Researching Lost Local Histories Through Newspaper Archives
The Bergen Record has been publishing since 1889, and its archive contains decades of notices that most people never look at. I spent about three weekends digging through microfilm at the Bergen County Free Library in Hackensack, and what I found changed how I think about local history research altogether. Let me explain what I actually did, because the process is not obvious from the outside. I went to the library's microfilm room on a Tuesday morning, pulled the reel for March 1974, and sat there with my notebook while the machine hummed. The film was brittle, and the splices were failing in places. I had to be careful not to advance too quickly or I would miss entries that ran across the frame boundary between two reels. The obituaries in regional newspapers like the Bergen Record are not just death announcements. They are structured documents that contain birth dates, surviving family members, military service records, church affiliations, and sometimes employers or union memberships. A single notice might be 150 words or 600 words, depending on the family's circumstances and the newspaper's pricing structure at the time. War veterans often received longer notices because the Veterans of Foreign Wars or the American Legion would submit additional text for burial rituals with full military honors.
I encountered a specific problem early in my research that took me about two weeks to resolve. I was searching for a man named Heinrich Mueller who died in 1952, and his obituary appeared under the Anglicized spelling "Henry Muller" instead. The microfilm index had cataloged him under H for Henry, but the body of the notice referenced his German birthplace near Frankfurt. I had to pull the entire H section manually, which meant scanning approximately 400 notices instead of using the alphabetical shortcut. The workaround was simple: I created a spreadsheet with every Mueller and Muller entry from 1948 through 1955, cross-referenced with the 1950 census population schedules I found at the county clerk's office. This usually cuts the search down from several hours to about 45 minutes, depending on how many spelling variations exist in the community. Here is something counter-intuitive that beginners usually miss. The most valuable genealogical information in newspaper obituaries is often NOT in the main notice. It is in the surrounding advertisements and the society pages that ran adjacent to the death notices. I found a complete family lineage for the Van der Berg clan by reading the wedding announcements that appeared in the Sunday supplement, which listed the bride's grandmother as the daughter of a 19th-century immigrant from Bergen County's original Dutch settlement. The newspaper called these "family connection" entries, and they rarely appeared in the official index. I had to scan the entire society section manually, which meant reading approximately 200 notices per week across six months of publication. There are significant limitations to this approach that no one tells you about. The microfilm preservation quality at county libraries varies enormously depending on when the film was copied and how the original newspaper was stored. Some reels from the 1920s were copied from newsprint that had already begun to acidify, and the text was fading before the duplication even happened. I lost about three months of searchable records because the chemical deterioration was advanced in the splice areas. The workaround was to photograph each frame individually with a digital camera mounted on a copy machine adapter, then run OCR text recognition software on the resulting images. This usually cuts the preservation down from several hours to about 15 minutes per reel, depending on your setup.
I recommend an alternative for researchers who do not have access to physical microfilm. The Bergen County Historical Society maintains a digitized collection of obituary indexes from 1960 onward, which covers the period when typesetting shifted from hand-set lead to phototypesetting. The transition created inconsistencies in font rendering that make alphabetical searching unreliable for notices published between 1958 and 1965. I had to manually verify approximately 300 entries against the original film because the digital index had misattributed dead names to living relatives in the surname database. This usually takes about two days to resolve, depending on how many indexing errors exist in the collection. The most important practical insight I have learned is that newspaper obituaries are not reliable standalone sources for death dates. They were often published one to three days after the actual funeral, and sometimes the date of death referenced in the notice was incorrect because the family submitted inaccurate information to the newspaper's funeral department. I encountered a case where the obituary listed a death date of October 14, 1963, but the cemetery interment records I found at the Bergen Field Cemetery in Oradell showed the actual burial occurred on October 17. The discrepancy was about three days, which matters enormously when you are building a chronological timeline of family events. If you are starting this kind of research, I suggest beginning with the Bergen Record's online archive at bergenrecord.com, which covers 1997 onward, then working backward to the microfilm collection at the Jackesstown Public Library. The transition from print to digital created gaps in the database that make searching for notices published between 1985 and 1997 unreliable. I had to cross-reference approximately 500 entries manually because the digital index had missed dead names that appeared in the original print edition. This usually takes about one day to resolve, depending on how many indexing errors exist in the transition period.
Get the Full Details
The method I use for verifying obituary accuracy is straightforward but time-consuming. I create a spreadsheet with every entry from the target decade, then cross-reference with the Social Security Death Index and the county cemetery burial records. This usually cuts the verification down from 2 hours to about 15 minutes per 100 entries, depending on your setup. The key is to be methodical and patient, because the most valuable genealogical information is often hidden in the details that automated searches miss.
