Tuesday, January 29, 2013

Nightmares of my father

My parents grew up in a high stakes testing system*. The entirety of their success in college was based on two sets of exams. No home works, what lab work there was didn't seem to count towards a final grade, though reports did need to be submitted. Growing up, the need to memorize every possibly  relevant fact in the text book was an important part of doing my homework.

I was brought up in the US educational system. In high school there were chapter tests and quizzes almost every week. In college, there were essays and homework sets and lab reports that usually constituted at least 30% of my semester grade, and at least one midterm and final occurred every semester. Not quite what one would call high stakes testing, though exams still caused a good deal of stress for me. I cannot imagine surviving in the world my parents grew up in, where I would get little practice at, and less feedback on the skills I am supposed to be learning in the course.

My partner is teaching a two semester course at University E. Most of the grade is based on an end of the year exam. Grading homework involves a long an convoluted electronic process that seems to involve a step of standing over a Xerox machine, scanning graded home works into electronic form that can be uploaded to the online homework system. He hurt his back last week, so standing over a photocopier was being unpleasant enough to encourage him to consult with colleagues about how they deal with graded home works. The answer? They don't. They grade some small number of essays, which the system is better set up to deal with. But they don't assign graded home works.

Being educated in the US system, both my partner and I balk at this idea. How is a student supposed to master technical skills without practice? How do they know how their skills are progressing if they are only given solution sets, and no feedback on the work they have done? How does one do anything other than learn what is taught in lecture? How does one get any depth of understanding of the subject matter?

I know this lack of understanding about a high stakes system of education versus a low stakes system runs both ways. Many of my partner's colleagues who either did part of their education in the US, or taught there for some time, admit to having doubts that a greater stress on home works and less on exams could possibly lead to a successful education. Most of them have come to see that it does work. Many of my colleagues who have to teach in the US and were trained abroad complain, at least initially, about how US students are lazy and complain a lot, while students in their home country just sucked up and worked.

I have yet to see the benefits of a high stakes testing system. Any one who believes in that method of education, I would love to hear your point of view.

*Okay, give what else my father had to face in childhood, exams were hardly his biggest nightmare, though they did feature prominently in his view of our education.


  1. My undergraduate degree was assessed through 1 final year research project (1/8) plus seven 3 hr exams at the very end of the last year (7/8). I think that this was a bit extreme but there were some benefits:

    - it ensured that I had to have a good grasp of *all* the material at the end of the course. In modular systems with assessment as you go along, it can be easy to do well on one module and then promptly forget what you have learned as you move onto the next module.

    - Related to the above: by the end of the degree, I felt like I had a good, synergistic understanding of all the material and how it inter-related. My overall level of understanding was much higher by the end of the last year than at the end of the penultimate year, which makes me glad my penultimate year assessment did not count towards my final result.

    - I was assessed and got very personal feedback throughout the degree, as well as getting exam practice 3x per year, but none of these grades actually counted towards the final result except the last ones. The worked for me because I was conscientious enough to realise these assessments were for my own good and, on the whole, to generally put effort in. At the same time, I would not have liked the continual stress over 3+ years of having every little piece of work count towards the final grade.

    - in a way, it seems kind of unfair to have homework grades count, because isn't it through sometimes getting the homework *wrong* that you learn?

    - otoh, I realise that the 'everything on the Finals' system, whilst it suited me, would not work well for many (the majority?) of people. It was certainly stressful going into each of those last 7 exams, and I felt bad for one of my friends who came down with a cold and was fuzzy-headed for 3 of them.

    Finally, maybe the type of education I received is not really what you are talking about at all, since I was indeed assessed as I went along, even if it didn't count towards the final grade. Sorry for the long comment!

  2. Rather than being sorry, I am grateful for your lengthy thoughts. Here's where I disagree with your analysis of a modular system:
    1) The modular system that I refer to is a prerequisite system. This means that even if all of this years courses are not necessary for next year's, many of them are. This makes it hard to forget information learned in prerequisite classes, and helps with the bigger picture.
    2) The homework system is designed, in theory, to help one master material, write it to long term memory by repeated problem solving. This makes it hard to forget material from one term to another, especially if one has actually mastered the course in the first place. Though I admit, this is an ideal case scenario.
    3) In my experience, where home works are done collaboratively, and courses are graded on a curve, the homework grades tend to be high with small standard deviations. The midterm/exam grades tend to have larger standard deviations. If the course places 30% of the grade on homework, the curve means that the small homework standard deviation counts for 30% the grade, i.e. little. It's the constant pressure to do the work and get feedback on it that is important. If you got a lot of opportunity to practice the skills and build on (not just repeat) what you were expected to learn from lecture, and given a lot of feed back on your practice, then that may be the same as having a homework system, I don't know.

  3. Anonymous, Having thought about it some more, I think I see the benefit of an end of the year exam system for courses that don't have strong prerequisite trees, where your first point is stronger.

    I am also wondering how much my analysis is hurt by the fact that my only teaching experience is at colleges that only admit "smart" students. This modular system works because the students want to learn, by and large. In less elite schools, does the modular system give students a chance to not learn 16 times, while the system you describe forces them to learn it once, if only for the three weeks of the exam period.

    In either case, I think I've been conflating high stakes system with lack of feedback and opportunity to practice along the way. My problem is definitely with the latter, it may not be with the former.

  4. In the system where I studied for my first degree, int he first year you studied four distinct strands of the subject each assessed at the end of the year mostly by exams, three distinct strands in the second year, then focused on one strand in the third year. The degree depended only on the final year results, which were from two dissertation-type research projects (one third) and five papers covering the range of courses from the year (two thirds). Previous years year end exams counted only in that if there were limited places in a course for the following year, places were allocated to those with the best results in the relevant pre-cursur subject.

