Computers and Software

Graphic History: From Colonial Demography to Civil War Soldiers

D'Ann Campbell, January 1993

Do people still go into history and the other humanities to avoid math? Many have speculated that, for some at least, the answer to that question is "yes." History journals have been known to cut tables to avoid offending the sensibilities of their regular readers—who sometimes boast they never read tables or study graphs. Oftentimes, assigned class readings omit articles that use these tools because they add one more type of analysis that has to be taught. Yet, at the same time many of us are avoiding these approaches, colleges are placing a new emphasis on basic mathematical skills. Fortunately, many of these same colleges are also placing an emphasis on computer literacy. It may well be that in this area of computer literacy, new possibilities exist for bridging the gap between those mathematically oriented disciplines, including the social sciences, and the humanities. As students learn many of their math skills on computers they are also finding those machines to be valuable tools for their classes in the humanities, and a new arena for interaction and exchange between the two approaches to learning has begun to emerge.

At the same time, large numbers of historians, unable to resist the wonders of word processing, now own their own powerful personal computers. Increasingly, they are also exploring other uses for their computers, from e-mail to on-line catalogs to manipulating research notes. By bringing these tools into the classroom, history can again serve its customary role as the bridge between the humanities and the social sciences by using all the new desktop computer power to illuminate historical problems. What follows is a description of a successful attempt to do just that in the honors U.S. history survey course at the U.S. Military Academy.

One could argue that I had a distinct advantage in trying my experiment at USMA, since the military academies are quite advanced in their level of computerization. All cadets and faculty have their own PCs, complete with standard word-processing and spreadsheet software, and all are linked together by a campus network. It is easy to send e-mail messages to all cadets in a class, rather like free telegrams (or free faxes, to bring the metaphor up to date). Plebes (freshmen) are pushed to master the systems they have. For example, a requirement that all English and history courses required papers to be word-processed in draft and final form provided strong motivation for all cadets to learn Microsoft Word quickly. In addition, over half the cadets will eventually major in math, science, or engineering fields.

Yet, on arrival at West Point they were not better prepared than the typical student at a good college. Their computer experience in high school was limited to programming (which was irrelevant for my purposes) and perhaps a little word processing. Services like e-mail were unfamiliar to them. Although I asked and answered questions and sometimes gave out assignments by e-mail, the cadets did not seem very comfortable with the system. Therefore, I could not assume my first-year cadets were familiar with the spreadsheet software I planned to use, Borland's Quattro Pro, and devoted two class periods to introducing its basics. (They were simultaneously taking a mathematics course that introduced Quattro as an algebraic tool, but none had the vaguest idea how the same program might possibly be used in a history course.) In this way, the challenge at the U.S. Military Academy was quite similar to that facing historians at other universities. A brief description of the assignments and how they succeeded might therefore provide insights on how similar activities could be used elsewhere.

Assignment #1: Population in the Colonies

Thanks to the major breakthroughs in the new social history, most textbooks now discuss the demographic history of the colonies and often provide appropriate tables, graphs, and maps. When I asked how many had studied the graph on page 132 in our textbook (Gienapp, Heyrman, Lytel, and Stoff, Nation of Nations) that compared black and white growth rates by region, none had given it more than a glance. I then challenged them to use their spreadsheet to uncover two major problems in the graph, one mathematical, one conceptual. None could eyeball the graph and find any mistake. In their computer lab they were given a data set on the population by race and colony, 1630 to 1780, taken from Historical Statistics of the United States. If we had more time, I would have asked them to enter the data themselves, but providing the data was a simple task. Together, we explored how to compute percentages and growth rates and how to construct and label bar and line graphs. They learned the necessity of using semi-log graphs to show the pattern of population growth. Their first homework assignment was to replicate the textbook's graph (a simple bar graph showing population by region from 1720–1760), find its flaws, and create a correct one. They were also to uncover some point about colonial demography from their data set and illustrate it with an informative and understandable graph. Quattro produces superb graphs very easily, and the program is so logically designed that student teams quickly figured out how to do the assignment.

The mathematical error in the textbook graph became immediately apparent when they built their graphs (the text subtracted two series instead of adding them). A few plebes figured out the conceptual problem as well. The designer of the textbook chart, perhaps more concerned with visual effect than accuracy, had overrepresented the black population and underrepresented the white. In fact, some offered further suggestions for improvement, such as organizing the data in a more natural north-to-south fashion.

