Working With Historical Obituary Archives
The Faces Of The Past Uncovering The Stories Of The Daily Item Obituaries Today
I spent three weeks last year trying to pull complete obituary records from a defunct local paper called the Daily Item that ran from 1962 to 1989. What I learned isn't worth a TED talk, but it will save you about fourteen hours of head-banging if you ever need to do this yourself. Start with the microfilm. Everyone assumes the digitized databases will have everything. They don't. The online indexes typically cover the last twenty years of any given publication. The earlier material, especially the pre-1990s stuff, lives on microfilm or in the basement of whatever library system once bought the physical copy. My first mistake was searching Ancestry and Legacy.com and then trying to figure out why my results were empty for anything before 2001. They simply don't index that far back reliably. The workaround is straightforward. Call the local historical society or the county library where the Daily Item had its distribution radius. Ask them to run the microfilm reader for the date range you need. Most libraries will let you do this for free, and they'll actually pull the film for you rather than making you stand there squinting at a tiny screen. One tip: bring your own digital camera or phone. The readers sometimes have a port you can feed through, but more often you just photograph the page directly. Lighting matters less than you'd think because microfilm contrast is already baked in.
When you get the images, don't expect them to be clean. Obituaries from the seventies and early eighties were typeset by hand on photo-composition machines. That means you'll see smudges, misaligned columns, and occasional missing words where the operator bumped the type case. I found myself spending twenty minutes just trying to read one particular obituary because the deceased's middle initial had been set slightly crooked and looked like a punctuation mark. The fix was to load the image into a basic OCR tool, run it, then manually correct the output. Tesseract handles this okay, but the real time-saver is just reading the image yourself. OCR introduces more errors than it saves on aged microfilm. Here's the thing nobody tells you about obituary research. The most useful data isn't in the text itself. It's in the margins. I've found phone numbers written in pencil on the proof copy, correction notes from the editor, and sometimes the names of surviving relatives who weren't listed in the formal obituary but appeared in the layout notes. The Daily Item archives I worked with had a binder of corrected proofs from 1974 to 1978 sitting in a drawer labeled "misc." I asked the archivist about it and she said she'd been meaning to catalog it for two years. Those correction notes gave me three family relationships that the published obituary completely omitted. If you're doing this for genealogy or family history, cross-reference the obituary with the death certificate. Obituaries are written by grieving relatives or funeral homes, and they contain factual errors about dates, locations, and even the order of survivors. Death certificates are filed with the county and contain different information that fills in the gaps. The overlap between the two sources is where the actual truth lives.
One edge case that burned me: the Daily Item published supplementary death notices separate from the full obituaries. The full obituaries ran in the main news section, but the compact death notices — just the basic vital statistics and service times — were in a classified column at the back of the paper. If you only search the digital index for "obituary," you'll miss the classified section entirely because it's tagged differently. I spent an afternoon going back through the microfilm specifically looking for the classified ads section and found about forty notices that the index had never captured. The workaround was to skip the database search and just run through the issue chronologically, noting which pages contained the classified section. For anyone building a database from these sources, the format consistency is the real bottleneck. Daily Item obituaries from the seventies use one layout. By the eighties they switched to a template that included standardized fields. The transition happened around 1983, and the articles before and after that date don't follow the same structure. If you're extracting data programmatically, build two separate parsers or write a single script that detects the layout pattern first. I wrote a Python script that checks for the presence of certain header strings to determine which parser to use, and it cut my extraction time from about six hours per issue down to maybe twenty minutes once I got it working. The files you'll need to preserve are the raw scanned images, the transcribed text, and a log of which source each fact came from. Don't skip the source log. I made the mistake of transcribing everything into a single spreadsheet and then realizing three months later that I couldn't tell which information came from the obituary versus the death certificate versus the family-supplied correction note. Re-triangulating that took another full day of work I didn't need to do.
Get the Full Details
If you're looking to download or access the Daily Item obituary collection directly, there's no single consolidated repository. The microfilm is held by the state library system and the digitized versions are scattered across commercial genealogy sites that charge per-record or require a subscription. The most cost-effective path is usually the microfilm route through your local library, which is free and lets you pull exactly what you need without paying per-document fees. The work is tedious but the payoff is real. Obituaries from this era contain details that no other record preserves: the names of siblings who moved away and lost touch, the specific street addresses where people grew up, the union or lodge affiliations that defined community life in small towns. Those details disappear from official records and live only in the text of the notice. I've found people tracing ancestry back four generations using nothing but the marginalia in obituaries that other researchers ignored because they were focused on the formal text. One final practical note. If you're working with fragile microfilm, the images will degrade over time with repeated viewing. The library's handling is generally careful, but ask about the condition of the film before you commit to a long project. I pulled one particular reel that had been spliced poorly and the damage appeared as a vertical black line running through about thirty percent of the frames. The archivist offered to pull a backup copy from their reference reel, which was in better shape. Always ask for the backup. It saves you from discovering the damage after you've already spent two hours reading through compromised frames.
