11/08/04

Are You A Deviant?

(Note: Many tests used by psychologists are little more than mumbo-jumbo masquerading as science. In this particular case concerning a purported test of sexual deviancy , I attempt to show you why such tests are often invalid, and why, in order to protect yourself, you would be better off by not telling the truth to the psychologist.)

Abel Assessment A test of sexual interest and deviancy

Well: I can categorically assure you that tests such as these are often little more than money-makers - but people - men - are very often assessed for deviancy on the basis of such tests. And so, for example, men in contested divorce cases might be denied access to their very own children on the grounds that such tests suggest that they are 'deviant'.

But there is a real problem with this.

 these tests ... have virtually no validity when assessing individuals.

These tests might have some validity in the sense that they can differentiate statistically between groups of people, but they have virtually no validity when assessing individuals.

For example, imagine a test that tries to assess people's weights simply by asking subjects two questions.

1. What is your gender?

2. What is your height?

On the basis of the answers to these two questions, such a test would be pretty good at assessing the average weights of the groups of subjects who fell into the different categories - men/women - tall/short.

This is because, for example, taller men tend to be heavier than shorter men, and because men tend to be heavier than women for any given height.

And so on.

And so, for example, if 100 men took this test and each said, "I am a man who is x foot tall," then the assessor would, on average, be able to assess quite well the average weight of the group.

Of. The. GROUP!

Thus, for example, the assessor would be able to assess reasonably well the average weight of a particular group of 'tall' 'men'.

But such a test would not tell us much that was reliable about the individuals themselves.

And so, for example, on the basis of such a test, a tall thin man would be much lighter in weight than the test would assess him to be. He might even be lighter than a short woman.

And the reason for this large error would be that such a simple test would have failed to take into account the fatness or thinness of the subjects taking the test. In other words, the test would not have taken into account all the variables that are relevant when it comes to assessing weight.

Thus, without knowing about the fatness or the thinness of a particular man, our test is quite likely to be way off the mark when it comes to predicting his weight.

Well. When it comes to 'psychological' matters, the variables that are likely to be relevant to them are not only positively vast in number and infinitely complex, but they are also very likely to be unknown.

And so while psychological tests might be of value when it comes to differentiating between groups of people, they cannot necessarily be relied upon to assess individuals accurately.

For example, one of the most thoroughly researched areas when it comes to psychological testing is that connected with the assessment of intelligence.

Some of these tests have been researched fairly extensively for almost 100 years. And some of them are very good indeed.

And in some 90+% of cases, these 'good' tests seem to predict pretty accurately how well individuals are likely to perform in cognitive tasks that require the use of 'intelligence'.

But there are many occasions where, for some reason or other, even these thoroughly-researched tests give results that are way off the mark for particular individuals.

 none of these tests should ever be used to justify serious consequences for any of the individuals who took them

As such, none of these tests should ever be used to justify serious consequences for any of the individuals who took them, because, as yet, they are nowhere near being valid enough nor accurate enough to warrant such consequences.

And, quite frankly, a quick glance at the Abel test website should tell you all that you need to know about the motive and the rationale behind the construction of this particular test.

It's called 'money'.

Look at the boasting (found on this page) ...

Half hour per client 

Easy to interpret – one page graph

Probability Values

Computer scored questionnaire summary

The whole thing is almost obscene.

The whole thing is almost obscene.

And in much the same way that Tarot card readers will wax lyrical about how accurate and valid are their Tarot card readings in order to persuade people to partake of their services, many 'psychologists' will do the same for the very same reason.

Unsurprisingly.

But there is a further important point to be made about this so-called test for sexual predilections.

And it is this.

Intelligence tests are also researched thoroughly with regard to how they predict future behaviour.

In other words, the researchers will check to see whether or not the people whom they have tested (e.g. children in a school) did, in fact, end up many years later performing as expected.

If not, they will jigger the various weightings and the factors so that future intelligence testing is likely to be more accurate.

And, as the years go by, more accurate will they indeed become - up to a point.

But the Abel test has not been developed in this way. It has not had time to develop in this way. And, further, this test appears to be based simply on comparisons between alleged molesters and alleged non-molesters who answered various questions and looked at sexy pictures

there is a fundamental flaw in the thinking behind this test ...

