In Quest of the Perfect Exam

By Nicholas Adams

I’ve never added up the blue-book examinations I’ve read in 20-odd years of teaching. Figure 40 semesters at an average of 100 blue books per semester–that makes 4,000; but double that for midterms, and the number climbs to 8,000. When I started teaching, it was common enough to give two midterm examinations. With overload semesters and a couple of visiting gigs on top of my regular job, let’s say I’ve graded 10,000 exams, give or take. At the end of each term, it feels like the entire weight of all of them is on my shoulders. Other faculty members seem to feel that way, too.

Over the years, I have tried desperately to devise examinations that will be interesting for students to write and a pleasure for me to grade–not an easy job. My examination styles fall into broad stylistic periods, more or less the kind of divisions you would expect from a historian of architecture. There was what one might call the classical quotation era, when I did nothing but offer quotations from original sources for identification and discussion. There was a baroque complex-paragraph phase, when I wrote a paragraph of quasi-scholarly gobbledygook and asked for an elucidation. I even went through a cartoon phase, when I used contemporary drawings to stimulate discussion. There have been open- and closed-book metaphases when, either bored by the answers I was receiving or the questions I was writing, I’ve tried one or the other examination style. For a brief period, I even gave a Zen-like two-hour Renaissance-architecture exam that presented just one image and asked students to write a history of the period based on what they could see.

But over all, the numbingly common examination phrases shuffle in and out of my brain in a somewhat meaningless patter. Discuss the main elements . . . compare and contrast (pointing out the similarities and differences) . . . illustrate your answer with examples drawn from . . . take into account so-and-so’s arguments . . . write a paragraph explaining . . . cite specific examples to prove your point. Talk about measuring out your life in coffee spoons! No, I keep telling myself, there has to be a better way.

And yet, whenever I have canvassed colleagues about other testing techniques, I have come away unconvinced. In some circles, the "dinner party" is popular. But while having students write hypothetical conversation among distinguished guests may work in some classes, the idea of Filippo Brunelleschi and Frank Gehry huddled together over linguine seems faintly ridiculous to me ("Wow, Pippo, that herringbone brick pattern in the cathedral dome means we were both inspired by fish!"). And I have attended glacial dinner parties in which conversation was as anguished as any that might occur between notable grumpies like Mies van der Rohe and Rem Koolhaas. Who would want to go through that again?

Some colleagues hold out hope for World Wide Web-based exams, in which students can pluck digital images from cyberspace like wildflowers from a forest and paste them into their virtual essay albums. Maybe that will be more fun for them to do (and for me to read), but I am not there yet. And even on the Web, one still has to ask a question, though at least one won’t have to puzzle out the handwriting.

This year I had another inspiration.For the past five years, I have been studying the historian Lucy Maynard Salmon (1853—1927), in preparation for editing a collection of her essays with Bonnie G. Smith, a historian at Rutgers University. Lucy Salmon was one of the most innovative historians of the 20th century, developing novel forms of historical analysis and a strikingly original modernist writing style at the same time that James Harvey Robinson was developing what he called "the new history."

When she came to Vassar’s history department in 1887, Salmon carried her boldness into the classroom. She favored methodology over narrative. She had her students visit and study the library, where they learned about the monastic sources of its gothic architecture. They had to identify the press marks figured in the stained glass and examine the nature of the card catalog as a device for opening the world of knowledge. In the library, they picked a book to study and evaluated its authoritativeness based on the sources listed in its bibliography, footnotes, and illustrations, as well as by its presentation: typography, quality of paper, binding, and so on. Salmon also had students study newspapers as historical documents–newspaper reading was still largely confined to men–to discover what one could deduce about the period from the articles, editorials, and cartoon.

One of her most radical exercises was to bring to class laundry lists–those multiform documents now limited mostly to hotel rooms, but then common to the new, commercial laundries–and ask students to interpret them as historical documents. What might the lists reveal about people dropping off their laundry? About those who ran such establishments? About systems of domestic service? About business organization? Her examination questions were equally memorable, and this [academic] year I decided to try one of them out in my class.

Salmon did not ask her students for the traditional summaries of events or even for exercises in comparison, but tried to turn students’ attention to the methodological issues underpinning history. In January 1914, for example, she asked students to evaluate the textbook they were using: "1. Give in bibliographical form the name of the textbook used. 2. State what principles can be applied to test the authoritativeness of any textbook. 3. Apply these tests to the textbook used. 4. Why is the study of the textbook the first step in the study of any period of history? 5. What classes of material are needed to supplement the textbook in the study of a. frontier life, b. the Stamp Act, c. the Philadelphia Convention?"

In another question, she asked students to discuss the Vassar College catalog as a historical document (June 1918); in another, to locate what historical records could be found in the students’ own back yards, or in the houses in which they lived (May 1911). Her course on municipal government culminated with a question that asked students to identify the chief defects in the plan for Poughkeepsie and how they could be remedied (May 1910). And in another examination dealing with the Vassar College campus (January 1915), she asked students to apply to the replanning of the campus the principles deduced from a study of city planning.

My favorite, from a course on American history, dates from February 1918 and was the one I used last semester in slightly modified form, in a course called "Architecture after Modernism." (Postmodernism, incidentally, is a category that would have been enormously interesting to Ms. Salmon, who wrote about the question of why history needs to be rewritten.)

Here is her exam question: "1. Make out a set of ten questions that will illustrate the work of the semester. The questions should have a logical sequence and express one central idea [in my version, I eliminated that requirement]; they may be framed as single, or as group questions; it is not necessary to consider the length of time or the preparation required to answer them; the textbook may be consulted in preparing the questions. 2. Why is this an examination? 3. What proportion of the questions could you answer?"

For my students, I explained on the examination sheet the source of the questions and a little bit about Lucy Salmon. After the students read the examination paper there was a collective nervous laugh, and one student, eyes raised to the ceiling, said, "Thank you, Ms. Salmon." Then they got to work.

The results were a pleasant surprise. The answers to the first question were not only interesting to read but relatively easy to mark. Bad or ill-prepared questions are much easier to spot than bad answers, it seems–they are shorter, and misunderstandings jump to the eye. Mundane questions (in the middling range) repeat the questions I asked in class. But students who had thought through the material, adding their own insights, asked incisive questions–often philosophical or speculative–that were rooted in their understanding of what we had studied and that integrated their learning from other courses.

The answer to the second question–Why is this an examination?–produced no surprises. But the third question, asking what proportion of their own questions they could answer, evoked interesting results. There was a certain bravura among some respondents, who thought they would have no problem answering their own questions. But those who asked the best questions often knew that they would be hard to answer and said so. And though there was one know-it-all, a history major, who told me afterward, "I knew you were going to ask us something like that," there were also comments of the sort that keep one in the classroom. One student, who had been highly critical of my dull midterm examination, wrote: "While you might not have written evidence of my complete thought process, this has been an exceptional experience of self-evaluation, free of regurgitation and useless binge-and-purge methods of studying and test-taking. In the end, this is one of the most effective exams I’ve ever taken."

Blue-book heaven? Well, this is probably as good as it gets. Thank you very much, Ms. Salmon.

Nicholas Adams is a professor of architectural history at Vassar and a coeditor of History and the Texture of Modern Life: Selected Essays of Lucy Maynard Salmon (1853—1927), published by the University of Pennsylvania Press in March. This essay was first published in The Chronicle of Higher Education.