Reading Between the Lines of Taos Death Notices
I've spent years compiling and cataloging obituaries from the Taos region. The Rio Arriba County courthouses, the old newspaper archives at the Taos Library, and various online memorial sites form a scattered record that most people don't realize exists. What I'm about to lay out is how to actually pull meaningful data from these sources, not some academic exercise, but the messy practical work of tracking vitality and decline in this corner of northern New Mexico. Start by gathering your sources. The main ones are the Taos Valley News, the Rio Arriba County Journal, and the Albuquerque Journal's regional obituary section which covers Taos deaths frequently. There are also funeral home websites like Baca Mortuary and Garcia's Sunflower Chapel that run memorials. Beyond the formal publications, you'll find Facebook memorial pages and the Taos Obituaries group on social media where families post their own notices. Nothing is centralized, so expect to spend significant time cross-referencing. The core challenge with any obituary dataset is inconsistency. A death notice in 1995 looks completely different from one published today. Early notices were often one paragraph with minimal detail. Modern ones run two to three pages with biographical essays, photo galleries, and lists of survivors that can include extended family across multiple states. This shift alone skews any trend analysis unless you account for it.
Here's what I found after tracking roughly fifteen years of data, and I mean actual hours logged reading every notice I could locate. The age distribution has been slowly shifting upward. In the early 2000s, a meaningful portion of obituaries were for people in their sixties and seventies. Now the median age at death in the record sits closer to mid-seventies, with a noticeable bump in deaths among people in their eighties and beyond. This tracks with broader New Mexico demographic patterns but the effect is more pronounced in Taos due to the retiree influx over the past two decades. The cause-of-death data is fragmented because obituaries rarely state causes explicitly anymore. That used to be common in the fifties and sixties. Now you'll see phrases like "passed away peacefully" or "after a brief illness." When cause of death does appear, it's usually something family-oriented like cancer or heart disease. You can infer partial data by looking at age ranges and using public health statistics from the New Mexico Department of Health as a proxy, but the correlation isn't tight enough to be comfortable relying on. I hit a specific wall when trying to compare pre- and post-2000 data because of how the format changed so dramatically around that time. Around 2003, the Taos newspapers began running longer, more sentimental obituaries funded by the families rather than charged by the line. Before that, length correlated with cost. After that, length correlated with how much the family wanted to say. This created an artificial spike in notice density that looked like increased mortality when it was just increased spending on print. I had to manually exclude anything published between 2003 and 2008 from my density calculations and treat that period separately until I could establish a new baseline.
The geographic data within obituaries is where things get interesting for understanding decline. I track where deceased residents lived at time of death, but also where they were from originally and where their surviving family resides. A pattern emerged showing that people who died in Taos but had families in Colorado or California were increasingly common starting around 2010. This suggests outmigration of the next generation, which is a signal of long-term community decline that obituaries capture better than any census estimate. Another thing most people miss: the seasonal variation in obituary publication carries real information. Deaths in remote parts of Rio Arriba County sometimes go unreported for weeks during winter. I've noticed a consistent pattern where February and March obituaries cluster noticeably, suggesting that deaths occurring in January in isolated mountain communities don't make it to print until roads clear and families can gather. This creates an artificial trough in winter death data that shouldn't be interpreted as fewer deaths actually occurring. If you're building any kind of analysis, you need a consistent coding framework. I use a simple spreadsheet with fields for date of death, age, location of residence, occupation if stated, veteran status, religious affiliation if mentioned, and surviving family geography. It takes about twenty minutes per obituary to code properly. Expect a full dataset for five years to require roughly eight hundred entries and sixty to eighty hours of work if you're reading each one carefully.
Get the Full Details
The biggest limitation anyone working with this data will run into is the incomplete record. Not everyone who dies in Taos County gets an obituary. People who are estranged from family, those without means to pay for newspaper notices, and residents of institutions like the Taos Community Living Center sometimes go unreported in formal publications. The social media memorial posts partially fill this gap but introduce their own selection bias toward people with younger, more connected families. I've found no reliable way to measure this underreporting, which means any trend you identify probably understates the true scope of whatever you're tracking. For someone starting this kind of project, my recommendation is to narrow the scope significantly. Don't try to analyze all of Taos County across all time. Pick a specific decade, focus on a particular demographic like veterans or long-term residents, or examine a single cause of death category. The data is there, but it's messy and uneven, and the narrower your question, the more useful your findings will be. There's no single download link or pre-built dataset for this because it doesn't exist in any organized form. The work of assembling it is the project itself. What I can offer is the framework I use, which you're free to adopt or adapt. Start with the newspaper archives at the Taos Public Library, pull microfilm or digital copies of the Taos Valley News going back as far as you need, and begin coding. The patterns emerge slowly but they're real, and they tell you more about who lives here and who leaves than anything else available.
