How Accurate Is an Obituary? Ask Who Wrote It

You have his obituary from the local paper. It gives the date he died, the town he was born in, both parents' names and the five children who survived him. Then the death certificate arrives, and it agrees on almost every point. Two sources, one story. It feels settled.

It may not be. In many cases the obituary and the death certificate were filled in by the same person, a daughter or a widow, in the same few days, from the same memory, at the same funeral home. When they agree, you may be reading one witness twice.

That is not a reason to distrust obituaries. They are among the richest sources a family historian has for the twentieth century. It is a reason to ask, of every fact in one, who would have known it and how.

The short answer. Unless your ancestor was newsworthy, the obituary you have is most likely a death notice, which The New York Times describes as usually written, submitted and paid for by the family. The personal details on a United States death certificate come from an informant, who the Centers for Disease Control and Prevention's handbook for funeral directors says is in most cases a member or friend of the family, so an obituary and a certificate that agree are often not independent. Obituaries are strongest on what the family knew first-hand, the death, the funeral and the survivors, and weakest on birth dates, birthplaces and parents' names from decades earlier. They also miss many deaths entirely. A 2025 study of Canadian online obituaries found they covered about half of recorded deaths in Quebec and a third in Ontario, and very few infants.

A death notice is usually written by the family

The word obituary covers two different documents, and the difference decides how much weight one can bear.

The news obituary is journalism. The New York Times describes its own process: editors scan news reports and readers' emails, then decide whether to assign a reporter to write about a life.

The death notice is the far more common kind. The New York Times, explaining how to suggest an obituary, puts the distinction in one line: a Times obituary is different from a death notice, which is usually written, submitted and paid for by the family. For an ordinary family, the notice you find is far more likely to be this second kind, whatever the column heading says, because the first kind is written only about the lives a paper chooses to cover.

The same is now true online. In a study of Canadian obituaries published in Demographic Research in October 2025, the demographers Pietro Violo and Nadine Ouellette of the Université de Montréal worked mainly from notices that funeral homes publish on their own websites. Quebec had 224 registered funeral service providers in 2024 and Ontario licensed 546 funeral establishment operators, and most of them, the authors note, publish death notices online for upcoming funerals.

So the first question about any obituary is the same as the first question about any record: who supplied the information, and were they in a position to know it?

The death certificate's details come from the same people

This is the part that changes how you should use the two documents together.

A United States death certificate has two halves. The cause of death comes from a doctor or a medical examiner. Everything else, the birth date, the birthplace, the parents' names, the occupation, the spouse, comes from an informant. The CDC's Funeral Director's Handbook: Death Registration and Fetal Death Reporting, in its 2019 edition, says the funeral director is responsible for completing and filing the certificate, and should obtain the personal information from the best-qualified person. In most cases, it says, that is a member or friend of the family, and it lists the order of preference as spouse, parent, child, another relative, or another person with knowledge of the facts. The informant's name, relationship and address must appear on the certificate.

Put the two documents side by side and the shared source becomes obvious. The family member who sat with the funeral director to answer the certificate's questions may well be the person who wrote the notice for the paper, or helped the funeral home write it, in the same week.

The judgement that follows is ours. When an obituary and a death certificate agree on a birthplace, you have not got two independent witnesses. You may have one, recorded twice. Look at the informant's name on the certificate, ask whether the same person was likely to have supplied the notice, and if so, count them once.

The useful corroboration comes from records made by someone else, at another time: a baptism written by a priest soon after the birth, a census answered while the person was alive, a marriage record where the bride gave her own parents' names.

What an obituary gets right, and where it slips

Not every fact in an obituary is equally weak. The test is how close the writer was to the event.

The death and the funeral are usually right. The family knew the date of death and arranged the funeral themselves, days before the notice was written. In Ireland the gap can be very short: the Central Statistics Office, which has used death notices on the website RIP.ie to track mortality since 2020, found that the average time between a death and the notice appearing online was about 1.3 days.

The survivors are usually right too. The person writing the notice knew who was alive and where they lived. The list of children, siblings and grandchildren, with their married names and towns, is often the most valuable thing in an obituary, because few other records gather a whole family in one place at one date.

Anything from decades earlier is secondhand. A daughter writing her father's notice in 1962 did not witness his birth in 1889, or know his parents as adults. What she wrote is what she was told, and it has all the weaknesses of a family story: a real core, compressed and sometimes rearranged in the retelling.

Ages are slipperier than they look. Violo and Ouellette found the phrase "in his (or her) XXth year" in more than 50,000 English-language obituaries. They point out that it is ambiguous: it can mean the person had completed that many years, or that they were one year younger. On the second reading, a man in his 80th year was 79. Work out both readings before you calculate a birth year.

Clerical slips happen in the notice itself. The same study found adults listed with identical birth and death years, the format usually reserved for infants, through simple errors in writing the obituary. The authors concluded that most of the mistakes they met came from inconsistencies or omissions in the original texts, not from their own methods.

Many people never got an obituary at all

An obituary search that comes back empty is weaker evidence than it feels, because a large share of deaths never produced one.

