Tuesday 30 August 2016

When it goes, it goes




My first ever proper job was as a research assistant at Guy’s Hospital Medical School, on a project which led to my PhD thesis:  “Cognitive effects of cortical lesions sustained in childhood”. After 7 long years I finished it, the largest sample of people with brain injuries in childhood who had been followed and fully tested, children who had been injured in car crashes, kicked in the head by horses, and in one case seen 31 years after he had been bombed in his bed as a baby in the London Blitz. I tested their intelligence on the Wechsler scales, tested their learning of new materials, and even their visual dot spans with a portable tachistoscope. My beautifully bound thesis was deposited at the University of London library, and when I happened to be at a meeting in that fortress of a building a decade later, on authorial impulse I went into the library and asked to see my own thesis. I found that it had achieved only one reader, myself in 1976, proudly signing in to check that my contribution to knowledge was available to the interested public. Sic transit gloria mundi

I was supervised by Dr John McFie, and one of the many things he taught me about was the work of Godfrey Thomson, whose 1939 Factorial Analysis of Human Ability had a great impact.   Now Prof Ian Deary has put together an exhibition about him and his work, which is displayed in the Edinburgh Main Library and runs until 29th October. Ian is a worried man. He is studying Scots men and women who took Thomson’s tests as 11 year olds in 1932 and 1947, and his now elderly subjects are often concerned that, as regards the power of their minds, when they go, they go. Out goes the light. However, in Ian’s case he is not worried about his own ageing but about the painful fact that, unlike a book or a published paper, when the exhibition goes out of the library at the end of October, it all goes. Gone, gone, like the proverbial baby down the plughole, in that terrifying music hall lament, redolent of maternal agony and traumatic loss, to be assuaged with coarse communal jollity:

'Your baby has gone down the plug-hole
Your baby has gone down the plug
The poor little thing was so skinny and thin
It should have been washed in a jug”

As Noel Coward tartly observed: “Strange how potent cheap music is.”

I digress.

How can these exhibits be saved? One way is for you to look at his introductory video, and then make your way to Edinburgh, to see the exhibition and take down your own notes for posterity. You might be able to suggest a permanent home for the exhibition.


You can also get the explanations, without being able to palpate any physical objects, first in a brief news item, and second in a more detailed lecture:



There you have it. One modest scholar puts together some intelligence tests and we eventually get the longest and most informative intelligence follow-up study ever conducted. Try beating a 66 year follow-up when assessing pre-morbid intelligence estimates.  Have the Scots made a greater contribution to modernity than any other people? I could not possibly comment.


Monday 29 August 2016

Author replies: Prof Scott O. Lilienfeld replies on myths about mental illness and violence


You may recall a previous post about myths regarding mental illness and violence.


All authors get immediate right of reply, so I am grateful to get Prof Lilienfeld’s comments in reply which I post below:



A colleague alerted me to this blog post and Dr. Thompson kindly informed me about it as well, and asked me if I’d care to respond. I’ll be more than happy to do so.

I apologize in advance if my tone is a bit harsh. But I have to confess that I found it extremely challenging to respond to Dr. Thompson’s blog post. Why?

It is one thing to respond to a substantive criticism of what one has written; it’s another to respond to a criticism of something that differs entirely from what one has actually written. In this case, Dr. Thompson’s blog addresses a very different question from that posed in our book. As a consequence – and I’m sorry that I can’t find any more delicate way to put it – his blog post is more or less one extended non sequitur. It addresses an interesting and important question, to be sure, but it is most certainly not the question we raised.

In short, Dr. Thompson’s method of argumentation appears to be as follows:

(1) The authors (in this case, us) claim that Statement A is a myth.

(2) I don’t agree that Statement A is as important or relevant as they do, and I wish they had instead addressed Statement B.

(3) Data indicate that Statement B is not a myth.

(4) Hence, Statement A is not a myth.

To be frank, I had to read the blog post several times to make sure that I wasn’t missing something, but I’m not. That is indeed Thompson’s line of argumentation.

