data across years. It's something every researcher should do; it has a lot of benefits, yet can be done in short sessions. It's the perfect Occasional Genealogist mini-project!
Census comparison is actually one of the things I used to do when I was an Occasional Genealogist without access to records. I didn't know it was so important and I rather thought I was taking a kindergarten approach to genealogy.
I like to lay things out in tables. Before having a computer, I would have done this on a wall, all over the floor, or with art supplies; that's what made it seem like a kindergarten project.
I was lucky I had a love of tables and did this by chance because there's nothing kindergarten about correlating evidence. And there's nothing wrong with using art supplies or doing this in a physical format (instead of digitally). What's important is that you do it and write down what you find (and that you can find what you wrote down!).
Since I'm featuring Evernote related posts as I get ready for my Evernote based lecture at the 2016 NGS Confernce in Ft. Lauderdale, Florida in a few weeks, I'm giving you another Evernote freebie to use for comparing or "correlating" census evidence. This is really just a very simple table and as you start to understand the purpose and advantages of comparing information across years, you should modify it to fit your needs.
That's part of the point of this form (and of "correlating evidence"). You are going to enter select data from every year your person appears in the census to see if it is consistent (you will compare data of the same type, from different years, to each other).
Also, the same technique is applicable to different records. As long as the different records provide the same type of information, you can compare them. That means you can compare a series of records giving your ancestor's age as long as you know when the record was created (i.e. you could estimate a birth year from that record).
The table already has each year from 1850 to 1940 entered in the first (header) row. I've included instructions telling you to enter "X" in the cells where the person was not alive so you can tell if you are missing census records or if the person should not be in them. An X means they shouldn't be in that record. A blank means you're missing a record. I enter "Xs" first, before I start entering data, to make entering the data quicker but also to make it faster to see if a census year is missing.
Another hint for "speed" is enter a column of data, i.e. the information from a single year. This will make it easy for you to do this in small amounts of time, even if you are interrupted. Working across rows will take longer as you need to look at each census record.
The next row is for the actual enumeration date. This is the date written on the census page; the day the enumerator should have actually come to your family's door.
When you are ready to analyze the data you've recorded, you will look at the instructions to see what the enumerator should have recorded for birth information. This is usually age at last birthday or age on official enumeration day with special instructions for infants.
However, enumerators don't always get it right. They may have been sloppy or been given the wrong information. Errors are more likely if the birthday and enumeration day are close together (have you ever given the wrong age shortly after your birthday?).
Tracking both enumeration dates can allow you to perform more involved analysis estimating a birthdate. This isn't extremely difficult so you may figure out how to do it once you start thinking about correlating data. I'm not going to describe it in this post.
Second, it's a double-check. Do these residences make sense? If not, you may have the wrong family.
Ancestry.com will suggest many wrong census records (usually there is only one "right" record per year, after all, and they suggest dozens of census records). Less experienced genealogists sometimes gather these up without even looking at them. You may have once been a less experienced genealogist of that type.
For 1850-1940 you will still mainly see movement westward. However, some moves are job-related. A family who's occupation is tied to shipping, shipbuilding, or are sailors will move up and down rivers and/or the coast. Farmers don't live in the city, some jobs are only done in towns/cities. Stop and think about the information you find.
You should be as specific in this field as the census is. In cities in the 20th century, you will get a street address (how early this starts differs by city). In rural areas you may only get the county. Later records are more precise, early records, less so.
Next, you have a choice. You can skip this row (calculated birth year) and come back later. However, for 1900 record the month and year written on the census, even if it doesn't agree with the age.
The other choice is to calculate the birth year. Later you can refine it with the enumerator instructions. Either way, you'll come back to this row so the choice is yours.
You can add notes and explanations under the table and then record the citation for each census record. You don't have to create a formal citation but someone should be able to find that census record from the information you give.
All the ins and outs of correlating could fill a book. Here are the basic goals the first few times you try this.
I also have an engineering-related background and working with other genealogists I've found those of us with engineering-related backgrounds tend to be all about tables---others, not so much.
It may be that you find setting up the table a chore, I've done it for you in this case. If you can't "see" the data better in a table, there's no reason to use it.
I will create a separate table if I am trying to compare a lot of different data points (a data point being an age, or a birthplace, or a naturalization status---in the census one data point is usually data in one column).