In Canada, about half or fewer. Violo and Ouellette web-scraped 236,290 obituaries from Quebec and 288,623 from Ontario covering 2017 to 2022, and compared them with every death in the official vital statistics. Obituaries accounted for only about half of recorded deaths in Quebec and a third in Ontario. In Quebec, coverage held at roughly 38 to 45 per cent from age 25 to about 70 and peaked at around 50 per cent between the ages of 85 and 90. Women were slightly better covered than men.

Infants and young children almost disappear. Only 4.2 per cent of infant deaths in Quebec appeared in obituaries, and about 6 to 7 per cent in Ontario, the lowest coverage at any age. For a child who died young, the absence of a notice says almost nothing.

Source: Violo and Ouellette, Demographic Research, 2025, comparing scraped obituaries with official vital statistics.

Other places look different again. The same paper reports a German study by Schnell and Redlich in 2019, which found that obituaries in one town's online newspapers represented only about 18 per cent of deaths and under-represented men and people born abroad, and an Austrian study by Mayer, Berger and Oberndorfer in 2023 that used the size of a newspaper obituary as a measure of the deceased's social standing.

That last finding points at an older problem. Violo and Ouellette describe the costly, restricted space of newspapers as a source of selection bias, and it is a fair inference that families with less money bought less of it. On the evidence above, the youngest and, in the German study, men and people born abroad were the least likely to get a notice at all, and they are often the people a family historian is hunting.

So if you search a paper and find nothing, record the search as negative evidence, with the paper and the dates you covered, and do not let it count as evidence that the death happened somewhere else.

The notice often comes before the official record

One genuine strength of a death notice is speed.

In Ireland, the Central Statistics Office explains that families have three months to register a death, and that an analysis by the General Register Office of deaths between 2016 and 2019 found about 82 per cent were registered within that time. Death notices, placed within a day or two because Irish funerals usually take place within two to three days, let the Office estimate mortality between 12 and 24 months earlier than official data.

For a family historian, the lesson is narrower. A notice can establish that a death happened, and roughly when and where, in cases where the civil registration was late, lost or filed in another district. It is a pointer to where the official record ought to be, and sometimes the only contemporary statement of a death that happened away from home.

How to weigh an obituary against everything else

The method below is ours. It is ordinary evidence analysis applied to one kind of document.

Name the likely informant. It may be the informant named on the death certificate, or the closest relative among the survivors. Write their name next to the obituary in your notes.

Sort the facts by distance. Separate what that person witnessed, the death, the funeral, the living family, from what they could only have been told, the birth, the parents, the immigration story. Give the first group real weight. Treat the second as leads.

Do not count the death certificate twice. If the certificate's informant is the same person, or the same household, its personal details are not independent of the obituary.

Look for the survivors in other records. A married daughter in another town is a lead to her own marriage record, which may name her parents in her own words.

Compare the age both ways. If the notice says "in his 80th year", calculate a birth year from 79 and from 80, and see which one the other records support.

The same discipline, weighing a record by who made it and why rather than by how official it looks, applies to a will: as we argued about what a pre-1858 English will proves, the dates in a probate file bracket a death without stating one.

What KleioBase does with an obituary, and what it does not

Disclosure, since we sell a tool.

KleioBase is an AI research companion that turns scattered historical records into one connected family archive. An obituary is one of the documents that benefits most from that, because a single notice can name a dozen people who each have their own records elsewhere.

Upload a scan or photo of the clipping, as a JPG, PNG or PDF, and the people, dates, places and relationships in it are extracted, so each survivor becomes a person you can link to their own baptism, marriage or census entry. Each record stays a fixed snapshot of what its document said, while the person's profile holds your conclusion, so when the obituary and a census disagree on a birth year you can see exactly what each one says instead of having one quietly overwrite the other. When a newspaper search comes up empty, you can record the search that found nothing against the person, with the paper and the period you covered.

Now the honest limits. Three of them.

It does not know who the informant was. It extracts what the obituary says. Deciding that the obituary and the death certificate came from the same daughter, and should count once, is your judgement.

It will not tell you an obituary exists. On the Researcher plan and above, Expected records suggests missing census, military, birth, marriage, death and burial records, but it does not suggest newspapers. Finding the notice is search work you do first.

It reads the page you give it. If a notice was cut from the paper without its date or masthead, KleioBase cannot supply them. Note the newspaper and the date on the image, or in the record, before you lose track of them.

Where to start with the obituary you already have

Take one obituary and write down, next to each fact, who could have known it first-hand.

Find the death certificate if you do not have it, and read the informant line. If it names the same person who likely wrote the notice, mark the matching facts as one source.

Then take the list of survivors and look for each of them in the next census and in their own marriage records. Those were made by other people, at other times, and they are where an obituary's family facts can actually be tested.

The daughter who wrote her father's notice, in grief and in a hurry, was not trying to deceive anyone. She was writing down the family as she knew it, for the neighbours who needed to know when to come to the funeral. Read her notice as that, a witness statement from someone who loved him, and it will tell you a great deal.