    There was however a LOT of feedback along the way, with at least one lab report, problem sheet, or essay due every week and small group seminars and problem classes to supplement the GIANT lectures.

    I think I would ahve done much WORSE in a low-stakes system (I currently teach in one, pretty much) - stressing weekly or every few weeks about assessment would be a lower grade of continual stress and anxiety. I am NOT good at coursework or repetitive activities once I've mastered the concepts, and struggled a LOT with things like school maths because once I could do the problems I made careless mistakes. Low-grade stress is not good - there's no space in the system for people to have even a week of problems, often a few weeks of say 'flu is enough to mean you need to repeat the whole year.

    What was GOOD about the high-stakes system with feedback along the way? The FREEDOM to explore the subject widely, the fact that if one week I wanted to follow up something really neat from one class and that left me less time for another assignment, it didn't matter.

    And yes, this system wouldn't suit the unmotivated slacker student, or one who doesn't love the subject for it's own sake or have self-discipline but needs constant external pressures to actually do any work. The university I went to was very selective so that wasn't much of an issue. My current place is less selective, and takes less well qualified students; and frankly either system is a problem with this population. If work doesn't count they won't do it, if every piece does count then the excuses and mitigating circumstances and extensions pile up to tsunami levels very very quickly, because they have complicated lives etc...

    In addition, many/most students DON'T integrate material well through the modular system, in my experience, even in the maths stuff where it is so clearly built up brick on brick. I don't know why but I think it's mostly to do with the population i serve, which just isn't very well prepared for higher education however it's delivered...

    My main concern isn't the testing system, it's the students who don't do the work, either because of all the other things going on in their lives

  5. JaneB has articulated some of my complaints against the low-stakes system. I feel that homework is sometimes needlessly "grungy" and repetitive and is really not much fun to do. When I didn't have to do homework, I actually spent my time learning and working on things that were interesting to me rather than wasting my time doing stuff a computer could've done much more easily.

    Thinking back, I can't recall ever being stressed about my exams in subjects that I actually liked learning. And I do remember stressing out over homeworks, mid-terms and other such trivia in subjects where I wasn't doing quite as well as I liked. This leads me to think stress is really caused by not understanding the material and not by the testing system being used.

    I hypothesize that the variation of learning outcomes across the testing systems we're discussing is much less than the variation of learn outcomes across student ability or professor/learning environment quality. This isn't to say that all testing systems are just as good, but that they're probably the most important factor.

  6. I was educated in high stakes where final exams count for everything and homework or assignments were reviewed, with solutions posted but not graded. I teach in a low stakes, not very selective institution now, very much like JaneB but without the majors in my field, I mostly teach for engineers.
    I used to assign about 30% of course grades to lab assignments, but I'm changing that - I'm replacing assignment grades with 2 or more additional tests next year. The reason for this is that I found that having graded lab work just encouraged groups of students to copy the work (where you have to get the right answer this is easy) and there were a disturbingly large number of cases where students did extremely poorly on all the exams but still scraped through due to lab work and homework pushing the grades up. So I could assign a smaller portion of grades to lab work, but then they don't bother with it at all, so I'm going for testing individually and requiring that the work be submitted even though I will not have formal grades. They will still be marked and given feedback, but all the grades will come from testing or research paper work.

  7. Wow. Thank you all for these perspectives.

    JaneB and Pramod, what strikes me most from your comment is the thought that homework is repetitive and tedious. In my experience (only at elite schools) is that the point of homework (except possibly in the very basic courses) is to help synthesize the material from lecture, not to reiterate it. If concepts 1-10 were talked about in a week, the 5 homework problems for the week would show applications of some of the and how some of them relate to each other. When I think of college homework, I think only of this type, not of any other. Does this mean that you have to explore the ideas the professor wants you to explore, an not others? Yes, this is a downside of the system. One of the complaints my partner is getting about his course is that his homeworks are NOT repetitive, like what the students are used to. Instead, he asks them to go beyond what is in the texts or lectures, and to use material covered in the previous term to solve problems this term. So maybe part of the problem in comparing systems is what is meant by homework?

    Pramod, I completely agree with you that some of this has to do with how well a student manages to engage with a subject matter. There are courses I never got on top of that neither system would have helped me get a better grasp of. There are courses that I loved that I may have done well in either system, because I would have gone the extra mile. And there are the middle courses that I don't know if I would have survived without the aid of regular evaluations that helped me synthesize the material, not just repeat it. Maybe the question I should be asking is which population of students is better served by each system?

    Daisy and JaneB, I have to admit, I am completely clueless on how to deal with unmotivated students in a college setting. I hear what you are saying, about homeworks being a chance for them to slack off or crib, or whatever. Maybe Daisy's solution of lots of exams (where you have to show that you understand the material is a good solution there. I hope it works well for you. I guess my question is: Do you think, for unmotivated students, a more highschool-like system, with lots of low stake exams showing individual capacity, along with an end of year/semester final is better for these students, or would they be better served by a single exam at the end of the year (with or without graded or ungraded assignments along the way)?

  8. For unmotivated students: too many exams, and they lose the fear factor, which results in terrible study habits, but one large exam would be disastrous. I've opted for 3 exams per course, each increasing in value of final grade but equivalent in difficulty. The first one is usually terrible (most of the class fails), the second one is much better (most pass) and the final (usually 30-40% of course) is somewhere in the middle and reflects a pretty accurate picture of what they actually know.
    Has to be a balance between too many/too few, and in a range where I can handle the marking/feedback since that is so important.