The follow-up post is about using and adjusting the form if you are unsure or if your initial comparison is wildly skewed. This possibly indicates you've identified several different families instead of one.
A great thing for Occasional Genealogists to do when they don't have a lot of time is correlate or compare census data across years. It's something every researcher should do; it has a lot of benefits, yet can be done in short sessions. It's the perfect Occasional Genealogist mini-project!
Census comparison is actually one of the things I used to do when I was an Occasional Genealogist without access to records. I didn't know it was so important and I rather thought I was taking a kindergarten approach to genealogy.
I like to lay things out in tables. Before having a computer, I would have done this on a wall, all over the floor, or with art supplies; that's what made it seem like a kindergarten project.
I was lucky I had a love of tables and did this by chance because there's nothing kindergarten about correlating evidence. And there's nothing wrong with using art supplies or doing this in a physical format (instead of digitally). What's important is that you do it and write down what you find (and that you can find what you wrote down!).
Since I'm featuring Evernote related posts as I get ready for my Evernote based lecture at the 2016 NGS Confernce in Ft. Lauderdale, Florida in a few weeks, I'm giving you another Evernote freebie to use for comparing or "correlating" census evidence. This is really just a very simple table and as you start to understand the purpose and advantages of comparing information across years, you should modify it to fit your needs.
This post also introduces the idea of correlation for those who aren't familiar with the term. Here's how to use the rest of the post depending on if you're going to use the Evernote template or not.
- If you're going to use the Evernote template, go ahead and get it so you can take a look at it as I describe how to use it. I do include some screenshots but only once I'm talking about specific sections (because you can't see the whole page easily as a screenshot). You can also scroll down to see the screenshots, first, if you need to.
- If you are NOT going to use the Evernote template, there are small amounts of information throughout this post that only apply to the Evernote template but most of it applies to either correlation in general or correlating using a table. You can create your own non-Evernote table using the information provided in the final sections of the post.
Errors, Instructions, and Adjustments
I've heard beginners lament about using census records because they seem so error-prone. I touched on this briefly in my blog post about using the enumerator instructions. The information might not be wrong. When you read the instructions you may find it is correct and correlates with other information you have.That's part of the point of this form (and of "correlating evidence"). You are going to enter select data from every year your person appears in the census to see if it is consistent (you will compare data of the same type, from different years, to each other).
You can do this with any information that appears in more than one year.
Also, the same technique is applicable to different records. As long as the different records provide the same type of information, you can compare them. That means you can compare a series of records giving your ancestor's age as long as you know when the record was created (i.e. you could estimate a birth year from that record).
Back to the situation where you think the information is wrong until you use the instructions, part of correlating is putting the information into the same format. You want to compare apples to apples, not apples to pears (because we know better than comparing apples to oranges).
An example of putting the information in the same "format" is having a birth year from one record and an age in a different record. We normally would take the age and figure out the approximate year of birth. The same applies when we have ages in different records. We don't see if 18, 24, 27, and 35 are the same number. We "format" this information into the same type, usually an approximate year of birth. You're already doing this for simple pieces of information, you just need to take it further.
When you use the enumerator instructions you may find what appears to be conflicting information is actually consistent when you "format" it to be the same type (the instructions would provide this if it's something more complicated than an approximate year of birth). If you've ever compared the age ranges provided on the 1790-1840 census, and read the descriptions on the form to get the birth year ranges, you're already familiar with how to do this.
Starting with the Census
The census is easy to start with and likely to give you a number of pieces of evidence of the same type to compare. That's why I've started you with birth information. Age and birthplace are requested from 1850 to 1940.The table already has each year from 1850 to 1940 entered in the first (header) row. I've included instructions telling you to enter "X" in the cells where the person was not alive so you can tell if you are missing census records or if the person should not be in them. An X means they shouldn't be in that record. A blank means you're missing a record. I enter "Xs" first, before I start entering data, to make entering the data quicker but also to make it faster to see if a census year is missing.
Another hint for "speed" is enter a column of data, i.e. the information from a single year. This will make it easy for you to do this in small amounts of time, even if you are interrupted. Working across rows will take longer as you need to look at each census record.
In the end it doesn't matter which way you do it but this is The Occasional Genealogist blog.
I recommend keeping one table for one individual in a note. You can put more tables in a note but it can quickly become confusing and there's no reason in Evernote to overstuff a note.
