Reading a Family Out of the Census, District by District
Individually identifiable census returns stay shut for seventy-two years. That single number decides everything else on this page: the newest census any member of the public can open was taken in the spring of 1950, and nothing after it exists for you at all. So the census is a historical instrument. It will not tell you where anybody lives today. What it will do, better than almost any other federal record, is put a named household at a named street address on a named day — and the way you actually get at that is not by typing a surname into a search box. It is by finding the enumeration district and reading it.
The short version
Say the unhelpful part first: the census cannot find a living person’s current address. Returns naming individuals are closed for seventy-two years, so the most recent one open to anybody is 1950, released on April 1, 2022. The 1960 returns open in April 2032. What the census is good at is putting a whole household on one line of one page in one year. Names, ages, birthplaces, relationship to the head of the household, sometimes the year of immigration and whether a person had naturalized, and — the part most guides skip — the street address. The search is geographic before it is alphabetical. Every schedule was written up inside an enumeration district, a patch of ground one enumerator could finish in two to four weeks, and the National Archives publishes the maps and the written descriptions that turn a street address into a district number. That is the search that works when the name index does not. And the name index frequently does not. Whole census years were never indexed by surname at all, whole states are missing from the ones that were, and the 1950 index was generated by machine from handwriting. A blank result is a statement about the index, not about your family.
Watch: why the census is searched by place, not by name
A walkthrough of the enumeration district, the finding aids that map one to a street, and what the schedule columns change from decade to decade.
Watch the walkthrough
The Newest Census Anyone Can Read Was Taken in 1950
Where the seventy-two years came from, why it is not written in the statute, and what it costs you.
The National Archives states the operative fact in one sentence on its own census research page: “Because of a 72-year restriction on access to the Census, the most recent year available is 1950. The 1950 Census was released on April 1, 2022.” The Census Bureau’s account of the rule itself adds the previous milestone: the 1940 returns opened on April 2, 2012, and the 1950 returns on April 1, 2022. And the Bureau’s genealogy reference answers give the next one plainly — “The 1960 Census records will become available in April 2032.”
Read those three dates together and you have the shape of the whole record. Ten-year steps, each one opening seventy-two years after the count. There is no queue to join, no early-access tier, no researcher exemption that gets you 1960 this decade. A locked census year is locked for everyone at the same time.
The rule is not a number in the statute. It is a letter.
This is the part that surprises people who go looking for the law. Open 44 U.S.C. 2108 and subsection (b) does not say seventy-two of anything. What it says is that with regard to census records “containing data identifying individuals enumerated in population censuses,” any release of that identifying information “shall be made by the Archivist pursuant to the specifications and agreements set forth in the exchange of correspondence on or about the date of October 10, 1952, between the Director of the Bureau of the Census and the Archivist of the United States, together with all amendments thereto.”
Congress, in other words, did not legislate a period. It legislated a correspondence. The seventy-two years live in an exchange of letters, and the statute points at them.
The National Archives’ own history blog, Pieces of History, sets out how that came about in a post titled “Census Records: The 72-Year Rule,” published January 20, 2022. Under the Federal Records Act the Archives had authority to open census returns at fifty years. On October 10, 1952, Roy V. Peel, Director of the Bureau of the Census, and Wayne Grover, Archivist of the United States, agreed instead on a seventy-two-year closure for population census records transferred to the Archives; under it the 1880 returns went out in 1952 and what survived of 1890 in 1962. Then in 1972, on the eve of the 1900 release, the Bureau reversed course and argued the 1952 agreement was invalid because it conflicted with Title 13, Section 9 of the U.S. Code — the Bureau’s position being that census records should be restricted permanently. The release was held up pending a Justice Department view, and the matter was finally settled by statute: 92 Stat. 915, Public Law 95-416, October 5, 1978.
Section 9 is worth reading, because it is the reason the argument was even arguable. 13 U.S.C. 9(a) forbids the Secretary and any officer or employee from using the information for a non-statistical purpose, from making “any publication whereby the data furnished by any particular establishment or individual under this title can be identified,” and from permitting anyone but sworn officers to examine the individual reports. That is an unusually flat confidentiality promise, and the Bureau made it to every household it counted. The seventy-two-year rule is the compromise that lets the promise expire without ever having been broken.
