koeneman@trex.rutgers.edu (Juergen Koenemann) (11/13/89)
I am looking for faculty member for some software which has the following functionality: Randomly selects a number of multiple choice questions from a test bank Student answers questions on terminal (with time limit) Program records performance, time, answers etc. and writes data to file. The identity of the student will be checked by personel, some protection against fraud (access to test bank, to data) would be nice Any hints (or source code) are appreciated. The system is not yet determined (either MAC Plus or IBM PC), the language isn't important either. The questions are currently not stored electronically. Since some money is available, programs which fulfill parts of the functionality or pointers to commercial products are also fine. -- Juergen Koenemann koeneman@paul.rutgers.edu
reggie@dinsdale.nm.paradyne.com (George W. Leach) (11/13/89)
In article <Nov.12.21.44.18.1989.2325@trex.rutgers.edu> koeneman@trex.rutgers.edu (Juergen Koenemann) writes: >I am looking for faculty member for some software which has the >following functionality: >Randomly selects a number of multiple choice questions from a test >bank [Other stuff, which could cause other discussions on security, etc.... have been deleted] Well this is refreshing! Normally we don't see these kinds of requests in the middle of a semester :-) How many times over the past several years have we seen similar requests that end up generating a rather lengthy debate over the validity of such an approach to testing what students have learned? I have refrained from joining into the debates for several years now. However, I feel compelled to jump in this time. Multiple choice tests are easy to administer, easy to take, and easy to grade. Perhaps they are even a fair means of comparitively judging student test scores. Why? Because there is no ambiguity. Something is either right or wrong (assuming there are no choices that are misleading). There is no judgement on the part of the grader as to how many points should be taken off or awarded for various aspects of an answer to a question. However, it is my feeling that these types of questions are not a good means of testing exactly what students do and do not know. Nor are they helpful in the process of showing students exactly where they went wrong with their answers. Remeber tests can be used to aid in the learning process, not just to evaluate. Some of the most valuable lessons that I have learned in my life have come from mistakes I made in formulating solutions to problems on test. If the work was not on the test when they were handed back when we went over the test, I may not have learned those lesson. I may not have saved my work or remembered how I approached the problem. Learning from our mistakes is a very powerful tool. I'll take this even further. I always give open book tests. I try to come up with problems that will make people think and apply what they have learned to solve problems. I do not want them to simply repeat back To me something that has been memorized. So on a programming problem I will put the emphasis on evaluating the problem, formulating a solution, and expressing that solution in the programming language that is being learned. I do not use simple questions that test a student's knowledge of syntax. That is not important. Syntactic knowledge can be looked up in a book. Also, as one uses a language more frequently, the syntax becomes second nature. However, the problem solving skills that are required to do something with the syntax is what students need to learn. They can be applied to many situations and with many languages. This does involve more work on the part of the instructor to formulate a test, set up the grading criteria (partial credit is difficult to decide upon and often is difficult to keep objective across all students' work), and actually grade the tests. It is also a bit more difficult on the students who must take the test. However, in my opinion it is all worth it. Students will learn more in the long run. George George W. Leach AT&T Paradyne (uunet|att)!pdn!reggie Mail stop LG-133 Phone: 1-813-530-2376 P.O. Box 2826 FAX: 1-813-530-8224 Largo, FL 34649-2826 USA
sullivan@aqdata.uucp (Michael T. Sullivan) (11/15/89)
