One of the more difficult tasks of the history teacher is to have students gain some sense of how we know what we know about the past. This is important for the undergraduate history major, particularly the student who intends to combine history with a secondary education preparation in anticipation of a career in teaching. Different strategies exist for introducing students to the study and interpretation of primary sources, and many such strategies have been discussed in the columns of Perspectives. We have found that one effective way is to have history majors work on quantitative projects in social history involving computer analysis of census data. The experience of studying history is enhanced when students can manipulate a computer to generate results and contribute to our knowledge about what happened in the past, when it did so, and why.

Our project began with an attempt to combine our interest in the teaching of history with the University of Wisconsin System's "Design for Diversity," the statewide effort to feature American minorities in every facet of university life. We wished to have our undergraduate students engage the new historical literature on race and race relations in American history, and apply the knowledge learned from reading modern secondary accounts to an investigation of Wisconsin Indian and Wisconsin African-American history. We felt that census data was the most appropriate primary source for our students to encounter, and that quantitative study of the census was the most efficient technique. We submitted to the university's system-wide Undergraduate Teaching Improvement Council a proposal that envisioned senior-level students at our two campuses undertaking an effort to study and write about Wisconsin Indians and Wisconsin African-Americans.

We met several times during the summer of 1989 to plan the work. In addition, we communicated with each other via BITNET on a regular basis. We identified two main sets of primary sources that we wished to make available to our students: (1) the decennial aggregate population, agriculture, industry, and housing census conducted from 1850 through 1970; (2) the annual tribal population census for each Wisconsin Indian tribe from 1885 through 1940. We gained access to the first source—the U.S. census—through the data archive of the Inter-University Consortium for Political and Social Research (ICPSR). The data reached us in machine-readable form from ICPSR on computer tape, along with printed codebooks that explained the variables in the data sets. The ICPSR census data, however, did not come with computer programs already written to interpret the massive accumulation of data contained on the tape. The nationwide 1940 census, for example, extends over 240,000 lines of computer data. Consequently, much of the summer 1989 work involved writing SPSS programs on our mainframe computer to gain access to the federal censuses for Wisconsin. We also prepared ten other smaller ICPSR data sets relevant to Wisconsin, such as social surveys from the 1960s and 70s that included Milwaukee. The final Wisconsin data sets were finished by September 1, 1989. These, when selected out from the larger national data sets, are sufficiently small in size that they can be manipulated on a microcomputer using the SPSS-PC package. Our goal was to help our students manipulate this vast source to learn about the black people of Wisconsin and, for after 1890, the Indian people of the state.

The second set of important primary sources—the annual tribal censuses—are nearly as voluminous as the federal decennial census. The tribal censuses were conducted by the Bureau of Indian Affairs and exist in manuscript form, available to researchers via National Archives microfilm. We worked during the summer of 1989 to identify the microfilm reels that cover all Wisconsin Indian people and to develop a model computer program for students selecting a small sample of the tribal censuses. Wisconsin has six tribes residing on eleven reservations, ranging from the indigenous Winnebago and Menominee people to such relocated eastern tribes as the Oneidas and Stockbridge-Munsees. Our goal was to help students see the tribal census as a source for investigating the historical changes in Indian families during the critical decades of federal assimilation policy and also how the tribes differed from one another in their response to change.

We also purchased reprints of all the presented papers and commentaries from the D'Arcy McNickle Center for the History of the American Indian, Newberry Library research conference, "Towards a Quantitative Approach to American Indian History" (Chicago 1987). These papers, available as an occasional paper from the Center, outline replicable models our students may follow as they do original studies of Wisconsin Indian data. McNickle Center staff are currently conducting a family history project which examines changes in household and family life from 1880 to 1930. For this project they have prepared a database for five Indian groups (White Earth Chippewa, Crow, Creek, Hopi, and Colville) which draws on tribal census and manuscript federal census returns. This database will be available to the public at cost in approximately thirty months when the project is done. According to Research Assistant Nancy Shoemaker, these data sets will allow students to compare Wisconsin Indians to these groups.

Because of our respective teaching schedules and other responsibilities, much of the initial attempt to implement the project in the classroom took place at the Eau Claire campus in two sections of the year-long senior thesis course known on campus as "Historiography." One section of sixteen students worked on a book-length collaborative study of aspects of the history of Wisconsin African-Americans, and the other section with fourteen students tackled a similar project writing part of the history of Wisconsin Indians. The fall semester of each section was devoted to learning quantitative techniques such as frequency distributions, crosstabulations, and regression analysis. In addition, students learned how to master the primary sources gathered for the projects. Throughout the fall semester and into the spring, the Eau Claire students had an opportunity to consult about their thesis topics with some of the leading scholars in African-American and American Indian history from such places as Carnegie Mellon University, Clark University, Columbia University, Northeastern University, the Smithsonian Institution, the University of Pennsylvania, and the University of Wisconsin-Madison. During the spring semester, the Eau Claire students worked on writing their individual theses, a paper of at least twenty-five pages, not counting notes, bibliographies, tables, maps, and graphs. In turn, these were combined in the summer of 1990 to make up two volumes published by the University of Wisconsin-Eau Claire History Department, "Studies in the History of Wisconsin African-Americans" and "Studies in the History of Wisconsin Native Americans."*

We did learn some lessons in the execution of this project. The summer time we devoted to the preparation of census data sets and programs was well worth the advance effort. It is important to have data files and primary sources ready for students to work on as soon as possible after classes begin, even in a two-semester course. There is an inevitable trade-off between devoting time to teaching history students statistics and computer use and spending time on how a historian analyzes a primary source. Where possible, we learned it is best to combine the two—teach statistical methods and computer analysis on the very data sets that students will be using in the writing of history. History majors on our two campuses often choose the major because of an aversion to math and science; overcoming such fears takes extra time and effort on the part of the instructor. Many of the initial problems the students faced involved file management on the computer rather than the analysis of data; we found it a great advantage during office consultations to have a computer on hand connected to the university mainframe to help our students. Our students worked on the university mainframe, not for the ease of operations, of course, but more for the availability of terminals. Our type of project is feasible on a microcomputer powerful enough to run SPSS-PC; history departments equipped with a sufficient number of such machines can take advantage of the greater flexibility of the personal computer.

The students' self-confidence rose when they began to generate output on their data sets. Many thrilled to see their first computer printouts, complete with their names in large print in the front-page "banner." More than one history student was seen swaggering around the computer center with hundreds of pages of crosstabulations, lording it over, at least for the day, the usually haughty business students. They produced good history, overcame the worst of their math and computer anxiety, and won a little self-esteem in the student pecking order.

We found that the first stage of our project confirmed our notions about the ability of advanced undergraduate students to do excellent work. We believe that our students have already made an important contribution to the history of Wisconsin. We also believe that our students have taken justifiable pride in the work they have accomplished.

Martin Zanger is a professor of history at the University of Wisconsin, La Crosse. James Oberly is an associate professor of history at the University of Wisconsin, Eau Claire.

* Copies of these publications are available at cost —about $20/volume—on an ad hoc basis by writing to Professor James W. Oberly, Department of History, University of Wisconsin-Eau Claire, Eau Claire, Wisconsin 54702-4004.