Friday, I spoke with Carlton Prothro and he had a question about the population of Minden in 1920 for an upcoming presentation he is making. While confirming those numbers for him, I was reminded of a particular quirk in that year’s enumeration for Minden, so today, I am rerunning a column from a few years back discussing some of the problems associated with using Census records for research and that unique Minden 1920 Census report.
In the course of doing genealogical research, hardly a day passes by without finding occasion to use the population schedules of the United States Census. The census is one of the best tools for not only tying parents to children, from the listings of family groups, but also of tracing families as they engaged in the westward migration that built our nation. From 1850 forward, the census schedules list each family member living in a household and depending on the year, includes various other data such as education, property and personal information. Thus, by starting with a name of ancestor you can go through the censuses of his or her lifetime and track their life. Today various subscription-based services on the Internet make a quick search of all the Federal Censuses just a mouse click away with the actual schedules displayed on your screen available for printing. It would seem a perfect answer to anyone trying to trace his or her family’s history in the United States. However, things are not always as they seem. Today’s column will deal with some of the quirks, foibles and outright errors in Census records and one major mix-up in Minden’s history.
The first problem with tracing a family member through the census is the time frame in which they lived. If you are looking for a family member prior to the 1850 census, you will often be frustrated. In 1840 and previous years, only the name of the head of the household is included on the sheet, other family members are merely numbers in the correct age column. The absence of names prior to 1840 stymies many researchers.
But, even having the names on the sheet does not always solve the problem. Another big problem with census records is the reliability of the information. Imagine, if you will, being the lady of the house on a sunny day in June 1850. Your husband and your older children have been working in the fields since early morning. You are struggling to finish your household chores and prepare the noon meal, when suddenly a man rides up in your yard. You recognize him as a well-known man in town, a man who probably does not work on a farm, but has a clerical job. Perhaps he is an attorney. Anyway he proceeds to come up to you with a large book and starts asking lots of questions about who lives in your house, what are their ages, when were they born, where your parents were born, where your husband’s parents were born, and other “foolish” questions. Most likely your answers are not going to be particularly accurate and perhaps you even delegate the responsibility of answering that pest’s question to the oldest child at the house. Because of these types of situation and simple miscommunication it is very rare for all the personal details about one person to remain the same in two consecutive censuses. Back then it was unimportant, today, when you are trying to get the facts about a family member it can be very frustrating. The only Census I will truly rely on for my family history is the 1910 Census for Columbia County Arkansas as the Census taker was my great-grandfather David Green Emerson, who was a bookkeeper and really concerned with accuracy. I think he got my ancestors right.
Spelling of names is subject to the whim of the enumerator. In my family I searched for my great-grandfather for many years. I knew that in 1880 he lived in Lafayette County, Arkansas. His name was Greenberry Agan and his wife’s name was Mahala. Unfortunately, there was no one by the name Agan in that section of Lafayette County that I could even find. Then the Church of Jesus Christ of the Latter Day Saints produced a new CD-ROM version of that Census. It allowed various forms of “intelligent” searches using phonetic and other types of clues. The first time I was able to use that tool looking for my grandfather it solved the mystery. There he was “Berry Agars” with his wife, “Hallie Agars.” Apparently, the census taker knew the couple well enough to call them by their familiar names, but not well enough to spell the last name correctly. Mystery solved. The introduction of these search tools has made it easier to solve some of the mysteries of the Census, but there are still other factors that make Census research difficult. Now, most of those same sophisticated search tools are available for subscription on the Internet and with no charge when using databases inside public libraries.
Of course, you cannot find your relative if they were not counted. The worst year for that problem in our area is 1870. This was the first census after the end of the Civil War and the beginning of Reconstruction. Many factors combined to make the totals from that year probably the worst undercount in our history. The census takers, for the most part, fell into two groups. Either white Republicans, many of whom were from the north and did a poor job of finding rural home sites; or newly freed slaves, who either feared approaching white homes or were hindered because they had never been allowed to attend school, rendering their reading and writing skills inferior. Another factor was fear sparked by the violence of the times. All people, black and white, had to be hesitant when a stranger appeared at their house. Random acts of violence, while not common, were perpetrated on rural residents in those years and many times the family simply didn’t answer the door when the census taker appeared. In our area, the 1880 census has some gaps, because of the creation of new parishes. Webster Parish was created in 1871, so it appeared in the census for the first time in 1880. In those years there were no road signs marking the end of one parish and the beginning of another. Many times, those living in the fringes of a parish were either missed by both, or on occasion, counted by both. The same situation was true in Red River Parish, also created in 1871, where another group of my ancestors lived. The Longino family (by the way that name appears in various census records as Longineau, Longineaux, and Lungeno) lived on the line between Natchitoches and Red River Parishes. After years of exhausting search, even using the modern tools and reading the entire censuses of both parishes page by page, I can confirm that according to the official count, my great-grandfather and his family did not exist in 1880.