Also, if you are missing some census records, make sure you record notes, if you have them, about why. Basically, correlating census information is quick. If you come back to this and see a missing census year there's a decent chance you'll immediately try to dash off and grab that missing info. If you've searched and searched and can't find that census (or it's a random lost census for your location), leave yourself a note on your table so you don't waste time looking for something you won't find (again). This can also be helpful if you share information with a research cousin (or for whoever inherits your genealogy and continues your work).
The Data
Enumeration Dates
The first row of data is "Official Enumeration Day" which I've entered for you. Because the person's age is dependent on when the census is enumerated, this is important. This obviously doesn't affect birthplace and other non-changing data so if you alter the table, you may be able to remove the enumeration rows.The next row is for the actual enumeration date. This is the date written on the census page; the day the enumerator should have actually come to your family's door.
When you are ready to analyze the data you've recorded, you will look at the instructions to see what the enumerator should have recorded for birth information. This is usually age at last birthday or age on official enumeration day with special instructions for infants.
However, enumerators don't always get it right. They may have been sloppy or been given the wrong information. Errors are more likely if the birthday and enumeration day are close together (have you ever given the wrong age shortly after your birthday?).
Tracking both enumeration dates can allow you to perform more involved analysis estimating a birthdate. This isn't extremely difficult so you may figure out how to do it once you start thinking about correlating data. I'm not going to describe it in this post.
Residence
The next row is residence. I've included this for several reasons. First, it's an informal citation in case you ignore the place I've given you at the bottom to cite each census record you use.Second, it's a double-check. Do these residences make sense? If not, you may have the wrong family.
Ancestry.com will suggest many wrong census records (usually there is only one "right" record per year, after all, and they suggest dozens of census records). Less experienced genealogists sometimes gather these up without even looking at them. You may have once been a less experienced genealogist of that type.
For 1850-1940 you will still mainly see movement westward. However, some moves are job-related. A family who's occupation is tied to shipping, shipbuilding, or are sailors will move up and down rivers and/or the coast. Farmers don't live in the city, some jobs are only done in towns/cities. Stop and think about the information you find.
You should be as specific in this field as the census is. In cities in the 20th century, you will get a street address (how early this starts differs by city). In rural areas you may only get the county. Later records are more precise, early records, less so.
Age and Birth Calculations
After the residence, it's time to finally get to the meat of this comparison, age. Record exactly what it says on the census. Later you'll use the census instructions to analyze this data and record that analysis in a written explanation.Next, you have a choice. You can skip this row (calculated birth year) and come back later. However, for 1900 record the month and year written on the census, even if it doesn't agree with the age.
The other choice is to calculate the birth year. Later you can refine it with the enumerator instructions. Either way, you'll come back to this row so the choice is yours.
Almost done... Birthplace, etc.
Record the birthplace, exactly as written on the census, in the next row. If you want, you can record more data points.You can add notes and explanations under the table and then record the citation for each census record. You don't have to create a formal citation but someone should be able to find that census record from the information you give.
All the ins and outs of correlating could fill a book. Here are the basic goals the first few times you try this.
- You will hopefully see the information from the census is consistent. At least write some notes, better yet, record what you are finding.
- You can simply write out what you are thinking ("why is the woman only seven years younger?" "is this really the same person?").
A Few Other Notes
"At-a-Glance" Consideration
There is something I want you to consider before you get carried away with this table. I did this as a table in the first place (years ago) because I wanted to see everything at-a-glance. I think a table is the easiest way to make a quick comparison and see red flags.I also have an engineering-related background and working with other genealogists I've found those of us with engineering-related backgrounds tend to be all about tables---others, not so much.
It may be that you find setting up the table a chore, I've done it for you in this case. If you can't "see" the data better in a table, there's no reason to use it.
Additionally, if the table gets too large, you can't see everything at-a-glance. Keep this in mind.
I will create a separate table if I am trying to compare a lot of different data points (a data point being an age, or a birthplace, or a naturalization status---in the census one data point is usually data in one column).
Remember, you can do this on paper, in a table format or any format you choose, as well as doing it digitally.
One Assumption
The table I've created is for the most basic census comparison, comparing birth information. This is the estimated birth year and birthplace. When I'm talking about how the template is set-up, I'm assuming you are confident you have identified the correct family in the census.The follow-up post is about using and adjusting the form if you are unsure or if your initial comparison is wildly skewed. This possibly indicates you've identified several different families instead of one.
