Kids will cheat on tests. Anyone involved in education knows that, as do most people who are not.
Sometimes, when there is an incentive to do so, teachers will even help their students to cheat on tests. That incentive could be financial, or prestige (or a lack of it).
This makes anything that depends upon standardized tests suspect… unless you can make those tests impossible to cheat.
Human ingenuity being what it is, that may not truly be possible… but you can make it incredibly difficult for them to do so. In my previous article on education I wrote a tiny bit about it, but I thought it was worth a more in-depth look.
If you want to change the world, you have to change the schools.
The first topic toward that goal that I’m going to cover is how to make it impossible to cheat on tests:
Writing The Tests
The first part of making tests that are nearly impossible to cheat is to move the writing of the tests a step back from people with a direct association with the person taking the test. That has the added benefit of keeping teachers from giving extraordinarily easy or difficult tests for the same class.
So how do you do this? You create groups, maybe at the district level, but probably at the state level, of people with knowledge in the field to be tested. These groups are likely to be mostly, if not all, people trained in education… either people who have been teaching for quite a while, or at least people with a degree in education, with a minor in whatever field.
You then have these groups write far more questions than will show up on the test. For a test that is 20 questions, for example, you should have at least 100 questions. Depending on the material, this could end up with similar questions, just written differently (subjects like history might require this), or it could be a lot of truly different questions (math, for example).
These questions would be entered into a computer, along with the right answer, if there is one (English questions may have more than one right answer, for example). This provides a master list of questions for the test.
Administering The Tests
When it comes time to administer the tests, the students would sit down at a computer, which would be connected to the central database (or a local copy) that contains all the questions. These computers would be essentially dumb terminals, only able to run the program that administers the tests, to prevent hacking and/or accessing the internet to try to use Google to find the answers. Each student would get 20 questions selected randomly from the 100 total available (from the example above).
That makes it nearly impossible to have a cheat sheet small enough to hide… because you would have to have the whole question AND answer written down. It also makes one a lot harder to create, since you can only put down the answers to the questions you received, though of course people would get together to compile larger lists when possible… it still increases the effort required considerably.
Having the test administered by a computer also makes it considerably easier to detect patterns of cheating. If, at one school, students get average scores until after one particular student (or group of students) takes a test, that could easily indicate cheating. If one teacher’s class always gets the same questions right and wrong, that might also be a sign of cheating… or an indication of what areas they need to teach better, thus improving education as a side effect of trying to detect cheating.
Grading The Tests
All of the answers to the questions would also, obviously, be entered into the computer. These answers could then be graded by someone who is not connected in any way to the student, quite possibly without even having any idea who the student is (ie only having an id number for the test they’re grading, not knowing whose test it is).
Having someone not connected grading the tests removes any possibility of the grade being affected by the teacher liking, disliking, or even feeling sorry for a student. It limits it to only grading answers to questions, with no social aspect involved.
The plan outlined above removes any chance for a teacher to let people slide through, or grade them extra harshly to make them fail. It also makes it easy to tell a teacher what area of their subject they may be neglecting, based on students missing more questions than normal in that area.
This is not to say, however, that all, or even the majority of teachers do these things… and many of those who do may do so subconsciously, not intentionally. A lot of teachers get into it because they have a passion for teaching, and the thought doing those things is repulsive. Even acknowledging this, though, it can’t hurt to remove the ability and the temptation to do so.
Freeing teachers from having to write and grade tests also gives them the ability to focus more on the actual teaching… not to mention increasing their free time considerably. They would still be able to easily see the resullts of their students, so they could know if someone needed extra help, or if they hadn’t covered an area well enough.
And in the mean time, people trained in analysis could look at the numbers, and quite possibly come up with solutions to various problems by looking at how different variables cause scores to rise or fall, things such as teacher experience, school policies, and possibly even from which school the teacher got their degree, not to mention standard things like student demographics (age, ethnicity, gender, etc.).
So… comments are welcome. I’d like this to serve as a starting point for discussion on how effective you, the readers, think this would be, and any improvements you could suggest.
Depending on the number and quality of suggestions, I will either update this post with the best suggestions, or write a follow-up… If you leave a website, and I use your suggestion, I will link to it. Otherwise, I will just use your name you leave in the comments.