To be clear, what we wrote is that the assertion that most mentally ill people are violent is a myth (people can legitimately question whether this is a commonly held myth among the general public, but numerous media and political conversations in the U.S. arising from recent mass shootings by apparently mentally ill individuals have referenced this claim or ones very much like it: http://time.com/4358295/violent-crimes-mental-illness/). We did not question whether mental illness is linked to a higher risk of violence; in fact, in our chapter we acknowledged that this link exists, and we described it as “modest” in magnitude (to his credit, Thompson acknowledges this point). One can legitimately question whether “modest” is too tepid a descriptor, but please see this recent (2012) study (https://www.infona.pl/resource/bwmeta1.element.springer-3ad5e3bd-3b5c-3703-9cf5-d072144333a6), and the authors’ conclusion:

“..there is a statistically significant, yet modest relationship between SMI [serious mental illness] (within 12 months) and violence, and a stronger relationship between SMI with substance use disorder and violence.”

In any case, I (we) did and do not question Thompson’s conclusion that severe mental illness is tied to a heightened risk of violence, although at least some of this link appears to be mediated by substance abuse. In fact, in a chapter that Thompson does not cite, I reviewed and discussed Teplin’s 1985 study in my 1994 book, Seeing both sides: Classic controversies in abnormal psychology:


There, I discussed some of the same sticky inferential issues regarding covariate control raised by Thompson (good lesson to blog posters – before criticizing authors, do your homework).

So, to summarize, in our book we argued that the claim that most mentally ill people are violent is a myth. Thompson then argued that he wished we’d discussed a different claim, namely, that mental illness isn’t associated with a heightened risk of violence, and he instead rebutted that second claim. In doing so, he reviewed evidence that actually supported our claim (!), but that rebutted his preferred myth, and then remarkably – in a bait-and-switch that would make advertisers envious - concluded that “So, the myth is not a myth: schizophrenic patients are roughly 5 times more violent than the general public” (!). The first part of that sentence is simply (and wildly) wrong: Our myth is indeed a myth.

In addition, to the logical errors (readers who enjoy collecting logical fallacies might also want to try to spot a striking example of the genetic fallacy in his blog post; see http://www.nizkor.org/features/fallacies/genetic-fallacy.html) in his column, Thompson seems to fall prey to the commonplace error of confusing relative with absolute risk ratios (see the following article for an excellent review of this and similar topics: http://pubman.mpdl.mpg.de/pubman/item/escidoc:2100705/component/escidoc:2100704/GG_Helping_2008.pdf). Specifically, the fact that schizophrenia is associated with a several times higher risk of violence does not imply, or come remotely close to implying, that most people with schizophrenia are violent. To do imply thus would be to fall prey to the related error of base rate neglect, as the base rate of schizophrenia in the general population is probably a bit under 1 percent.

Similarly, imagine that one were to maintain that (a) “Most people who smoke cigarettes do not get lung cancer” (certainly true, by the way… https://www.verywell.com/what-percentage-of-smokers-get-lung-cancer-2248868) and that Thompson (or someone else) were to take issue with this claim on the grounds that (b) cigarette smoking is associated with a heightened risk of lung cancer. Again, such an argument would constitute a glaring non sequitur. The data presented in b in no way refute a, because the base rate of lung cancer in the general population is relatively low on a statistical level (although of course extremely high from the perspective of tragic human and societal costs – alas, statistical formulas are heartless and do not care about such matters).

Finally, Thompson’s argument about invoking myths, following the late Buzz Hunt of intelligence research fame, strikes me as quite unconvincing (one hears such arguments against invoking myths from time to time, although I’ve never been persuaded that people who advance them actually believe such arguments when the rubber means the road in actual practice). If one can make a claim that X is true, as Thompson himself is doing here, one can also make an equally compelling claim that not X is false. To be sure, some “myths” contain a kernel of truth (hence, such myths fall on a continuum with accurate information, as we acknowledge in our book), but many are flatly wrong. The sun does not revolve around the earth; the brain does not serve as a radiator, as the great Aristotle believed; people do not use only 10 percent of their brain power; water does not retain a “memory” of molecules that are no longer present in it (as homeopathy proponents claim); and so on.

Thanks again for the opportunity to respond; I’m sorry that I found the quality of Thompson’s blog posting to be disappointing.

Wednesday 24 August 2016

More diversity illusions


Cover of the Race report 2016

It is unkind to mock the afflicted, but reasonable to ask them not to sneeze in our faces.

The Equality and Human Rights Commission in the UK has published a document entitled: Healing a divided Britain: the need for a comprehensive race equality strategy. Described by them as the biggest ever review into race inequality in Great Britain, providing a comprehensive analysis on whether our society lives up to its promise to be fair to all its citizens, the full report can be found here:


The Equality and Human Rights Commission does not regard itself as afflicted,  yet it is afflicted by significant bias: a tacit acceptance of a world view they believe to be inherently correct: that unless outcomes between ethnic groups are equal, that is prima face evidence of bias, also known as racial discrimination.

They state their objectives thus: The Equality and Human Rights Commission is calling on the UK, Scottish and Welsh Governments to address race inequality and discrimination experienced by people in Britain in a comprehensive and coordinated way.

They give details of inequality in education and learning; work, income and the economy; health and care; justice, security and the right to life; and the individual and society.

I started by looking at the references, and they mostly relate to government reports, trade union reports, a Guardian article and some books and working papers. There was nothing I could find from the peer-reviewed literature on scholastic attainment or intelligence, and nothing on genetics or the heritability of behavioural characteristics. This is a prime example of a confection which is lawyerly rather than scholarly. It is written by lawyers who ignore the world-as-it-is for the world-as-they-want-it-to-be; nay, the world as they command it to be.

At this point it might be wise just to turn away, because it is a weak publication, but to do so would make it seem that we have received the sneeze in our face without demur: that we had agreed that Britain is divided, and needs healing, and that new policies are required to make Britain more equal, and the underlying cause is racial discrimination by White British people.

The rejoinder is that Britain is doing its best to cope with large-scale unselective immigration, is paying the price with relative good grace, but cannot overcome genetic differences and deep-seated immigrant cultural norms. Furthermore, immigrants who are as bright or brighter than the locals do well, those who are less bright do worse, and the average intelligence of the country of origin is the best predictor of outcomes for immigrants to the UK.

Immigrant competences:


The suppression, or “ignoral” of these findings makes it possible to insult the White British majority with imputations of foul character, and to demand further taxation for race-based equality compensation schemes.

I will pick only a few areas for comment.

Here is a sample comment (page 5 bottom paragraph): It is indefensible that in 21st century Britain, Black workers with degrees earn over 23 per cent less on average than White workers with degrees; and if you are Black in England you are more than three times more likely to be a victim of murder and four times more likely to be stopped and searched by the police.

Degrees: University degrees in Britain are now obtainable from institutions of very different quality, ranging from the Russell Group who require very high exam results to institutions which require nothing more than good intentions. The latter weak institutions are numerous. Weaker students go to the weakest institutions, and gain less valuable qualifications. A good degree from a good university boosts income by about 60% over those without degrees. A bad university does not boost income to anything like that amount, if at all. The above percentages would be informative if the complication of heterogeneity of degree quality was acknowledged. Low scoring minorities are disproportionately represented in low achieving universities. Employers know all this, and many restrict recruitment to those universities who require high entry standards. A wise policy.

Victims of murder; and stop and search: What you will not find in this report is the descriptions victims give of their assailants. This would allow researchers to see whether Police are searching for suspects in line with the reported ethnicity of the assailant. Unless we know the race of the assailant the other statistics are misleading. If Black Britons are doing more of the crime, and offering more violence to those closest to them, other black males they compete with; they will be stopped and searched more often, and put in prison more often. This is a proportionate response, and would only be disproportionate if the data of the assailants were available. In the absence of UK data, here is some from the US. The overwhelming majority of black homicide victims (93 percent from 1980 to 2008) were killed by other black men.

Here is a study looking at criminal justice procedures, but controlling for IQ and previous criminality.


On education, the Commission report says (page 10, point 3): Ethnicity has been shown to impact a child’s educational attainment at GCSE level in England and Wales, and the GCSE equivalent (Standard Grade) in Scotland.6 Data in England shows that Gypsy/Roma, Irish Travellers and Black Caribbean pupils have the lowest attainment. In addition, ‘Is Britain Fairer?’ found that White boys receiving free school meals in England and Wales had the lowest educational attainment at the age of 16 in 2015.

“Ethnicity has been shown to impact…” is a curious construction. A more usual description would be “some ethnic groups get low examination scores”.

In criticism of Government initiatives they say (page 12, point 4, second paragraph): By addressing individual issues in a piecemeal way, without consideration of causes, drivers and levers for change, actions taken are unlikely to be effective in the long term or provide significant and sustainable change.

“Without consideration of causes” raises a wry smile, because they do not consider inherent behavioural differences at any stage. I think they would consider it morally wrong to do so. Blithely, they continue: For example, we will not make sustainable progress in reducing the ethnic minority employment gap unless we address the educational attainment gap.

Well, let us look at that. The US has tried to do that since the middle of the last century. Despite best and expensive efforts, the scholastic gaps at the end of schooling, though reduced, remain substantial, with commensurate impacts on later occupational achievements. Asians are ahead of Whites, who are ahead of Hispanics and Blacks. Using these data the proportions of the different racial groups who would be 2 sigma above the European white mean (and thus gain entrance to a highly selective university) are: Asians 4.78% , Whites 2.28%, Hispanics 0.82%, Blacks 0.47%.




For example, here is a bald statement which, in the Commission’s view, shows “inequality”:  Just 6% of Black school leavers attended a Russell Group university, compared with 12% of Mixed or Asian school leavers and 11% of White school leavers.

The predictions above would suggest that if entry were based on ability alone the rates of attendance would vary even more (this assume US results reasonable proxy for UK, which is close enough for illustrative purposes) . The main point is that differences in group level intelligence will have big impacts on participation at elite universities.

What they could have said is that since black school leavers have lower exam scores they go to less good universities. They could also have said that another genetic group, Asians, do better than Whites, so the big difference is between the scholastic levels of Blacks and Asians, not the fact of them being ethnic minorities in Britain. They could also have added “And this is in line with international scholastic ability, as shown in repeated PISA, TIMMS and other assessments.” The authors either intend to mislead, or have misled themselves so comprehensively that they cannot give a balanced account.

The Commission does not give tables or figures, perhaps because the greater scholastic abilities of the Chinese would be all too apparent. Instead there is a litany of percentages, from which it is difficult to draw any conclusions other than that minorities are hard done by. Often the use of percentages of percentages leads to confusion. Those seeking details are referred to other websites. It is the old “Supplementary Appendix 3” problem all over again.

Although the Commission is against racial stereotypes they cannot resist the temptation to use a new label “White boys FSM”. I wondered idly if this was white boys who, against the trend, were carrying out Female Sexual Mutilation, but it turns out to refer to Free School Meals: in other words, Poor Whites. It is now generally admitted that they are doing badly. Quite how this fits in with the narrative is unclear. The story seems to go like this:

White people are biased (hence the need for more anti-racial bias policies), and are prejudiced against other races. Formerly this bias was blatant, now it is subtle. Admittedly, they are not prejudiced against yellow skinned Chinese, and against brown-skinned Indians, but definitely against Black skinned Africans from the colonies, but not against Africans direct from Africa, who are doing well. But Whites are generally prejudiced anyway, so these exceptions need not concern us. Whites are even prejudiced against White people if they are poor because, I think, white people are racist and also class-ist, and hate foreigners and all poor people. Brexit has made them worse, because they now hate foreigners even more. Britain has had a spike of “hate crimes”. However, not given prominence in this report, Britain is still one of the world’s most favoured destinations for all immigrants. Nonetheless, more anti-racism policies are required.

You may need to help me here, because I have had to infer the argument from the tone of the report, and the contradictions are not discussed.

 One of their policy recommendations is interesting:

4. All governments improve the range and scope of the disaggregated ethnicity data available – including intersectional data – and ensure that ethnicity statistics and research findings inform their race equality strategies.

This is fine if they collect racial data on a broad range of issues, including the perceived race of assailants in reported crimes. It would be good if they could include data on racial differences in intelligence in their next report. None of this is likely to happen, but I am simply pointing out the obvious gaps in their arguments.

The Commission are entirely un-reflexive about immigration. It is presumed to be a good thing. They are secure in their opinions, and woe betide you if you do not share them. Criticising them may eventually get registered as a hate crime.

I know that I should track down every single thing they have left out, but it is August, the silly season, when people release silly stories to fill up the newspapers. Strangely, the sun is shining, so one should not get too agitated about biased, partial argumentation being used to strongly imply that racial differences must be the fault of one race, not the result of the different characteristics of all races. For example, the Olympic 100 metres final has been 100% West African for decades. This is not a black conspiracy, any more than East African dominance in long distance races is a conspiracy. It is a result of underlying genetics, plus some sponsorship.

Anyway, too much sunshine to push all this further.

I do not wish to speak ill of anyone, but I suspect this report was written by a lawyer.


Wednesday 17 August 2016

Philosophical Printers: a study in complexity



I make a habit of confessing my cognitive errors, as a warning to others, an illustration of my inevitable limitations, and the better to understand them. You may remember the difficulties I have getting myself to conferences, because of having to coordinate travel times, country zone time changes, car rental and return times,  to say nothing of getting the dates of travel to correspond to the early commencement of welcoming lectures the following day.

