Placement Tests are an important part of the process at the vast majority of community colleges, especially relative to mathematics. Over on the
MATHEDCC discussion list, Fred asked an honest question about finding an online placement test that was not a commercial test. Most of the public responses to his query have been critiques of placement tests in general (some would say toilet-emptying, self-affirming statements). Under the surface, is it possible that we do not need placement tests?
Some readers will have an extreme reaction to a question raising the possibility that placement tests are not good. Let me clearly state my opinion after working with them for 39 years: Most placement tests are reasonably good assessments of the content that they were designed to measure. Given the limitations that users place on them (users being most of us), the tests achieve the best measurements possible. Of course, these statements don’t tell you if I think placement tests are ‘good’ or not … and that is my point. Our use of placement tests might be good or not; the tests themselves are just what they are designed to be.
The use of placement tests involves several issues. The largest issue right now is whether placement test results are the only factor in initial math placements. The best research I have seen suggests that we should supplement test results with other information, especially high school performance (overall) for recent ‘graduates’ (whether they graduated or not). Some states have a common data system for K-16 which makes this relatively easy; others (like my state) have significant barriers.
Another issue deals with the content definitions for placement tests. Some of us see the companies involved as ‘evil’, with a higher priority on money or prestige than on helping students. I suspect that this point of view is held by people who have not been involved with the companies work. Although it is true that some of the field representatives of the companies are not helpful academically … the people with actual control at the companies are focused on academic success. Personally, I fault ‘us’ more than the companies. We have been telling the companies that the content for the tests needs to identify skills that the student does or does not have; skills are forgotten, and are vulnerable to trivial details. If we would focus more on comprehension, application, and reasoning … the placement tests would have more meaning for us and our students.
A related issue is the use of placement tests in a deficiency model, such as some modular programs. We sometimes expect a placement test to indicate whether a student ‘knows percents’ (alternatively, ‘does not need the module on percents’). We should not use placement test for diagnostic purposes. We might use well-designed diagnostic tests for this purpose, though I actually have more concerns about diagnostic tests than placement tests. Diagnostic tests involve the effective ‘waiving’ of instruction; as a profession, I do not think we can support a 20-item diagnostic test as being equivalent to the instructional value of 3 weeks of class. I digress!
Another issue with our use of placement tests is ironic: We do not apply number sense to the results of a test. For measurements of objects, we know that there is no signficant difference between 3.1 meters and 3.2 meters — if the measurements are made with a meter measure. However, for placement tests having essentially 1 digit of precision, we often make a distinction between a 64 and a 61. Take a look at the standard errors for your placement tests, and remain humble. If tests are the only measure used, a ‘line’ needs to be drawn somewhere; this line might separate the 64 from the 61, but that does not mean that they are really different. Too often, we look at placement tests as if they were precision calipers when they are really meter measures.
The title of this post has two meanings (at least). The obvious would be ‘will placement tests go away?’; I do not think so, and I would not advocate that. Another meaning is ‘placement tests as ‘take-out’ process’ (like drive-in restaurants) … “would you like fries with that algebra test?”. In other words, ‘give the student what they want’; over on the MATHEDCC discussion, this was the point for some people — just let students decide on their best math course. Self-selection has been the subject of research, which I think has generally found less effective than placement based on tests for the entire population of students.
We do not need to get rid of placement tests. We need to support changes in the content of those tests, and we need to show a better understanding of the measurements resulting from placement tests. The ‘placement test problem’ is more about us as math faculty than it is about the tests.
Join Dev Math Revival on Facebook: