warning

Essay or Multiple Choice

When I was in high school and the teacher announced a test for Friday, most students would ask what would be on the test. I, on the other hand, would ask if it would be multiple choice, short answer or essay. Knowing the format for a test determined how I would study (and even how I would learn). Over my lifetime, I have cultivated a deep-seated suspicion of tests, pretty much because I was good at taking them. I've taken a series of IQ tests of the years and gotten scores ranging from a low of 98 and a high of 174 (these numbers purport to measure something called "intelligence."). I once took a programming course which I essentially failed even though my grade on the final exam was the highest in the class (and even I agreed that I deserved a failing mark; I know all the concepts and could answer artificial exercises, but I couldn't get programs to actually work). I once almost failed a literature course (my best subject) because I was unable to recall very specific details from novels (people's names, hair color, etc) even though I was able to write great essays about them. On standardized tests like the GRE, I would study hard one time and get modest results, and on other times not at all and get remarkable results. I worked at a career testing and aptitude testing foundation for four years and began to appreciate how the manner in which a test was presented often affected the outcome, how an oral exam of a subject bought often vastly different results from a written exam.

A number of educational readings suggest what implications a particular learning theory might have on an assessment method. The thinking seems to be that a single competency-based test is better, or if not that, then one method is to be preferred over the other, whether it be "mental maps" or "expert tutoring." Even when the articles lambaste "accountability testing" or the old-fashioned multiple choice test, they overlook the need for a variety of testing instruments to be used simultaneously.

Using a particular test method poses two problems. First, some people's natural abilities tend to put them at a disadvantage on a certain method. Some students are good at memorizing; others are better at using charts and presentations; others are good at interpreting texts. Second, and more importantly, students quickly develop metacognition strategies for "beating the test" which affect not only how they study but how they learn in general. For a class requiring only papers, you didn't need to know names or details or even terminology but simply to know the different theories or schools of thought and present them as arguments. For that literature class where I failed the weekly comprehension tests, I learned to read the original text very carefully without doing outside reading or understanding the context or cultural background. The first time a test is delivered, everyone is at a disadvantage, and person whose learning style was most suitable for the particular testing method would tend to excel. But after the first time, what matters is whether students changed their learning style to be able to perform the tasks called for on the test. In other words, the choice of method dictates what kind of learning actually occurs. Over time, students' ability to pass the test will increase, but the results will not necessarily indicate general mastery, and indeed the test will still be biased against certain students.

The mental maps method described in one of the articles was fascinating, and I suppose that the research behind it might demonstrate its efficacy in the classroom. But mental maps seems to work well only in some domains of learning, especially those hierarchically arranged, or dealing with concrete objects rather than ideas, or dealing with interactive systems. I found myself trying to compose maps for certain subjects: literature, the Holocaust, this class, ESL, linguistics, computer programming. In a field like biology or chemistry you are dealing with taxonomies, so of course these maps would be better. A person without a good talent for diagrams might prefer mnemonic devices or understand things in terms of an example, for which mental maps mean nothing. In the case of the hard sciences, that deficiency may be unfortunate because the structural arrangements of these hierarchies is so critical to understanding the subject itself.

But even in the sciences, it is not always possible to diagram thoughts into clear subject and predicate charts. Perhaps a simpler unidirectional flow charts might capture a student's sense of cause and effect. But of course once students learn that making more nodes is better, they will start making trivial nodes and connections simply for the sake of quantity. At this point, the test method will be rendered irrelevant.

When teaching TEFL, I noticed that some student did better on writing paragraph essays while others did well on memory-oriented tasks (knowing vocabulary, recalling things from a passage). The easy answer is to try to divide the subjects into several subjects and test on each skill individually. The problem is that 1)a teacher is expected to give a single grade, not many and 2)the individual masteries tend to be interrelated, and often one competence can carry off to another mastery. Take the one paragraph essay for example. I gave two "sub-grades": Grammar and substance. Easy enough, except first of all, students with the natural writing gift tended to write longer essays and make more mistakes. Or students recognizing their own grammatical incompetence would write simple subject-verb sentences. They tended to avoid situations where they could make mistakes. And what about spelling? Spelling is separate from grammar, isn't it? Doesn't that mean I need to divide grammar up into 2 categories?

Assessment should be an opportunity to display knowledge or competence. These opportunities should be diverse and allow for people to succeed in more than one way. Teachers should use the same teaching methods over time so students can develop a comfort level , but they should vary the methods often enough so students will be forced to prepare for several types of tests. One grade for a subject is not a good idea (and my preference would be not to give a grade at all). Maybe a teacher can give 5 criteria, and identify the area with the best performance and the area in which the student needs work). This can be done for everybody because yes, even the best students need to know where they need improvement. Also self-assessment should be a necessary part of any assessment. Students, more than anyone will know if they get it.

Further questions remain. Should a test include a variety of teaching methods (dictation, vocabulary review and multiple choice in the same test) or should each method be administered in a separate test? For convenience, it might be better to test everything at the same time, but that results in a composite score, and composite scores can be meaningless. As education seems to move toward contructivist approaches to learning, it becomes more necessary for a teacher to present opportunities to demonstrate learning or mastery, and to present as many opportunities as possible.