What that means for a search for a living person
Plainly: the census is the wrong tool, and no amount of technique fixes it. A person who is alive today and old enough to be findable was, at most, a small child in April 1950. Anybody born after that has no public census entry at all and will not have one for decades. Where a searcher wants a current address, the census contributes nothing directly and the work belongs to entirely different records — which is what our records-based locate practice exists to do.
There is one narrow federal exception and it is not a research service. The Census Bureau’s Age Search Service will search closed returns for the named person, their heirs or legal representatives, and the Bureau’s own notice states that effective March 4, 2026 the service is on pause and new requests will not be processed. It has never been a way to look up a third party, and it is not one now.
So use the census for what it is: a decennial photograph of a household in a place. It answers “who was living with whom, where, in 1940” with a precision nothing else of that era matches. It does not answer “where is she now,” and a page that lets you believe otherwise is wasting your month.
Before There Was an Index, There Was an Enumeration District
The census was collected geographically and filed geographically. Searching it that way is the skill worth having.
An enumerator in 1950 did not receive a list of names. He received a piece of ground and a map of it. The National Archives defines the unit exactly on its enumeration district maps page: “An Enumeration District was an area that an enumerator (census taker) could completely cover within two weeks in cities and within four weeks in rural areas.” That is the whole definition. It is a workload, converted into a boundary.
Because it is a workload, the districts are wildly unequal in size. The Archives’ 1950 finding aids put it this way: enumeration districts “varied in size from several city blocks in densely populated urban areas to an entire county in sparsely populated rural areas.” A rural district may be a whole county; an urban one may be four streets. The number of people is roughly constant. The acreage is not.
How to read an ED number
Every district carries a two-part number with a hyphen in it, and the two parts mean different things. The Archives explains it with a worked case: “Enumeration District numbers had two parts, such as 48-69, where 48 prefix was typically for the county and 69 suffix was for the specific area within the county.” The prefix is the geography above you; the suffix is your patch inside it.
The exception is the one that catches people out. Independent cities and very large cities were given prefixes of their own rather than borrowing the county’s. The Archives’ example is Chicago: although Chicago sits inside Cook County, its enumeration districts carry the prefix 103, while Cook County’s own prefix is 16. If your family lived in a big city and you have been searching under the county prefix, that is very likely the entire problem.
Turning a street address into a district number
The procedure is four moves and it has not changed in decades.
First, get an address for the census year, not for the year the person died. The Archives’ 1950 guidance is explicit that “City directories, telephone directories, and other sources from circa 1950 may provide addresses from around 1950.” A city directory listing from within a couple of years of the count is the standard input, and this is the step that actually takes time; everything after it is mechanical.
Second, find the map. The 1950 enumeration district maps sit in the National Archives Catalog as a searchable series, and the Archives tells you how to query it: type the year, the county and the state — its worked example is “1950 Sussex Delaware” — and narrow with a town name if the county returns too much. Not every town has its own sheet, so you may be reading a county map for a small place.
Third, read the pencil. The maps are not printed products. The Archives notes that the base maps were obtained locally and include postal route maps, General Land Office maps, soil survey maps and maps produced by city, county and state offices as well as commercial printers, and that census officials then drew the district boundaries and numbers on top of them by hand. The Bureau annotated them in red, orange, yellow and green pencil, and enumeration district numbers were usually written in orange. Once you know that, a map that looked like scribble resolves into a system. The Archives’ illustration is a detail covering parts of Washington and Sylvania Townships in Lucas County, Ohio, with the Michigan state line across the top, a neighborhood called Trilby, a golf course and three named ditches; the township boundary is red and the district boundaries are red, orange or yellow.
Fourth, cross-check against the written description, then browse. Every district also has a text description — a paragraph naming the streets, wards or townships that bound it — held as a separate Catalog series and searched the same way. Between the map and the description you should be able to satisfy yourself which side of a street the family fell on. Then you browse the schedules for that district, page by page, and read.
The finding aids by name, because you will need to ask for them
Federal finding aids have publication numbers, and quoting the number is the difference between a helpful reply from an archivist and a shrug. The Archives’ 1930 census reference answers name the three that matter most.
T1224, Descriptions of Enumeration Districts, 1830–1950, in 156 rolls, arranged by state and then by county, holds the written descriptions of every district across more than a century of censuses; the 1930 descriptions occupy rolls 61 through 90. M1930, Enumeration District Maps for the Fifteenth Census of the United States, 1930, in 36 rolls, holds the 1930 maps. And M1931, Index to Selected City Streets and Enumeration Districts, 1930 Census, in 7 rolls, does the conversion for you: it cross-references street addresses against enumeration districts for more than fifty cities. If your 1930 family lived in one of those fifty, M1931 turns a four-hour job into a five-minute one.