In all my examinations I have never found a case like the census flaw that affected Minden in the Census of 1920. I first became aware of an anomaly with that year while researching in old local newspapers. In May 1930, the population totals for the 1930 census were announced. In Minden, these numbers produced a shock. Minden had boasted a population of 6,105 in 1920, but by 1930 the official total only showed 5,622 residents. The Minden Signal-Tribune in a front-page editorial offered the following explanation:
“Figures don’t lie. That’s an age-old proverb, but it seems almost disproved by the late census returns which give Minden a population of 5,622 as compared with 6,105 in 1920; but causes us nevertheless to hail a gain.
“The explanation isn’t complicated. In 1920, Minden was in the midst of one of the biggest oil booms in the history of the parish. People literally poured into the town from every possible direction. Houses were overcrowded; room and board went at a premium. Even the vacant lots and alleys were utilized by the transient crowd and tents sprang up like mushrooms in every available space. The boom was accompanied by a hike in prices of all commodities, naturally, but that is another story.
“1920 was a census year, and the census takers list people where they find them if they have no permanent home. The transient crowd that occupied Minden then had no permanent home, nor did they intend to make Minden their home unless they struck it rich. Few did; the boom dragged to a close, and the people began their trek for easy money again. Pipe liners completed the work they had to do in this vicinity and the workers moved on.
“But Minden’s population remained 6,105, an increase of over one hundred percent since 1910. Let us estimate the transient problem in Minden in 1920 was 1,500, a conservative figure, bringing the real population down to 4,605. Minden has been growing; there is no denying that. But the city limits have not been extended as the population advanced, and much of the city’s real population was not included in the 1930 census. Our suburbs, which are an integral part of the city in everything except name, have grown and are adding to Minden’s progress, but the 1930 census will not show this. These figures will appear only in the parish and ward reports.
“It is unfortunate that the transient population of 1920 should be officially counted as belonging to Minden and that the suburbs, which have not yet been included in the city limits, should be excluded but the condition is one which was almost entirely unavoidable and for which no person can be blamed.”
Apparently, this explanation sufficed for the time. The same argument was raised in a news story in the Signal-Tribune and in those days, before census totals were tied to government funding, it was really not a major issue. The strange census totals were soon forgotten.
From a historical perspective, however, the odd count remained a problem. The extent of the oil boom of 1920 in Minden was grossly exaggerated in the editorial explanation. While the scenes described were common in Haynesville and Cotton Valley; Minden did not have a boom on the same scale. In addition, the editorial left out a couple of pertinent points relating to local industry. In 1920, Minden had been without a major employer for two years, since fire destroyed the Minden Lumber Mill in May 1918. Three years later, in 1923, the shops of the L& A Railroad were relocated to Minden bringing hundreds of families (including my Agan family) to Minden. The boom was such that new school construction was mandated and new residential areas, along the Sibley and Shreveport Roads were constructed. How was it possible that during its greatest boom time in history, the 1920s, Minden lost population?
Again, the answer came from the use of new census tools and the simple availability of information. Census population schedules are sealed by law for 72 years. In 1930, the only data available to the public were the statistics, not the personal details, listing by name all residents that are found on the schedules. The 1920 census schedules were first made available to the public in 1992. A few years later, as a by-product of more technical innovation, these schedules became available on the Internet. One day, while researching some local families I was using the 1920 census for Minden, when I was struck by a sense of “déjà vu.” Was I imagining things, or hadn’t I just seen these same names earlier. On page 8B, I found the family of R. H. Miller, prominent local businessman and grandfather of Thad Andress. He was listed as being 51 years old and having been born in Arkansas, both of which were incorrect pieces of information. In the household with him were his wife, Treeby; children, Joan, Weston, Christopher, Robert, and Treeby; and sister-in-laws, Jeannette and Elizabeth Chaffe. Yet, on page 26A, I found another family, headed by R. H. Miller and including his wife Treeby C.; children, Joan C., James Weston, Christopher C., Robert H., Jr., and Treeby; sister-in-laws Jeanette and Elizabeth Chaffe; and three boarders. On this page, Mr. Miller was correctly identified as a native of Alabama and with the correct age of 55. Looking more closely at the two pages I found that the Miller family was not the only family counted twice. Neighbors such as the Almond, Fitzgerald and Nelson families appeared on both pages. Examining more closely I discovered that large sections of Broadway and the area around Minden High School had been counted twice. In fact, part of the area of West Union was triple counted by the census taker, an older lady.
I have not yet had the time or energy to fully evaluate the distortion caused by the excess counts. It appears that some sections of the Warsaw and Shiney (which the lady oddly called China) neighborhoods may have also been counted twice. I assume the cause of this was simple human error as further examination shows a few families were actually counted three times. I have omitted the name of the lady who took that census, because the significant part of the story is the miscount, not the person making the error. In those day funding from government was not tied to the census count so that leaves no motivation for fraud. Although in some areas the census taker was paid on the basis of people counted, but I don’t know if that policy was used here.
So, when you start your genealogical research you will find the census to be a very good tool, but also keep in mind the computer anagram, GIGO, garbage in – garbage out. The information you will find is only as good as the person gathering the information. At least one Echo of our Past is a puzzle of population caused by simple human error in the census.
Webster Parish Historian John Agan’s column appears Tuesdays in the Minden Press-Herald