Assignment #2: Virginia Regiments during the Civil War

We devoted approximately one-third of the course to the Civil War, paying special attention to both its military and social aspects. The topic fascinated the cadets, as it did my civilian students at Indiana University. To understand the historical issues, they read such social historians as William Rorabaugh and Maris Vinovskis. Then the cadets divided into teams of four, and each team chose a Virginia regiment to study for which a complete published roster was available in book form. These books provided capsule histories of the formation, organization, and battlefield experience of the regiments, as well as basic demographic and social information on each of their members. The class worked out a standard coding scheme that applied to all teams, a situation made possible by the fact that each roster had the same format. To prepare the data for this exercise, we used the database features of Quattro Pro. Although not as powerful as many database packages for linking files or preparing mailing labels, spreadsheets are, in many respects, superior for entering and correcting data. They also challenge statistical packages for certain data manipulations like changing dates into numbers and computing moving averages, and surpass most statistics programs in creating graphs.

Each team had the mission of coding two hundred names sampled from the rosters, a task that took about four hours, and then using the data to answer some basic questions about Lee's army—who served, how much did they fight, who became casualties, who deserted, and who survived. Not all teams sought the same answers. Some described the age distribution of enlistees; others examined occupational or geographical backgrounds. The more adventurous cadets began asking original social history questions such as what sorts of soldiers suffered the first casualties? Were they the old men or the young ones, farmers or townsmen? Likewise, they wondered what types were promoted first? They discovered, to their surprise, that the older family men did not disproportionately desert to return home. When they found that younger solders were wounded first and promoted last, they tried to figure out why. Were the younger men more reckless? Were they sent on more dangerous missions? The results were even better than I had anticipated. Not only had the groups worked together successfully, but they also helped each other greatly in mastering the historical and technical issues of the assignment.

Conclusions

What did students learn from these computer assignments? Their comments on the course critique (which arrived on e-mail) suggested they learned many things. First, it was essential to work in teams. Having someone watch over your shoulder as you enter data or work a program is no longer disconcerting when you realize how essential it is. Although they all had exactly the same prior experience with spreadsheets, some cadets caught on much quicker to the logic of the program. They discovered they had developed a better understanding of the spreadsheet program—a giant checkerboard, with each cell holding labels, numbers and formulas. The spreadsheet is electronically alive, in that each cell pays attention to every other cell and adjusts the numerical results of formulas instantly. Graphs can be built up from rows of data, one for each colony and each race, in which each column corresponds to a date. Summary rows are easily constructed that add together the entire white and black populations of the Chesapeake and compute the proportion black. Multiple graphs can be set up for each combination of rows, showing (in full color if wanted) the basic patterns inherent in the maze of numbers. Mistakes that could be buried deep in a table stand out glaringly on a graph and are easily corrected within the spreadsheet. If the textbook designer had used a spreadsheet, perhaps the flaws would have been caught before going to press. In a team of four, someone always figured out how to generate the graphs. Another mastered the coding scheme. Second, the students said they learned that textbook makers are fallible too, not just in expressing opinions or stating facts, but in the subtle ways that quantitative patterns are expressed graphically.

More importantly, students learned some of the secrets of how historians do research. Most had never examined a primary source in detail. By the end of the semester they could replicate the textbook graphs and knew how to replicate the studies done by social historians of the Civil War. The famous battles like Gettysburg were still there, but now they were seen from the perspective of privates and lieutenants, not just generals and presidents. They learned the valuable knowledge that interpretations are inherent in coding schemes and that coders are responsible for much more than data entry. They came to appreciate that the types of sources used will necessarily influence interpretation and presentation of the past. These students also struggled with one of the main problems of social historians—how to represent the history of millions of individual Americans. In addition, some students decided to take more history courses, because they have developed a new appreciation of the work of historians and what they could learn from it.

Finally, they also learned another lesson. Tools such as Quattro Pro, which they used in their mathematics courses, were also valuable tools in their history courses. Issues such as composition of categories and data transformations could be as critical in their history courses as in their math and social science courses. Perhaps when the students who took the course take their subsequent history classes, they will bring to them their experience with coding forms, spreadsheets, graphs, and their knowledge from other classes that will lead to further insights into how history happened.

—D'Ann Campbell is dean of the College of Arts and Sciences and professor of history at Austin Peay State University.