Related Weblinks

In a Time Far Far Away...

...or perhaps, in all times that testing operates...

A fundamental question is this: what if my decision is wrong? What if my test tells me that a particular student can (or cannot) perform the skills I've assessed, when in fact, the opposite is true?

We approach this problem in our book through a philosophical device called a 'truth table', which is actually an adaptation of an idea from statistics:

For example, suppose we run an ESL placement test, like the UIUC EPT. A 'Yes' inference means: this person has sufficient command of English to take full-time coursework at the university, where all instruction is in English. A 'No' answer means: this person does not, and further ESL classes are needed.

If we get it wrong, there are two possibilities: we might say that the person's English ability is strong enough for full-time study, but subsequent interviews with the student show that English was a problem. This is a 'false positive'. Our 'yes' decision turned out to be wrong.

A false negative would happen when we hold the person back and require further English classes. We find out that the classes did not help the person, or more accurately, that teachers (in those classes) later report that this individual did not belong in further ESL instruction.

Strictly speaking, we never know the 'state of the world' – we can only estimate it. We can interview the student, his/her teachers, obtain whatever data we wish, but the state of the world is always somewhat of a guess.

False positives and false negatives have varying consequences, and any effect-driven testing system should bear those in mind. One thing is probably certain. There are areas of great social concern where the danger of a false positive or a false negative has far greater consequences, such as predicting whether (or not) a criminal will be dangerous.

To Boldly Go

Our book seeks to stimulate and disturb. We hope to bother the reader about language testing and its relationship to two world : that of broader educational and psychological testing, and second, to the world at large.

We close our book with an exercise we repeat here, we hope in the same spirit that we present it in the book, have fun. Only by having fun can the trek be optimistic, at all:

• Keep a testing diary. Track your own awareness of testing and seek the pragmatic effect of this book. Continue to become a testing citizen, to participate in the dialectic that is testing, for it is a conversation to which all are welcome, and unless we all participate then it is not a dialogue at all but rather a monologue – and therein resides the same fears and powers which (we suspect) brought you to critical awareness of tests in the first place.