From article <6741@pdn.paradyne.com>, by reggie@dinsdale.nm.paradyne.com (George W. Leach): > > However, it is my feeling that these types of questions are not a > good means of testing exactly what students do and do not know. Nor are I don't know. I had a GE class where the multiple choice test was HARD and I felt it did test what I knew. Multiple choice tests are like anything else, they're just a tool that can be properly used or abused. > This does involve more work on the part of the instructor to formulate > a test, set up the grading criteria (partial credit is difficult to decide > upon and often is difficult to keep objective across all students' work), and > actually grade the tests. It is also a bit more difficult on the students who > must take the test. However, in my opinion it is all worth it. Students will > learn more in the long run. This is fine if you aren't teaching a large GE class with hundreds of students. In cases like that, it makes sense to give multiple choice tests because they are easier and objective. A student in Psych. 1 doesn't need the attention to detail that a student in an upper-level CS course does. ... > wrong with their answers. Remeber tests can be used to aid in the learning > process, not just to evaluate. Some of the most valuable lessons that I have > learned in my life have come from mistakes I made in formulating solutions > to problems on test. If the work was not on the test when they were handed This is just a separate note on the above. In some classes, tests count too much to learn from them. What I mean is that missing 30% of the questions and learning from the corrections on your test _is_ a great learning experience. Problem is, if that test was worth 40% of your grade for the quarter it's not much help in the grade department. Another reason I disliked school so much. -- Michael Sullivan uunet!jarthur.uucp!aqdata!sullivan aQdata, Inc. San Dimas, CA
reggie@dinsdale.nm.paradyne.com (George W. Leach) (11/15/89)
In article <1989Nov14.161530.6224@aqdata.uucp> sullivan@aqdata.uucp (Michael T. Sullivan) writes: >From article <6741@pdn.paradyne.com>, by reggie@dinsdale.nm.paradyne.com (George W. Leach): >> However, it is my feeling that these types of questions are not a >> good means of testing exactly what students do and do not know. Nor are >I don't know. I had a GE class where the multiple choice test was HARD >and I felt it did test what I knew. Multiple choice tests are like anything >else, they're just a tool that can be properly used or abused. Yes, I guess that is true. However, just like determining the appropriate questions to ask on a survey and how to ask them, they are not easy to design well. I have seen more poorly designed multiple choice tests than good ones over the years. An essay type or problem solving type question can also be poorly designed. It is difficult to come up with a problem that will accurately test the knowledge of a student, present a challenge, and not take up the entire test time. Furthermore, if students can not see how to take a specific problem and solve it, there is no guessing. On a mutliple choice test, the process of elimination can reduce the chances of a wrong guess and one might just get lucky. They teach these methods for the SATs!!!! Of coures, one could argue that this is applying a type of problem solving :-) >> This does involve more work on the part of the instructor to formulate >> a test, set up the grading criteria (partial credit is difficult to decide >> upon and often is difficult to keep objective across all students' work), and >> actually grade the tests. It is also a bit more difficult on the students who >> must take the test. However, in my opinion it is all worth it. Students will >> learn more in the long run. >This is fine if you aren't teaching a large GE class with hundreds of students. >In cases like that, it makes sense to give multiple choice tests because they >are easier and objective. A student in Psych. 1 doesn't need the attention >to detail that a student in an upper-level CS course does. Yes, I have heard from some others via e-mail about the problems of managing a large class. My opinion is that teaching large classes is not a desirable situation for learning. However, the economics of many departments dictate such an approach for certain types of courses. I don't teach in those settings. I also feel that most CS courses are not taught that way either. >... >> wrong with their answers. Remember tests can be used to aid in the learning >> process, not just to evaluate. Some of the most valuable lessons that I have >> learned in my life have come from mistakes I made in formulating solutions >> to problems on test. If the work was not on the test when they were handed >This is just a separate note on the above. In some classes, tests count >too much to learn from them. What I mean is that missing 30% of the >questions and learning from the corrections on your test _is_ a great >learning experience. Problem is, if that test was worth 40% of your >grade for the quarter it's not much help in the grade department. Another >reason I disliked school so much. (1) Are you there to get good grades or learn? (2) As I stated in my original note, I provide people with the opportunity to take on extra credit assignments that will help in the grade department. I try to tie them in with test problems that presented folks with difficulty. So, they have a chance to demonstrate that they have learned. George W. Leach AT&T Paradyne (uunet|att)!pdn!reggie Mail stop LG-133 Phone: 1-813-530-2376 P.O. Box 2826 FAX: 1-813-530-8224 Largo, FL 34649-2826 USA