Textbook Test Banks and Online Exams

[bio-nonymous]

Are your tests timed?  If so, you are putting folks who are simply slower readers at a real disadvantage.  That doesn't mean they aren't smart, but it makes time spent reading weigh heavily on their score.  They may be doing very well and then simply run out of time to answer all the questions. Now, in some fields, reading speed is a very necessary skill (e.g. law), but that's not true for most jobs.  It's also unfair to English-language learners to "hide" a simple question in a complex story.
[/quote]

I teach to professional students whose board exams are timed at 1 minute per question, I give them about 25% more time than that. Students with accommodations get time and a half. It is true that complicated questions take more time to comprehend than straight forward A-D multiple choice, so I try to counteract very difficult involved questions with a few True/False--you either know it or you don't and the question gets done in a few seconds...

[Baldwinschild]

Are your tests timed?  If so, you are putting folks who are simply slower readers at a real disadvantage.  That doesn't mean they aren't smart, but it makes time spent reading weigh heavily on their score.  They may be doing very well and then simply run out of time to answer all the questions. Now, in some fields, reading speed is a very necessary skill (e.g. law), but that's not true for most jobs.  It's also unfair to English-language learners to "hide" a simple question in a complex story.

I teach to professional students whose board exams are timed at 1 minute per question, I give them about 25% more time than that. Students with accommodations get time and a half. It is true that complicated questions take more time to comprehend than straight forward A-D multiple choice, so I try to counteract very difficult involved questions with a few True/False--you either know it or you don't and the question gets done in a few seconds...
[/quote]

I restrict the time for tests (humanities) because I think spending too much time on an exam is not beneficial.  It the exam time is limited, they are more likely to read the study materials and take notes.  More important, after about two hours, they are mentally exhausted.  To the point about non-native speakers and complex questions:  the questions should address the course materials, so non-native speakers should not be encountering unfamiliar terms when they are presented with complex questions.  I think if they can handle the course reading materials, they will able to handle complex questions. 

In my experience, students do not perform better on untimed tests.  I have seen a few students meltdown after spending a long time on an exam.  That said, the time limit should be reasonable.  I try to give them way more time than they need.  A time limit doesn't necessarily mean they will be in a race to the finish.  It just means they have a certain amount of time to complete the exam. 

[HigherEd7]

If giving them a complicated stem and long answer choices tests their ability to read process information in addition to their memory of content or concepts, I think it makes a better exam.

But then students complain that the exam is unfair because smarter people have an advantage.

Are your tests timed?  If so, you are putting folks who are simply slower readers at a real disadvantage.  That doesn't mean they aren't smart, but it makes time spent reading weigh heavily on their score.  They may be doing very well and then simply run out of time to answer all the questions. Now, in some fields, reading speed is a very necessary skill (e.g. law), but that's not true for most jobs.  It's also unfair to English-language learners to "hide" a simple question in a complex story.

Yes, my exams have a time limit. And from my understanding this is a standard in online learning. Thoughts?

[Hegemony]

Yes, you certainly need a time limit.

[the_geneticist]

You do need a time limit and a due date.  But don't use the time limit as a way to try and make an easy test harder.  I have colleagues that think giving 60 multiple-choice, recall-level questions in 60 minutes is a "good test".  Why? 60 minutes is too short to look up most of the answers (true); it limits the ability to cheat your way to a high score (also true).  But it's testing students on 2 things: 1) basic knowledge 2) reading speed.
I think it's better to ask application/prediction/design questions so their score is based more on their overall understanding.  Fewer questions, same time allowed.  That way sheer panic and reading speed aren't confounding factors.

[Aster]

If giving them a complicated stem and long answer choices tests their ability to read process information in addition to their memory of content or concepts, I think it makes a better exam.

But then students complain that the exam is unfair because smarter people have an advantage.

Are your tests timed?  If so, you are putting folks who are simply slower readers at a real disadvantage.  That doesn't mean they aren't smart, but it makes time spent reading weigh heavily on their score.  They may be doing very well and then simply run out of time to answer all the questions. Now, in some fields, reading speed is a very necessary skill (e.g. law), but that's not true for most jobs.  It's also unfair to English-language learners to "hide" a simple question in a complex story.

Yes, my exams have a time limit. And from my understanding this is a standard in online learning. Thoughts?

Yes. Timed tests are standard for virtually all forms of education. It does not matter if the format is online or not.

If you are concerned about "slow readers" (which would imply someone requiring special accommodations and registered with your disabilities office), then either make your timed window longer, or program in extensions for whatever students are requesting the special test taking accommodation.

[in2ny]

I always write my own questions.  Test bank questions are also typically based on memorizing definitions (easy to memorize, easy to cram).  I like to give students a figure/diagram/graph and ask questions about it.  Super easy to swap out the graph or write variations on the questions on my end.  Students have to apply their knowledge to do well.

Example: graph of goldfish weights over time with 2 types of food.

Question:  Examine the above graph.  Based on what you have learned, these data _______ the null hypothesis "There is no difference in fish weight gain between fish fed Brand A or Brand B food".  Therefore, diet _____ an important variable in fish weight gain.
A. support; is NOT
B. support; is
C. do NOT support; is NOT
D. do NOT support; is

Looking at days 5-10, what is a the average rate of weight gain during this time for Brand A?
A. 10 grams
B. 0.5 grams/day
C. 5 days
D. not enough information to answer

These are both fantastic examples and techniques that I use. I also use "impossible to tell" similarly to "not enough information to answer"; the mere presence of this option helps sort out the weak students, who tend to over-select that answer rather than guessing one of the others (i.e., they assume that if they can't figure out the answer, it must be impossible; in reality this is the answer usually only once in the whole exam). Finally, I often give questions that can be answered based on a formula but don't give all the numbers involved so the student can't actually calculate the answer-- they have to really understand the formula.

Needless to say in an online exam, I don't give any questions that simply ask the definition of a term as I know they can just look it up (unlike in-person exams, where I throw in a few of those just to see who doesn't have a pulse). 

As a result, my online open-book exams tend to have an average score that's about 10 points lower than my in-class closed book exams (yeah, I curve it up so they're comparable across sections/semesters).

[fosca]

For gen-ed survey courses, since I'm awful at writing questions I've been known to use the test bank for a different text.  I go through the questions to make sure that the information is covered in my course (using the same terms, particularly for things like statistics) and that it's a good fair question, and also scramble the questions/answers and select from a pool so students don't get the same test. 

I also use 1 minute per question for "simple" multiple-choice questions (very few of which are just definitions--most are application or understanding), and add time on for more complicated items.  Students tend to finish in plenty of time, so it hasn't been a problem.