My relationship to the U.S. history survey is quite unique, at least in comparison to my colleagues. Most of my colleagues began their careers teaching the survey as grad student T.A.s, grading tests, leading discussion sections, perhaps giving a lecture or two. As teachers, they model their survey courses after the professors they assisted, copying lectures, assignments, and even exam questions. Like their mentors, they talk about teaching the survey as an inconvenience. By contrast, I have never T.A.ed the U.S. history survey (the department chose our T.A. assignments in my school, and the Director of Grad Studies frequently assigned me to twentieth-century courses). My first teaching job, as an adjunct at a community college after finishing my Master’s coursework (at an M.A. program that didn’t require any T.A. work), I taught the second half of the U.S. history survey. By myself. Given that I hadn’t studied the twentieth century since my undergrad days, the experience was, to say the least, a baptism by fire. I quickly understood the challenges my senior colleagues often complained about: lack of student interest, the hurried pace that forces you to skim over so much material without adequately exploring any of it, class sizes that made anything but multiple choice tests prohibitively time-consuming. My next teaching gig, at Washington University in St. Louis as I finished my dissertation, included teaching two sections of the survey, which at Wash. U. is the whole thing—pre-contact to the modern day—in one semester. I expected the experience to be even more grueling. But by the end of the semester, the course was my favorite class to teach. I began to wonder: was something wrong with me?

I still don’t have a definitive answer to that question, but what I do know is, I like teaching the American history survey, and I’m not bashful about saying so to my friends and colleagues. It forces me to revisit books outside my research field that I may not have read since I prepared for comps, and to consult more recent scholarship on subjects I find fascinating but which are outside the chronological scope of my field. It requires that I make broad connections between themes and eras responsibly, and push my students to make their own connections and then defend them using evidence. It also gave me a chance to try out an approach to assigning readings in surveys that is a little unusual, at least when compared to many of my colleagues’ pedagogical styles.

Most of the colleagues I’ve talked to or observed assign a combination of primary and secondary source texts in survey courses. The thinking is, I suppose, that doing so is an effective way to introduce students to both history and historiography in the same class. When I taught at the junior college, I did this, too; I assigned a textbook and supplemented it with both primary and secondary source readings. (I tried to be careful not to overload the students with too much reading, though I’m not sure how well I succeeded). I found that students often parroted from the secondary sources, rather than form their own opinions. Discussions of historiographic nuances went over their heads, and failed to generate any interest. And the textbook reinforced the belief that history was an endless parade of names, dates, and facts to be memorized. In short, the secondary sources did more harm than good. When I got set to teach the survey at Washington University, I decided to change my approach.

I should say here that I benefitted from the freedom that Washington University gave me to teach the class. At the junior college, I was required to teach a textbook, but at Wash. U., no such requirement existed. I was free to assign, or not assign, whatever I wanted. So I decided to assign only primary sources. My lectures—which, since I had to cover all of American history, often had to cover whole eras in one fell swoop—provided a basic outline of the material, with only a few scant references to historians or seminal works of history throughout the semester. The primary source readings were designed to show how the broad themes I highlighted in my lecture played out “on the ground.”

To my delight, the format worked very well. Of course, some primary sources failed to elicit much enthusiasm, but many primary sources worked exceptionally well. Because students did not have secondary source readings to lean on when responding to questions from me or their fellow students, they had to make their own claims, and then point to specific passages and explain their analysis when pressed.

For many readers of this blog, this approach might be old news. But to me, it was a revelation. Everyone I knew assigned secondary sources in the survey. It was, pardon the pun, a matter of course. By taking them out, I found I wasn’t sacrificing anything that the richness of the primary sources did not more than make up for. The experience inspired what is currently my approach to assigning primary and secondary sources in all my classes. I omit secondary sources from the survey, and assign a roughly even amount of primary and secondary source reading in my more advanced classes. So far, I find this approach a significant improvement over what I perceive to be the more common approach of combining primary and secondary source texts in survey courses, and assigning mostly primary sources in their advanced classes. I have found that secondary source texts work well in my advanced courses (I have taught upper-level courses on colonial America, the ratification of the Constitution, and the presidency, among others), because students in those classes have taken at least the survey, and in most cases other history classes as well. They have experience analyzing primary sources, and can better draw upon the secondary sources they read without simply replicating another scholar’s argument. The papers they produce are more like scholarly papers in that they combine analysis of primary sources with an appreciation for how the student’s argument fits within an established conversation. By removing secondary sources from the survey and assigning them relatively equal to primary sources in advanced classes, I introduce my students to basic historical analysis first, and then build historiography on top of it. The result, I think, is better classes, and particularly a better survey course, for me and my students.