The Archives also points researchers to the free unified enumeration district finder built by the genealogists Stephen P. Morse and Joel Weintraub, which covers the 1870 through 1950 censuses and accepts a street address directly. We name it rather than link it because it is a private site; the Archives lists it among its own 1950 finding aids, which is the endorsement that matters.
Two abbreviations that will save you an afternoon
Browsing a district list, you will meet districts that appear to hold nothing. The Archives’ published definitions explain both cases. NP marks an enumeration district where no one lived — it exists to close a gap in the map, not to hold people. Void marks a district that was merged into another one, so its number survives in the catalog while its contents live elsewhere. Neither is a scanning failure and neither means your search went wrong. Separately, institutions — hospitals, schools, jails — were given enumeration districts of their own in 1930, which is why a relative in an institution can be entirely absent from the district covering the street the institution stands on.
What Each Census Year Actually Asked
The schedule changed every decade. Which year you open decides whether the record can answer your question at all.
Researchers lose weeks to a false assumption: that the census is one form asked repeatedly. It is not. Columns were added, dropped and reworded at nearly every count, and several of the most valuable questions were asked in one decade only. Choosing the year is half the research. The National Archives publishes a column-level guide to this under the title Clues in Census Records, 1850–1950, and the table below distills the parts that change what a search can do.
| Year | What it gives you that no earlier year does | What it still will not answer |
|---|---|---|
| 1790–1840 | Nothing about individuals. The Archives is blunt: only the head of household is listed by name, and everyone else is a tally inside age brackets. | Every other person in the house. Useful for placing a surname in a county, useless for reconstructing a family. |
| 1850–1880 | Every person in the household named, with age, birthplace and occupation. Value of real property is recorded 1850–1870, which is the pointer into county deed books. | The 1890 count that should follow it. Most of it burned in a Department of Commerce fire in 1921. |
| 1900 | Month and year of birth in column 7 — the only census that asks it. Years married (col. 10), children born (col. 11) and children still living (col. 12). Year of immigration (col. 16) and naturalization status (col. 18).Densest single year | Where the family was five years earlier. Residence history does not appear until 1940. |
| 1910 | Years married (col. 9), children born and surviving (cols. 10–11), immigration year (col. 15), naturalization status (col. 16), and Civil War service in column 30 coded UA, UN, CA or CN. | Month of birth, dropped after 1900. And no Soundex index at all in 29 states. |
| 1920 | Year of naturalization in column 15 — the only census ever to ask it. Immigration year sits in column 13 and status in column 14. | Age at first marriage, and any measure of what the home was worth. |
| 1930 | Age at first marriage; value of the home if owned or monthly rent if rented (col. 8); whether the household owned a radio; and which specific war a man served in, coded in column 31 as CW, Sp, Phil, Box, Mex or WW. | Year of naturalization, reduced to a yes-or-no. The unemployment and supplemental schedules no longer exist. |
| 1940 | Where the person lived on April 1, 1935, in columns 17–20, plus a circled X marking whichever household member gave the answers. | Parents’ birthplaces and age at first marriage — moved to a supplemental schedule covering only two names per page. |
| 1950 | Twenty questions for everyone, then residence one year earlier, parents’ birthplaces, schooling, income and military service for the six people per sheet who fell on a sample line. | Anything for the other thirty-four people on the sheet. Whether your relative got the long form was pure luck of the line number. |
The date attached to an age is not the day of the visit
Ages in the census are recorded as of the official census day, whatever date the enumerator actually knocked. The Archives lists the days: June 1 for 1850 through 1900, April 15 for 1910, January 1 for 1920, and April 1 for 1930 through 1950. Two consequences follow. A child born after census day should not be on the form even if the enumerator called in July, and the 1930 instructions say exactly that — include every person living on April 1, 1930, omit children born since. And an age that seems eleven months adrift between two censuses is usually not an error at all; it is the census day moving from June to April between 1900 and 1910.
