CousinChart — family connections made simple

State Census Records: Which States Counted, and When

Almost every census guide tells you to "try a state census" and then walks away without saying which states took one. Here is the short answer: roughly half the US states counted their own people at some point, most often in the years ending in 5, which drops a snapshot neatly into the middle of every federal decade — and for several states it also puts a name-by-name list either side of the destroyed 1890 federal census. New York, Iowa, Kansas, Minnesota, Wisconsin, Florida, New Jersey and Rhode Island have the strongest runs. Some of them asked questions Washington never thought to ask. If a state list hands you a household of half-placed relatives, the cousin calculator will sort out how they connect.

Handwritten state census schedule page listing families with ages and birthplaces

Which state census should you try first? Tap the closest match.

Tap an option above. State censuses sit in the middle of the federal decade, which is exactly where families move.

Work out how two relatives are related →

What a state census actually is

A state census is a head count run by a state or territory under its own law, on its own timetable, for its own reasons.

The reasons were usually money and seats. States needed population figures to apportion their legislatures, draw school districts, and share out state funds. Waiting ten years for Washington was no use to a state filling up with settlers, so many wrote their own count into the state constitution.

That is why so many landed on the years ending in 5. Splitting the federal decade gave a state a fresh number roughly every five years. For you, the effect is a second set of eyes on the family halfway between two federal counts — the exact point at which a household is most likely to have changed shape.

They are separate records, held separately. A state census was never sent to Washington and is not part of the National Archives federal series. The originals sit with the state archive or the state historical society, which is why they are missing from the big census search boxes that only cover the federal returns.

And they are not bound by the federal 72-year privacy rule. That rule applies to the federal count alone. Florida 1945, South Dakota 1945 and Rhode Island 1935 are all open, which makes them some of the most recent named lists of ordinary people you can read.

Quality varies enormously. Some are as good as a federal schedule or better. Some are a single tally sheet counting males over 21. Knowing which you are dealing with before you go hunting saves a wasted evening.

Which states counted, and when

This is the reference table the rest of the census cluster keeps pointing at. Years shown are the substantial ones that name people or heads of household; a few statistical-only years are left out.

StateMain census yearsWorth knowing
New York1825, 1835, 1845, 1855, 1865, 1875, 1892, 1905, 1915, 19251855–1875 name everyone; 1892 covers the 1890 gap; pre-1855 are heads of household with tallies
Iowa1836–1856 (several), 1885, 1895, 1905, 1915, 19251925 names both parents of every person, plus the mother's maiden name
Kansas1855–1859, then 1865, 1875, 1885, 1895, 1905, 1915, 1925Asks the state or country the family came from before Kansas
Minnesota1849, 1853, 1857, 1865, 1875, 1885, 1895, 19051905 is the fullest, with birthplaces of parents
Wisconsin1836–1895 (mostly heads only), 19051905 is the one that names everybody
Florida1867, 1875, 1885, 1895, 1935, 19451935 and 1945 are open and unusually recent
New Jersey1855, 1865, 1875, 1885, 1895, 1905, 19151895 lists names but groups ages in bands
Michigan1845, 1854, 1864, 1874, 1884, 1894, 1904Survival is patchy and varies county by county
Rhode Island1865, 1875, 1885, 1895, 1905, 1915, 1925, 1935One of the longest unbroken runs anywhere
Massachusetts1855, 1865Full detail, right in the middle of two federal decades
South Dakota1885, 1895, 1905, 1915, 1925, 1935, 1945Later years ask church affiliation
Illinois1810–1865, roughly every five yearsHeads of household with tallies, not full names
Colorado, Nebraska, New Mexico, Dakota Territory1885Federally funded, so they read like a federal schedule

Treat this as a starting map, not gospel. Coverage within a state can be complete for one county and lost for the next, so confirm the specific county and year in the state archive's own catalog before you conclude anything is missing.

How state censuses bridge the 1890 gap

The 1890 federal census burned, and the loss is the single biggest hole in US research. State censuses are the nearest thing to a patch.

