Secrets of the Census

by Donna Przecha

The United States Census is an extremely valuable tool in genealogy research. Several censuses give not only names, ages and birthplaces, but also state the relationship of people within a household. Depending on the questions asked in that particular census, you may also learn when your ancestors came to the U.S., if and when they were naturalized, how many children a woman gave birth to and other vital pieces of information.

Background Information

In the U.S., a census has been taken every 10 years from 1790 through 1990. All censuses taken after 1930 are still confidential and the information they contain is not open to the public. The census from 1790 through 1840 only named the head of the household and the numbers of people in categories — 3 males over 16, 2 males under 16, etc. The census from 1850 through 1930 lists each member of the household and usually gives the relationship to the head of the house. It also gives age (later years give birth month and year) and place of birth (usually just the state or foreign country). The 1890 census was almost totally destroyed by fire so there is a 20 year gap between 1880 and 1900.

All U.S. censuses are available at the 11 offices of the National Archives. Many of the censuses are also available in many genealogy libraries and Family History Centers. All censuses are available on loan from the Family History Library in Salt Lake City through your local Family History Center.

A beginning genealogist will probably have information on at least some relatives who were alive in 1930 and will start with the 1930 census and work back as more names are found and linked.

People are enumerated in a census by year, then state, then county, then enumeration district. You must know the state to begin a search. Within the 1850–1930 period, unless you know a fairly exact location, the most important consideration is the existence of an index.

Later Censuses and Soundex

The 1920 census has been indexed by the government by surnames within a household using the Soundex system. An index is a major undertaking. Consider that the Soundex (index) for the 1920 census takes 8,585 rolls of microfilm while the actual census only uses 2,076 rolls. Since this microfilm must be purchased, some libraries may only be able to afford the actual census return, but it will probably be more important for you to obtain a copy of the Soundex first. The Soundex coding system is an amazingly efficient system for handling names, especially in this country where you may have Smith, Smyth, Smythe or Cain, Cane, Caine, Canne. Soundex will reduce most of the variation of a name to one code. Many genealogy programs or utility programs will Soundex code a name, but doing it by hand is a fairly simple process. Just assign numbers to letters using the chart below. The first letter of the name comes first followed by 3 numbers as follows:

1   B P F V
2   C S K G J Q X Z
3   D T
4   L
5   M N
6   R

Letters with the same code are pronounced somewhat the same and when names, especially foreign ones, are recorded, one letter could have been substituted for the other. By using numbers, it usually eliminates "losing" a name because a different letter was used.

To create a Soundex code, first write down the surname you wish to encode. Keep the first letter, but cross out all vowels (A E I O U) as well as the letters W, Y and H. If there are any double letters, cross out the second letter. With the remaining letters, assign the codes as above. Thus, Johnson would become J525. One final rule: If two or three different letters side by side have the same code, disregard all but one letter. Schultz, which you might think would be S243 is actually S432 because the initial S and the C are both represented by a 2. If you do not have 3 letters left to code, fill in with zeros.

The Soundex is arranged by state so you next select the film number covering your code number for the state you wish to check. All names with the same code number will be grouped together, arranged in alphabetical order by first name. Names like Shirinian, Sherman, Schermoon and Szermankowitz will all have the same Soundex code — S655 and will be mixed together. In spite of similar coding you will be able to see that some names are so different that you can ignore them. Sometimes a range of codes will be grouped together if there are not many names in one code.

When you look in the index, you will see that it gives abbreviated information including the county where the listing is found, the age and birthplace of the head of the household as well as others living there. Once you locate the right person, you need to note the volume number, enumeration district, sheet and line. You then go to the listing for that state, then the county and find the reel number for the enumeration district you need (Reel numbers will vary according to the library you are using.) When you find the enumeration district on the film, look for the page, and the family will begin on the line noted. In the actual census, the enumeration district and sheet number are in the upper right hand corner.

The 1900 census also has a Soundex index for all states. The 1910 census has only been indexed for the following states: Alabama, Arkansas, California, Florida, Georgia, Illinois, Kansas, Kentucky, Louisiana, Michigan, Mississippi, Missouri, North Carolina, Ohio, Oklahoma, Pennsylvania, South Carolina, Tennessee, Texas, Virginia and West Virginia. Unfortunately, this omits some of the most populous states such as New York and New Jersey. The Soundex index to the 1880 census only covers households that had children under the age of 10.

The Soundex indexes were created by copying the original handwritten entries from the census. The index cards for some years are also handwritten. This means that you are reading the handwriting of one person who wrote down what he or she thought another person had written. Needless to say, errors can occur. In all indexes, if you do not find the name you are looking for, you should look under other letters that might be written similarly — F for T, S for L, etc. You should also assume that an L (4) and a T (3) within the name might have been mistaken and search other Soundex codes. Be alert for unexpected spellings of the name you are looking for.

Earlier Censuses and Printed Indexes

Early censuses have been indexed by private companies, usually in the form of bound volumes. Since these books are quite expensive, most libraries will have a limited collection. (The Family History Library in Salt Lake City has an extensive collection, but they do not circulate.) It may pay to visit different libraries in your area to find out which indexes they carry. An index will usually cover one state, but some states have been divided. All states have been indexed for 1790-1850. The last censuses being indexed are 1860 and 1870.

These indexes are printed books, but generally computer generated, the majority being done by Accelerated Indexing Systems. The data used to produce these books has been reproduced on CD-ROM disks available from There is a master disk which combines all the states that have been indexed by Accelerated Indexing as well as names from other CD-ROMs in the collection. The master index is a part of's Family Finder Index, which is available for searching online and also comes with CD-ROM versions of their Family Tree Maker software. Once you have located the name on the master disk, you can go to the individual disks for the complete reference. The individual census disks are available for purchase or are available at some libraries. The CD-ROM index is handy because it combines all the states. In addition, many errors, typos and misspellings in the original books have been corrected.