The immigration and citizenship columns, and where they lead
Four censuses running record the year a person arrived in the United States — 1900 column 16, 1910 column 15, 1920 column 13 and 1930 column 22 — and the same four record citizenship status, coded Al for alien, Pa for first papers filed, and Na for naturalized. In 1950 the question shrinks to yes or no, with AP for a person born abroad of American parents. Treat these columns as a pointer, because that is what they are: “Pa” in 1910 means a declaration of intention existed somewhere, and finding the court that held it is a separate discipline with its own rules. If the arrival year is what you are chasing, the companion piece on picking up an immigrant relative’s trail takes it from there.
The military columns, and one specific trap
Column 30 of the 1910 schedule marks survivors of the Union or Confederate Army or Navy, and it is a genuinely valuable pointer into service and pension files. The Archives attaches a warning to it that is worth repeating exactly, because it has misled a great many people: columns 30 to 32 are often overwritten with numbers such as 2-1-0-0 or 6-9-0-0, and those numbers are not answers. They are data summaries added later by Census Bureau tabulators in Washington for statistical compilation. A “2-1-0-0” in that column tells you nothing about anyone’s war. Where the trail does run through a veteran, the descendant work is a different problem, set out in the piece on tracing living descendants of a Civil War ancestor.
One boundary worth naming rather than blurring: for a family enslaved before 1865, the population schedules are not the right instrument and no technique on this page makes them one. That research runs through Reconstruction-era federal records with an entirely different structure, and it is its own discipline rather than a variation on this one.
Why the Name Search Comes Back Empty
Four separate reasons, and only one of them is about your ancestor.
The commercial search box is a thin layer over a set of indexes built at different times by different people to different standards. When it returns nothing, the useful question is which of those layers failed.
Reason one: the year was never fully indexed
The Archives’ guide to using its census microfilm catalogs lays out the coverage, and it is much patchier than most people assume. 1880 is indexed only for families that included a child aged ten or under; if your ancestor’s household had no young child, the Archives’ own instruction is that you will have to search the community line by line. 1910 has a Soundex index for twenty-one states only — 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. Live anywhere else in 1910 and there is no surname index to fail; there never was one.
1930 is narrower still. There are Soundex indexes for Alabama, Arkansas, Florida, Georgia, Louisiana, Mississippi, North Carolina, South Carolina, Tennessee and Virginia in full, plus seven Kentucky counties (Bell, Floyd, Harlan, Kenton, Muhlenberg, Perry and Pike) and seven West Virginia counties (Fayette, Harrison, Kanawha, Logan, McDowell, Mercer and Raleigh). The Archives explains why in a single line: in the 1960s the Bureau of the Census prepared the Soundex cards, but only for the Southern states. A New York family in 1930 is not missing. It was never carded.
Reason two: Soundex is a sound code, and it has rules
Where an index does exist for the years 1880 through 1930, it is usually a Soundex, and understanding it turns a dead end into a hit. The Archives publishes the coding system in full. Every code is a letter and three digits: the letter is the first letter of the surname, and the remaining consonants are numbered 1 for B, F, P, V; 2 for C, G, J, K, Q, S, X, Z; 3 for D, T; 4 for L; 5 for M, N; and 6 for R, with A, E, I, O, U, H, W and Y disregarded and zeros added to fill. Washington codes to W-252. Lee, having no codable consonants after the L, codes to L-000.
The refinements are where searches go wrong. Doubled letters count once, so Gutierrez is G-362. Different letters sitting side by side that share a number also count once, so Pfister is P-236 and Jackson is J-250. A vowel between two same-coded consonants makes the second one count again, which is why Tymczak is T-522; an H or a W between them does not, which is why Ashcraft is A-261 and not A-226. And a prefix — Van, Con, De, Di, La, Le — must be coded both ways, because the card may be filed under either: VanDeusen is V-532 or D-250. Mc and Mac are not treated as prefixes. Anyone whose surname has drifted in spelling across generations should read that list twice; the same problem from the other end is covered in the note on relatives whose surname changed on arrival.
One practical detail: a Soundex card is a finding aid, not the record. What it hands you is four numbers in its upper right corner — volume, enumeration district, sheet and line. Those four take you back to the schedule, which is where the actual information lives.
Reason three: the 1950 index was written by a machine
The 1950 release did not arrive with a hand-made index. The Archives states that it used Amazon Web Services’ artificial intelligence and optical character recognition tool to extract handwritten names from the digitized population schedules, that the resulting index “is not 100-percent accurate,” and that it asks the public to submit corrections through a transcription tool on the 1950 census website. That is a candid and unusual admission from a federal agency, and it should change your behavior: for 1950, a name search failing is close to expected, and the enumeration district route is the primary method rather than the fallback.