The gap is twenty years wide, not ten. Without 1890 you jump from 1880 to 1900. A child born in 1882 and dead by 1898 can pass through the whole federal record without ever appearing. That is how whole siblings vanish from family memory — the 1880 census has them as infants and the next one has nothing.

A state count in 1885 or 1895 halves that gap. Iowa, Kansas, Rhode Island, Florida, New Jersey, Michigan, Minnesota, South Dakota and Wisconsin all ran one or both. Two of those states — Iowa and Kansas — have both, so you get 1885, 1895 and 1905 as well as the federal years.

New York 1892 sits almost exactly on the missing year. It is not the strongest New York schedule, and the ages are rough, but it puts a named household on paper within two years of the burned count.

Colorado, Nebraska, New Mexico and Dakota Territory ran federally funded 1885 censuses. Congress paid for them, so they were taken to federal standards and are held with the federal records rather than the state ones. For those places, 1885 reads like a proper mid-decade federal schedule.

Use them the way you would use the missing census. Take the household you know from 1880, the household you know from 1900, and read the state schedule between them for anybody who appears in one list and not the other. What our page on the destroyed 1890 census calls the substitute strategy is really just this, done systematically.

The questions the federal census never asked

This is the part most people miss. Several state schedules asked things Washington did not, and a couple of those questions are worth more than the whole federal line.

Iowa 1925 asks for both parents by name. Every person on the schedule gets their father's name, their mother's full maiden name, and the place where the parents were married. There is no other US census that hands you a maiden name for an adult woman as a matter of routine. If your line touches Iowa in the 1920s, search this first and everything else second.

New York 1855 asks how many years the person had lived in that county. Subtract, and you have the year the family arrived — which tells you where to look for them before that, and roughly when they moved.

New York 1855 also asks New York natives for their county of birth. For a state that people crossed steadily from east to west, that single column can move you a hundred miles in the right direction.

Kansas asks where the family came from. Every schedule from 1865 onwards records the state or country of previous residence, so a Kansas farmhouse full of children becomes a migration trail.

New York 1865 asks about marriage and war. It records whether a person had been married more than once and whether a woman had been widowed by the war, which flags a remarriage you would otherwise miss.

Later schedules often ask about citizenship and years in the country. New York 1905, 1915 and 1925 do, which narrows an arrival date before you ever open a passenger list.

What state censuses will not do

Being honest about the limits saves you from concluding a family was not there.

Many early ones only name the head of household. Illinois before 1865, Wisconsin before 1905 and New York before 1855 mostly give one name and then tally the rest of the household by age and sex. Useful for placing a family, useless for naming a child.

Some are purely statistical. A handful of years survive as county totals only, with no household detail at all. The state archive's catalog will tell you which — the phrase to look for is "aggregate" or "statistical returns".

Losses are common and unevenly spread. Michigan is the classic example: some counties have a good run, others have almost nothing, and the reason is usually that a county clerk threw out the local copy after sending the state its figures.

Ages and spellings are often worse than federal. State enumerators were frequently local men paid a small fee per name. Expect approximate ages, phonetic surnames, and households recorded from a neighbor's memory. Compare across years rather than trusting one line.

And a good half of the country simply never counted. Most of the South after Reconstruction, and much of the interior West, has little or nothing. If that is your state, the substitute is not another census — it is tax lists, city directories and church registers.

Where to find them free

Start with the state archive or state historical society. They own the originals and they publish the definitive list of what survives. The Kansas Historical Society, the New York State Archives and the State Historical Society of Iowa all put finding aids online, and several host free images.

FamilySearch has the largest free digitized set. Search the catalog by state and by county rather than using the main search box, because a great deal of it is browsable image-only and never appears in a name search. A free account is required and worth having.

Ancestry holds indexed versions of the big runs, including New York, Iowa and Kansas. The Library Edition at most public libraries includes them at no cost to you, which is the cheapest route if the free images are unindexed.

Some states have their own free portals. Minnesota's historical society, for instance, offers a free searchable index to its state censuses. It is always worth ten minutes with a search engine on "[state] state census index" before you pay for anything.