As such, there is a fundamental flaw in the thinking behind this test that could easily lead those who might be investigated by it to incriminate themselves when all they are doing is trying to demonstrate their innocence by being honest.

Basically the Abel Test tries to determine whether or not an individual is likely to engage in 'deviant' behaviour. And so what the constructors of this test have basically done is this.

They have asked those convicted of deviant behaviour to complete the test and they have compared the answers of these individuals to those who have not been convicted of any such offence. They have then determined which questions discriminate best between the two groups.

And so, for example, if those who have been convicted of deviant behaviour mostly answer Yes to questions 1, 2 and 3, whereas those who have never been convicted of such a thing mostly answer No, then the thinking is that if psychologists test someone about whom they know nothing, then they will be able to determine (with some degree of probability) which group they fall into.

And so if an individual is asked to take the test and he answers Yes to questions 1, 2 and 3, then he will be categorised in the same group as the convicted offenders (i.e. it will be argued that he is likely to engage in deviant behaviours) whereas if he answers No, then he will be categorised as someone who is not likely to offend.

But there is a truly fundamental problem here. And it is this.

Those who have already been convicted of deviant behaviour have no real incentive for lying, or for deceiving the psychologists. Indeed, they might even be getting help from the psychologists and are, therefore, only too willing to help them construct their tests.

 those who have not been convicted of anything might well have good reasons for lying

On the other hand, those who have not been convicted of anything might well have good reasons for lying and for trying to deceive the testers.

For example, imagine that questions 1, 2 and 3 are as follows.

1. Do you have fantasies of having sex with young girls?

2. Do you enjoy fantasies of rape?

3. Do you dream about orgies?

Well.

Convicted offenders might happily answer Yes to such questions.

After all, they have already been caught!

But how would 'normal' men who have not been convicted of any such things answer such questions?

Might they not lie?

Might they not say to themselves, "No way am I telling the truth about my fantasies to these people. They'll think I'm a pervert!"

And so the point is this.

The constructors of the test will have been misled into believing that their test can be used to discriminate between those who are likely to engage in deviant behaviours and those who are not.

But the truth of the matter might be that their test merely discriminates between those who have been caught engaging in such behaviours and, perhaps, ordinary men who have not engaged in such things (or who have not been caught engaging in such things) - but who are not going to tell the truth about their fantasies in the tests - for obvious reasons.

In other words, the test might have been constructed by comparing men who have no reason to lie (because they have already been caught) with men who have every reason to lie (because they do not wish to be seen as deviant).

As such, if you are tested, and you happen to have some of the fantasies above, and, for the sake of honesty, you admit to them, then you will be categorised as someone who is likely to engage in deviant behaviours whereas, in fact, you might be no different from ordinary innocent men! - who, for the most part, do not engage in such behaviours but who, in fact, might fantasise about them without being prepared to admit to this.

the test is not actually discriminating between 'deviants' and 'normals',

In other words, the test is not actually discriminating between 'deviants' and 'normals', it is discriminating between, ...

1. Those who have been caught engaging in deviant behaviours and those who have not been caught doing so.

and/or

2. Those who do not feel the need to lie and those who do feel the need to lie.

Well. OK. The three questions above are a gross simplification of what the Abel test involves. But the criticisms of the test - as outlined above - give some indication of just how flawed this test is likely to be.

Imagine, for example, that you wanted to construct a test to figure out which men were likely to commit adultery.

Imagine, for example, that you wanted to construct a test to figure out which men were likely to commit adultery. And, just for the sake of argument, assume that all men have fantasies about committing adultery.

You take a group of convicted adulterers and, in the presence of their wives, you ask them this question.

Have you ever had thoughts about committing adultery?

They all answer Yes.

Unsurprisingly.

You now take a group of men who have never been convicted of committing adultery and you ask them, in the presence of their wives, the same question.

Have you ever had thoughts about committing adultery?

They all answer No.

Unsurprisingly.

A perfect correlation!

So. Do you now have the basis for a valid test that determines the likelihood of a man committing adultery?

Hmm.

Let's see.

A wife brings her never-convicted husband along and says to you, "Please give him the test."

And you duly oblige.

You ask him, "Have you ever had thoughts about committing adultery?"