Reason four: the handwriting, the informant, and the year that burned
Names were written down by an enumerator hearing them spoken, often by a neighbor rather than by the family. Spelling is phonetic. Given names appear in forms a family never used. And in 1940 the Archives notes a circled X marking whoever supplied the answers, which is the single most underused mark on the schedule: it tells you whether you are reading a mother’s account of her household or a lodger’s guess at it.
Then there is 1890, which is simply not there. Most of it was destroyed after a Department of Commerce fire in 1921, leaving a twenty-year hole between 1880 and 1900 that no amount of searching closes. Families that moved during those two decades are the hardest cases in American genealogy for exactly that reason, and the general approach to a stalled line is set out in the guide to working a genealogy brick wall with locate technique.
Six Ways a Census Entry Misleads a Researcher
Each of these is a real document behaving exactly as designed, and being read wrong.
Treating a relationship label as a legal fact
The relationship column records what one person told an enumerator on one morning. Grandson, ward, boarder, niece and adopted daughter were all used loosely, and the same child can be described three ways across three decades. It dates an arrangement. It does not prove one.
Assuming the household was asked the long questions
In 1950 only six people per sheet fell on a sample line, and five versions of the form existed so the sampled lines varied. Income, schooling, parents’ birthplaces and military service exist only for those six. An empty column is usually not a refusal to answer.
Searching a city under its county prefix
Chicago’s enumeration districts start with 103, not with Cook County’s 16. Independent and very large cities were numbered separately, so a methodical search under the wrong prefix reads dozens of correct pages about entirely the wrong people.
Reading a tabulator’s arithmetic as an answer
The strings overwritten across columns 30 to 32 of the 1910 schedule, of the form 2-1-0-0, are Census Bureau statistical codes added in Washington. They are not military service, not children, not anything about the household. The Archives says so on its own reference page.
Concluding absence from a null index result
Twenty-nine states have no 1910 surname index. Most of the country has no 1930 one. An 1880 household with no child under ten was never carded. In each case the index is silent because it does not cover that ground, not because the family was elsewhere.
Expecting a college student at the family home
In 1950 students were counted where they lived while attending school rather than at a parental address, and service members sleeping off post were counted where they slept. A missing eighteen-year-old is frequently on a page two hundred miles away, in the right year, correctly recorded.
How We Work a File That Begins in the Schedules
Four stages, in this order, because each one produces the input the next one needs.
Pin an address to a census year
Not a city and not a county — a street and a number, dated to within a year or two of the count. City and telephone directories are the usual source. Without this the geographic search has nothing to consume, and everything downstream is guesswork.
Resolve the district from map and description together
The map gives the boundary and the number; the written description gives the streets in words. We read both, because a family on a corner lot can sit either side of a line, and a district resolved from one source alone is a district you will re-do.
Read the pages, then read the neighbors
We read the whole district rather than stopping at the first matching surname, and we record who else is on the page. Married daughters, in-laws and siblings cluster on adjacent lines far more often than chance allows, and that adjacency is frequently the finding.
Carry the citation with the fact
Year, state, county, enumeration district, sheet and line, with the publication number where the source is microfilm. A name without those six items is an assertion. With them, anybody can order the same page and check us.
The Limits This Particular Record Puts on Us
Historical research that ends with a living person’s address needs its rules stated at the front, not in a footer.
We are a skip-tracing and public-records research firm, and nobody here holds a private investigator’s license. We are not licensed private investigators and we never describe ourselves as investigators to an archivist, a librarian or a county clerk. We read records, we file written requests under our own name, and we say who instructed us. Every file we take begins with a purpose the law permits, written down before the first search runs: an estate that has to identify its heirs, a lineage application that needs a generation proved, a probate court that requires evidence of a diligent search, a family reconstructing itself. Curiosity about a stranger’s address is not on that list, and we would rather say so at the first email than take the work.
We do not pretext. Nobody from this firm telephones an archive claiming to be a relative, lets a convenient wrong assumption stand, or misrepresents who is paying for a request in order to get a file opened. On census work the temptation is small and the principle is not: an archivist who knows exactly what we are doing helps us, and one who has been misled tells the next caller. Nor do we go after what is inside a private financial account — balances, statements, account numbers. A bank owes its customer confidentiality, we are not the customer, and the standard workaround is the impersonation refused a sentence ago. Nineteenth and twentieth-century federal population schedules are the opposite kind of record: open by statute, available to anybody, and cited openly.
