18
$\begingroup$

Ideally, I'd like my students' grades to be based on more than just exams, because I think that certain tasks, such as writing proofs, aren't best performed in a test environment. Using homework for assessment seems like an obvious way to accomplish this. But my idealized view of homework, formed when I was a student, seems to have little to do with modern reality. Educators on this site suggest that instructors solve homework problems during office hours, and my students seem to expect that I will do this, but the ability of students to transcribe solutions that I provide them with does not demonstrate understanding. Working on homework in groups seems to be nearly universally recommended and very widely practiced, but how do I differentiate the signal of Student A's understanding from the noise of his/her study partners' understanding?

Are there any studies that compare students' perception of how homework problems are to be solved (e.g, by the student him/herself or by others) today vs., say, 30 years ago?

Are there "best practices" for making valid assessments of student understanding outside the environment of an exam?

$\endgroup$
  • 4
    $\begingroup$ It depends whether you mean freshman classes, like calculus, or upper level / grad classes. I've had plenty of "Topics" classes with no exams, just homework. $\endgroup$ – pjs36 Sep 15 '15 at 16:56
  • 3
    $\begingroup$ Personally I struggled for years trying to assess student homework (based on direction from a mentor I had in college), and it just drove me crazy all the time (it was frankly bad advice, at least for me). So glad I finally aborted that. Perhaps for longer, high-stakes items like proofs you could give an exam with several problem options from which to choose (as is done in many essay exams). $\endgroup$ – Daniel R. Collins Sep 16 '15 at 6:04
  • 1
    $\begingroup$ “No man who worships education has got the best out of education... Without a gentle contempt for education no man’s education is complete.” -- G.K. Chesterton $\endgroup$ – EsperantoSpeaker1 Oct 11 '15 at 16:33
5
$\begingroup$

This is not a full answer to your question, but these comments do not fit in the comment box.

Homework and exams measure different things. If you want to make sure that it is the student's own thoughts that you are reading on a paper, you should have a controlled exam. I think such control is often justified, so an exam is a good way to assess students, but it need not be the sole method. Taking homework as a part of your grading scheme gives value to the students' activity with the problems. Even if a student is copying the solution from someone else, he is putting at least some amount of thought into the subject and going through a solution. This is good; the student could do better, but this is far better than nothing.

And of course if you want to grade the homework, you should refuse to give much help with at least some (if not all) of it. Having two sets of exercises, ones that the instructors help with and others that have to be handed in for grading, seems to work well here. After solving some problems with help, the students can be expected to solve more problems without help.

I think activity with homework is a sound criterion for assessment. This for (at least) two reasons: (1) activity correlates with the competence you want to create in your students, and (2) giving points for activity motivates activity and causes learning as a side effect. Student assessment is not only for seeing who knows their trade and who doesn't, but also for pushing students in the right direction.

There is often some social pressure against copying solutions from others. And in group homework most students will master at least some part of the topic, and that is often enough. If you want to make sure that everyone masters the same thing, an exam is in order. It seems that my answer to your first question is this: it is not always that important to differentiate the signal from the noise (the student's understanding from his friend's understanding).

If you use both an exam and graded homework, the role of the exam can be very different. You can just test the understanding of the basic concepts and ideas in the course and the homework project; leaving the goriest details to homework changes the tone of the exam quite a bit.

$\endgroup$
  • 1
    $\begingroup$ To counteract the "just copy it from somewhere", the homework grade is for what was handed in, or for a "randomly" selected (small) sample an interrogation on what was handed in. I'm not so interested in them doing the work as in them understanding the stuff. $\endgroup$ – vonbrand Sep 17 '15 at 11:27
3
$\begingroup$

Making your students to learn something is more important than grading.

Grading homework is good for one thing: it makes your students to do the homework. No matter how much they cheat, it helps them to learn something. Therefore you have to grade it.

$\endgroup$
3
$\begingroup$

I like @Joonas explanation of why homework is valuable, and how it can philosophically fit in with exams.

I believe that there are few students who go to the effort of copying an assignment, but not put in enough effort in understanding the assignment that they deserve no credit. Most students will at least attempt an assignment, or will make zero effort and not produce anything. So I am comfortable with giving a completion credit to any assignment that is turned in.

I generally do not like grading homework for more than completion because I feel that places too much of an incentive to copy answers. I have wanted to make homework more effective and worth a greater portion of students' grades. In an effort to reduce the incentive on copying assignments and make homework more valuable I tried a couple strategies:

I have a tried to make grading work by checking homework at the end of sections, allowing time for students to complete work and fix mistakes if they didn't get the homework assignment at first. This worked reasonably well but was a great burden on my time. Grading tests and detailed notebooks at the same time was difficult.

I required students to choose a problem from a subset of their homework to copy and submit. I liked this and have continue to use it for certain classes. Works well with other entrance/exit card strategies.

One idea I have had, but never tried, would be to create quizzes that are similar to homework problems and allow the students to complete the quizzes with open notes. I have not tried this because of the class time required. I don't like the idea of how long it can take some students when presented with the option of using their notes, and I don't like the idea of making it timed and cutting them off.

$\endgroup$
0
$\begingroup$

No. In high school, you do homework. In college, you study. This is precept #11 in the list of differences between high school and college. For the full list: http://www.public-domain-materials.com/folder-the-difference-between-high-school-and-college.html

$\endgroup$
  • $\begingroup$ Link is now broken. Unfortunately, this makes this a mostly non-answer at this time. Delete? $\endgroup$ – Daniel R. Collins Dec 30 '17 at 18:37
  • $\begingroup$ I mostly disagree with this attitude and believe it is a copout for colleges (which have beautiful lawns and research programs and fancy buildings but poor pedagogy). I did post bachelor's training in the Navy (pretty intense, Navy nuke school) and they did not have the "this isn't high school" attitude. Instead...it was important to them that you learned and they used strong structure in many ways similar to HS. And taught a heck of a lot of info in a very short time. $\endgroup$ – guest Jul 7 '18 at 18:00
0
$\begingroup$

I always felt that I got better mastery from courses with a lot of drill and testing. I think graded homework appeals to professors but that says more about how professors think than if it is best pedagogy. In particular, I found that (in grad school sciences) that graded homework tended to be very hard and that then not enough progression and easy problems were completed. Can lead to big gaps on mastery. And can be time inefficient if all you are doing is b$%|-buster problems.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.