If he says Yes, does this mean that he is extremely likely to commit adultery? - after all, your test validation procedures showed a perfect correlation - and from this it follows that he will certainly commit adultery.

But this is clearly nonsense. He might just be a rare soul who is being honest about his thoughts. 

And what if he says, "No, I have never had any thoughts about committing adultery."

Well, according to your test, this man is definitely not going to commit adultery.

But do you necessarily believe him? - especially as his wife squints her eyes with menace and scrutinises his face closely.

The reality, of course, is that whichever way he answers, you will most likely have learned precious little about the likelihood of him committing adultery, even though when you constructed your test it showed a perfect correlation between the answers to the test and the (apparent) behaviours of the subjects who gave them.

And the main reason that a test such as this would be invalid is because it has been developed by means of comparing the responses of a category of people who have a large incentive to lie (viz, the group of people who have not been convicted of anything but who, nevertheless, would not likely wish to reveal their private thoughts) to those who have already been caught and who have far less of an incentive to lie.

 if you find yourself being assessed by the Abel test then giving honest answers is probably not a wise thing to do!

And, regretfully, as such, if you find yourself being assessed by the Abel test then giving honest answers is probably not a wise thing to do!

In other words, it is probably best to respond to the test in much the same way as did those supposedly non-deviant men likely respond to it; viz, by lying - if they are, in fact, 'deviant' - or by not lying - if, in fact, they are non-deviant.

Either way, these men said, "No Siree. Not me. I never have such thoughts."

And the upshot is that if you are asked to take the test then, whether or not you have deviant fantasies, it is best to give the impression that you don't - because honesty could easily land you in serious trouble even though you have not, and will not, ever engage in deviant behaviour.

Part of the test appears to involve subjects looking at various sexy pictures

Added Note 1:

Part of the test appears to involve subjects looking at various sexy pictures - some of them allegedly 'deviant' and some of them not - and assessing how much the subjects like them by observing how long they choose to look at the various categories of pictures.

Well. Apart from the fact that it must be blindingly obvious to subjects what the whole thing is about - which means that they can easily 'lie' by, for example, looking longer at the non-deviant pictures - there remains no real justification for assuming that the longer that one looks at a picture, the more is one being sexually aroused by it. For example, it might simply be the case that a subject has so little experience of viewing material contained in 'deviant' pictures that he is far more interested in looking at such pictures than at the ones which contain more 'normal' material and, hence, material which he has probably seen many times before.

In other words, his 'interest' in 'deviant' pictures as reflected solely by how long he looks at them might be nothing more than a reflection of his added curiosity over something that he has rarely encountered before.

Human brains are much more complex than psychologists often assume them to be.

Added Note 2:

The type of video games that I have most enjoyed in the past are those wherein various warriors of old beat the hell out of each other. They impale each other with axes and swords. They beat each other ceaselessly with fists and feet. They knock each other unconscious. And, allegedly, they are 100% to do with 'violence'.

But, in fact, they are all fantasy, totally unrealistic, and no-one actually gets hurt.

Indeed, human beings are perfectly capable of enjoying fictional representations of situations that they would positively hate to encounter in real life.

But the real point here is this.

If the public was ever indoctrinated successfully with the view that playing 'violent' video games was a valid mark of serious 'deviancy', then I would not be admitting to my enjoyment of 'violent' video games to psychologists who were assessing me for such deviancy.

And nor would most men who enjoyed such games.

Unless, of course, they had already been caught playing such games.

...

Also see the Ink Blot Test - which is a completely bogus test that is used by many clinical psychologists to assess, for example, the suitability of a man to be a father to his own children.

To gain some further insight into just how discreditable the results of testing for psychological 'deviancy' might be, the following article describes how a test that simply tests for drug abuse with a purported 99% accuracy level will, in fact, lead to about one-third of those who fail the test being inaccurately labelled as drug abusers. (A basic understanding of probability theory is required.)

We like to think these tests are at least 99% accurate, and yet, horror stories of spurious results seem to abound. Take company-wide drug testing, opponents may claim that at least a third of those identified as drug users will actually be innocent. If we assume the test is 99% accurate, this claim sounds ridiculous. But is it?

Also see,

Fathers Who Kill

 Depressed Females

Dealing With False Accusations