Where the FCRA line sits, and why genealogy crosses it
A family tree is not a consumer report, and this firm is not a consumer reporting agency. A pedigree assembled here is not an input to a decision about anybody’s employment or promotion, their tenancy or rental application, their credit, their insurance, or a professional license or benefit they have applied for. The Fair Credit Reporting Act governs that entire category and assigns it to regulated screening firms, which are required to ask permission, disclose what they hold, send an adverse-action notice and staff a dispute process. We do none of those things because we take none of that work. Where a request turns out to be one of those decisions wearing a genealogy costume, we decline it.
The safety decline, and why this subject invites it
We decline any file where the person at the end of it appears to have stepped away deliberately. Genealogy is a very easy thing to hide behind: it sounds harmless, it supplies a warm reason for wanting an address, and an outsider has almost no way to test it. So a census inquiry that ends in a request for a living person’s current address is questioned harder at intake here, not waved through. Where what we hear points to a stalking pattern, an order of protection, a custody dispute being run around a courtroom rather than inside one, or a relative who has already said in plain words that they want nothing to do with the enquirer, the file closes. Rewording the instruction more carefully the second time does not reopen it. Nobody is owed a relative’s address. Distance is sometimes a decision, and it is not ours to undo.
Two variants come up often enough to name. One is the descendant approach: somebody has matched a surname in a 1920 schedule and wants an introduction to a household that has never heard of them. We will carry a letter and make it easy to ignore; we will not hand over an address so that a doorstep visit can be a surprise. The other is the estate angle, because unclaimed inheritances cluster around exactly this research and attract people who are not what they say they are. If a stranger has approached you about an estate, put three questions to them in writing before anything else — whose estate, which court, what case number — and never sign a document you have not read to the end. Anybody doing this legitimately will answer all three on the spot.
Everything here is general information about federal records and is not legal advice. Whether a census entry establishes a legal relationship, who inherits, and what any given court treats as sufficient proof of descent all belong with an attorney licensed where the question has to be answered. We produce the page and the citation under it. What that proves in law is not ours to say.
Who Brings Us a Census Problem
Six kinds of client, one set of schedules, six entirely different questions put to it.
Lineage society applicants
Needing one generation documented to a standard a membership committee will accept.
Land title examiners
Reconstructing occupancy and ownership across decades where a chain of title has gaps.
House history researchers
Working an address forward and backward through the districts that covered that block.
One-place study groups
Documenting every household in a village or parish rather than a single family line.
Local history librarians
Fielding the same district question from patrons and wanting a repeatable answer.
Documentary producers
Verifying that a family really was in the place a story says they were, in the year it says.
The house history researchers are the group who get the most out of the enumeration district method, because they arrive with the one input everybody else is missing: they already have the address. For them the census reverses. Instead of hunting a family and hoping to learn where it lived, they take a known street number, resolve the district, and read out every household that occupied the property across seven decades. Whether that address happened to be owned or rented is on the schedule too, coded O or R, with mortgaged homes marked M and unencumbered ones F. It is one of the very few historical searches that runs cleanly from place to person, which is also why it is the one most likely to turn up a name nobody in the family remembered — often a married daughter under a surname the family lost track of, which is the situation the note on searching under a maiden name addresses.
Every entry comes back with its district, sheet and line
No finding leaves this office without the coordinates that let you check it: the year, the state and county, the enumeration district number, the sheet, the line, and the publication number where the source sits on microfilm. Order the same page and you should see the same thing. Where the honest answer is that the district was searched and the family is not in it, we write that up as a result with its own citations rather than describing the file as still open — and a present-day locate that does not produce a person is not billed, because nothing was located.
Questions About Reading the Federal Population Schedules
Can I use census records to find where someone lives now?
No, and it is better to hear that on day one. Returns that identify individuals are closed for seventy-two years, so the newest census open to the public was taken in the spring of 1950 and released on April 1, 2022. Anybody born after that has no public census entry at all, and anybody who does have one was a child when it was written. The census is a historical instrument for reconstructing a household in a place and a year. A current address comes from a completely different set of records, and any page that suggests otherwise is selling you a subscription.