Locating the Entry

The index will list the name of the head of household, give the town and county of residence and page number. Some indexes include everyone in the household, but usually it is only the head of the household. Since an elderly parent may be living with a married child, you may not find him or her in the index. Once you locate the entry, note all the information from the index — county and township — not just the page number. Then go to the list of microfilm rolls for that state and county and then turn to that page number. Be sure you have the right county before you look too closely at page numbers. You may have the correct film for the county you need, but there may be another county first on the roll and each county may have its own set of numbers.

A page number is generally a bold number stamped in the upper right hand corner. Sometimes there will be handwritten page numbers and there may even be more than one set of stamped numbers. Sometimes every other page is numbered so your name may not appear on the page with the number, but a preceding or following page. If you are having trouble finding the entry, be sure you are in the right township.

Utilizing the Information

The census gives a lot of information and you should try to find your ancestors in as many censuses as possible. If you have the name of your great-grandfather, but nothing else except the names of all his children, try to locate as many children as possible in census returns. You may find him living with a married child in later returns. From this you will know where he was born and approximately when. Take note of the names near your ancestor's. If you find a 35 year old man and family and two or three households away, a 65 year old man of the same surname with his 60 year old wife, there is a good possibility they are related. If you are looking for the parents of the younger man, it would be worthwhile checking the property and probate records of the older man and looking for an obituary for him. Even if you find your ancestor as a 13 year old in a census, you should look back to the previous census where he was 3. Sometimes people would adopt a nickname or middle name at an early age and use that exclusively throughout life. If you were not able to find a birth record for Hank Smith, you may find that as a 2 or 3 year old child he was listed as Charles Henry Smith. This indicates that the birth was probably registered as Charles and you won't ever find a record for Hank. Looking at as many returns as possible may also give you a pattern of migration. The oldest children might be born in Pennsylvania, the second 2 in Ohio and the younger ones in Indiana.

Dealing With Errors

Keep in mind how the census was taken and how the indexes were created. It was a very inexact science! The census enumerator took the information orally. He was probably in a hurry, many of the foreign immigrants had heavy accents and most families were large. He didn't always ask how to spell a name — and many people were illiterate. The Polish letters prz are pronounced something like "pzsh" and an enumerator who didn't speak Polish might transcribe them as psz or pzsh which means they will never turn up under the expected Soundex code. Many letters can sound alike — G and K, C and K, C and S, F and S and many others that are unexpected to English speaking people such as GR being mistaken for CR.

The census taker usually interviewed at least one person in the household, but if no one was home and the farm was a long way from his home, the enumerator may have obtained the information from a neighbor. In any case, the person supplying the data may not have known the facts. The neighbor may not have known the children's ages or birthplaces or a husband may not have known his mother-in-law's place of birth. Many, many people are recorded as being born in the wrong state. Anyone may assume a person was born in the U.S. because he or she has no accent when actually the individual may have been born in a foreign country and come to the U.S. as an infant or young child. When an individual was asked when he came to the U.S., he probably searched his memory and said "oh, about 1883". He didn't consult records and the actual year may have been 1887. They probably recited their children's names rather quickly and son Georgie may have turned into daughter Georgia. Many people could be recorded as the wrong sex if they didn't have traditional names. A child who had gone to stay with a relative for a month, perhaps to help out with a new baby might be counted twice — or not at all. Because of all these potential human errors, it is important to check as many censuses as possible.

Unable to Find?

If you cannot find your person in an index and you feel reasonably confident that he should be there, or if you have no other leads, do not hesitate to check the whole county if it is a rural area. A county often takes about one reel and can be done in 2-3 hours. If you are researching a name that is not extremely common such as Jones, Smith or Brown, it might be worthwhile noting everyone with the surname you are looking for — at least the name of the head of household and wife, ages and page number. Then you will not have to go back if you find your family in that county. If you are looking in a large city, it is very difficult to check the entire city. You should use city directories and street indexes to try to narrow the area. If the people belonged to one ethnic group, a history of the city will tell you what areas which nationalities settled in.

Once you have checked all census returns back through 1850, the census becomes less valuable but still can provide evidence. If you know your ancestor was born in 1821 and a family of the right name in the 1830 census has a 10 year old male, you know that he may belong to this family. If the family only consists of people over the age of 16, then it is unlikely that the head of this household is his father.

Slaves were recorded by numbers, usually without names, and Native American (Indian) censuses are a separate set of records.

Other Censuses

Many states also had their own census, sometimes at more frequent intervals. There are census returns in other countries also. Canada and Great Britain have census for every 10 years since 1841. There are not nearly as many indexes for these census as for the U.S. ones, but there is an online index of the 1871 census for Ontario. Because of their 100 year confidentiality rule, only 1841–1891 are open to the public. They can also be obtained from the Family History Library. No matter what area you are researching, one of the first resources you should look for is a census.

About the Author
I began genealogy in 1970 when we were living in Ogden, Utah for a short time. I was immediately hooked when, on my first visit to the local Family History Center, I found my great-grandparents in the 1850 Ohio census. I have been researching ever since on my own family and for others. I soon recognized the value of computer programs for keeping track of the data. I was a founding member of the Computer Genealogy Society of San Diego and editor of the newsletter. I have written a third party manual on ROOTS III and, with Joan Lowrey, authored two guides to genealogy software. Using ROOTS III and WordPerfect, I have written several family history books for others, but have yet to stop researching long enough to complete my own family history!

Home | About Us | Site Index | PRIVACY
© 2011