Do not assume the microfilm is the only copy. County clerks sometimes kept their own set, and those local copies occasionally survive where the state's has not.

How to search when the index fails

Unindexed does not mean unsearchable. It means you search by place instead of by name.

Work out the exact county and township first. State schedules are filed by county, then by minor civil division. If you know the family was in one township, you are browsing perhaps thirty pages, not thirty thousand.

Get the place right for the year, not for today. A family that appears in one county in 1875 and a different one in 1885 has very often not moved at all — the line moved. Check county boundary changes before you decide they migrated.

Use the federal census either side as your map. The federal search method gives you the township and often the neighbors. Those same neighbors will be a few lines away on the state schedule.

Read the whole page, then the page before and after. Married children, elderly parents and hired hands with the family surname cluster within a page or two. This is where the extra relatives come from.

Search the surname the way it sounds. A local enumerator writing at speed produced spellings no index will match. If you are browsing anyway, your eye is a better tool than a search box.

Turning a state census into a broken wall coming down

The mechanics matter less than the sequence. Here is the order that tends to work.

Fix the family in the two federal years either side. Note every person, their stated age, and their birthplace. You now have a before and an after.

Read the state schedule in between and list every difference. A new child, a missing child, a widow now heading the household, an in-law who appeared. Each difference is a dated event you can go and find a record for.

Convert the extra columns into searches. Years in the county becomes an arrival year. Previous state becomes a place to search deeds and tax lists. Parents' names in Iowa 1925 become a marriage record to order.

Treat a disappearance as a fact, not a dead end. A man in 1885 and gone in 1895 died, moved, or was missed. Probate, the local newspaper and the cemetery will tell you which, and that is often the moment a stubborn line finally gives, as our guide to breaking through a brick wall works through in more detail.

Write down what you searched and did not find. Negative results in an unindexed browse are genuinely valuable, because in six months you will not remember which fourteen townships you already read.

FAQ

Which states have the most useful state censuses?

New York, Iowa, Kansas, Minnesota, Wisconsin, Rhode Island, New Jersey and Florida have the longest and most detailed runs. Iowa 1925 is the single most informative state census in the United States, because it names both parents of every person listed, including the mother's maiden name.

Do state censuses cover the destroyed 1890 federal census?

Partly. Iowa, Kansas, Rhode Island, Florida, New Jersey, Michigan, Minnesota, South Dakota and Wisconsin took censuses in 1885 or 1895 or both, and New York took one in 1892. They do not replace 1890 nationally, but for those states they cut a twenty-year gap roughly in half.

Are state census records free to search?

Often, yes. FamilySearch has the largest free digitized collection, many state archives publish free indexes or images, and Ancestry's Library Edition is free inside most US public libraries. Indexing is patchy, so free browsing by county is frequently faster than paying for a name search.

Why can't I find my state on the list?

Roughly half the states never ran a meaningful census of their own, and much of the South and interior West has nothing after the territorial period. In those places, annual tax lists, city directories and church registers cover the same middle-of-the-decade gap.

Are state censuses as accurate as the federal census?

Generally a little less so. Enumerators were often locals paid per name, so ages are rounded, surnames are spelled phonetically, and households were sometimes recorded from a neighbor's account. Use them for placing and dating a family, and confirm details against vital or church records.

Does the 72-year privacy rule apply to state censuses?

No. That rule covers the federal census only. Several state censuses from the 1930s and 1940s — Florida 1935 and 1945, South Dakota 1945, Rhode Island 1935 — are fully open, making them among the most recent named lists of ordinary people available to researchers.

The five-year snapshot worth building into your routine

Once you know your state took a census, the habit is simple: every time you place a family in a federal year, check whether a state schedule sits between that year and the next. It costs ten minutes and it catches the children who were born and died inside a decade, the widow who remarried, and the grandmother who moved in and then out again. For New York, Iowa and Kansas researchers especially, the state schedules are not a fallback — they are better records than the federal ones, with extra columns you cannot get anywhere else. When a new household turns up full of people you cannot immediately place, drop the names into the cousin calculator and see exactly where they sit on your tree.