Why can’t I see anything after 1950, and when does the 1960 census open?
April 2032, according to the Census Bureau’s own genealogy answers. The interesting part is where the rule comes from. 44 U.S.C. 2108(b) does not state a period at all; it directs the Archivist to release identifying census data “pursuant to the specifications and agreements set forth in the exchange of correspondence on or about the date of October 10, 1952, between the Director of the Bureau of the Census and the Archivist of the United States.” That correspondence, between Census Director Roy V. Peel and Archivist Wayne Grover, set the seventy-two years. The Bureau disowned it in 1972, arguing it conflicted with the confidentiality duty in 13 U.S.C. 9, and Congress settled the question by statute in Public Law 95-416 on October 5, 1978.
What is an enumeration district, and how do I find the number?
It is the patch of ground assigned to one census taker — formally, an area an enumerator could completely cover within two weeks in cities and within four weeks in rural areas. Districts run from a few city blocks to an entire rural county. The number has two parts separated by a hyphen: the prefix is usually the county and the suffix the specific district inside it, except that independent and very large cities get their own prefix, which is why Chicago’s districts begin 103 while Cook County’s begin 16. To find yours, get an address for the census year from a city or telephone directory, then search the National Archives Catalog for the year, county and state to pull up the district map and the written description.
The name search returns nothing. What do I do next?
Stop searching by name and search by place. The National Archives’ own answer is that you can locate people by identifying the enumeration district in which they lived and then browsing the population schedules for that district. This is not a workaround; before the commercial indexes existed it was the only method there was. You need an address from roughly the right year, the district map, and patience for a few dozen pages. For 1930, check first whether your city is one of the fifty-plus covered by microfilm publication M1931, which cross-references street addresses directly to district numbers and removes the map step entirely.
What is Soundex, and why isn’t my state indexed?
Soundex is a surname code — first letter plus three digits derived from the consonants, so Washington is W-252 and Lee is L-000 — designed so that variant spellings file together. The coverage is the problem. The 1880 index only covers households containing a child aged ten or under. The 1910 index exists for twenty-one states and no others. The 1930 index covers ten Southern states in full plus seven Kentucky and seven West Virginia counties, because the Bureau made the cards in the 1960s for the Southern states only. If your state is not on those lists, no surname index was ever created, and the geographic search is the only search.
Which census year should I look at for immigration or citizenship?
It depends on which fact you need. Year of arrival appears in 1900 column 16, 1910 column 15, 1920 column 13 and 1930 column 22. Citizenship status appears alongside it in each of those years, coded Al for alien, Pa for first papers filed and Na for naturalized. But the year a person actually naturalized was asked in one census only, 1920, in column 15. By 1950 the question is reduced to yes or no, with AP for someone born abroad to American parents. Treat these as pointers rather than proof: they tell you a court record should exist and roughly when, and finding the court that holds it is separate work.
Why is my ancestor’s age different in every census?
Partly because people misremembered or misreported, and partly because of something structural that catches everyone. Age is recorded as of the official census day rather than the day the enumerator called, and that day moved: June 1 for 1850 through 1900, April 15 in 1910, January 1 in 1920, and April 1 for 1930 through 1950. A gap that looks like a ten-months error between the 1900 and 1910 entries for the same person is often just the census day sliding from June to April. Use age to track a person across decades and to bracket a birth year, never as a date of birth.
Can you do the census research for me, and can you find the living relative at the end of it?
Yes to the first, and conditionally to the second. The archival work — pinning an address, resolving the district, reading the schedules and citing what we find — is measured in weeks, because it runs through catalog searches, microfilm and library holdings. Locating a living person on a line we have documented is separate work with its own rules, and where we have a full name and a reasonable starting point that usually comes back within 24 hours. A lawful purpose has to be stated first; any approach to the person names this firm and the party who instructed it; and saying no is made the easy option rather than the awkward one. Where the circumstances suggest somebody has stepped away on purpose, we decline the file rather than deliver the address.
You Have a Street and a Year. Let Us Find the District
Send the surname, the town, and whatever address you have for any census year between 1880 and 1950, and we will tell you which years are indexed for that state, whether the district can be resolved from the address you have, and what the schedule for that decade will actually contain — before anybody commits to anything. If the answer is that your state has no surname index and this will be a page-by-page read, we would rather say that first. Tell us which household you are trying to place and we will tell you where the paper is.
Open a records research file