Last week I had a struggle with a printer which would not print. Printers used to be insensate objects which merely printed. They knew their place: slaves to the computer master. Now they have developed philosophical doubts about their purpose in life, and will not print unless handled with extreme tact. They can hold your work hostage, and know how to exploit Luddite bargaining power. They demand registration, your personal details, and the right to tell you in subsequent communications how they are developing as a product. That is the unkind interpretation. A more gentle view is that they are suffering from identity anxiety and polymorphous existential doubt: they do not know if they are scanners, copiers, glossy photograph booths or even, in their atavistic past, fax machines. One gains the impression they have been reading Goethe’s “The Sorrows of Young Werther”.

Newly bought, my printer seemed benign: a harmless helper without sharp edges. It arrived without a full set of printing cartridges, so it sulked at being asked to work without the required complement of colours, particularly the missing magenta and cyan (it was quite peeved about those) and there was a delay till those were delivered the following week. The proffered installation disc was rendered obsolete by my new laptop no longer catering for this out-of-date medium, so I downloaded the software and then the problems began.

The printer had its own set of priorities. It was willing to connect by Wifi, but demanded that any device have a username and a password. My router had a username and a collection of passwords, though not labelled as passwords, on the underside of the router. I tried a variety of these, but the printer rejected them petulantly. Eventually I found a new potential password on the edge of the router, and that partially mollified it. It printed a test page, and then nothing else. Documents it rejected, keeping them in the Documents Printing queue for ever. This is the printer equivalent of the Passive-Aggressive ploy.

After many attempts, I decided to resolve the high level of complexity by going for a very simple solution: a cable between printer and computer. After a week the cable arrived, and when attached the printer still sulked, printing nothing, other than a page saying it was connected, but had elected to go Offline. It remained Offline thereafter, whilst still occasionally moving the print head in a suggestive and reassuring manner.

There were two problems: the connectivity problem, whatever it was; and my mental model of the problem, which was probably wrong in one or several of its parts. For example, should the computer and printer talk to each other through the router, or to each other? When the printer asks for a password, is it the password to the router or the computer? Why doesn’t the direct cable work,  and why does the obviously connected printer say that it is “offline” and remain sulking offline? Why do the many instructions distinguish between different types of cable, USB, Internet, etc. but not explain why? Why do the advice forums say that you should re-install the software? Why do so many advice forum questions get answered by people who have surface plausibility and zero knowledge?  Why are the simple installation manuals so simple, and yet so complicated in their ambiguity? Why does my iPhone have no problem connecting with the printer, even though it has never met it before, let alone never having been introduced properly?

You may say, quite rightly, that I should have solved this problem simply and quickly. True. I know it has not happened to you. I also confess that I do not know how I solved it, or if I have solved it permanently. It printed something, and I am grateful, as any supplicant should be. One of my many downloaded connectivity solutions may have done the trick, or they may yet interact to thwart me once again. It is a fraught relationship, ripe for psychoanalytic interpretations.

I simply illustrate that even trivial problems contain some complexity, and understanding some of the complexity does not necessarily help, because some complex understanding is itself imperfect, and because some things have become more simple as a new solution leapfrogs an old problem. That is the joy of competence, as Bryan and Harter found in 1897, in their psychological studies of the emerging technology of telegraphy. Once operators understand the overall meaning of a message, the details of the Morse codes of individual letters can almost be ignored. Key presses give way to a higher grammar, with a commensurate increase in speed and power of communication. Similarly, we will soon let computers connect to each other with their own handshakes, troubling us no more with their quaint requirements, local customs and digital shibboleths.

Till that day, there will be the odd confusion, born of imperfect understanding, a discrepancy between model and reality, in which the latter always triumphs, and the printer refuses to print. May it not happen to you.


Thursday 11 August 2016

In a free state


The evening was English, overcast with a  brisk breeze, and the club garden seemed just warm enough for a drink, but not for a longer dinner. Then the wind abated, some blue sky appeared among the clouds, and we decided to risk it.

Almost at the end of our meal, with pink clouds glowing on a fine August evening, a new party came to dine, gathered round an older man with a face which seemed familiar from book covers. I was excited but also abashed, and wondered if it would be best just to leave him to his meal, uncertain that it was really him, and not wanting to intrude. I dawdled over a protracted coffee to consider what to do.

When it could no longer be delayed we rose to leave, saying goodbye to two separate tables of friends, and I gathered courage and approached his wife, saying: “Do you always get your meals interrupted by strangers paying their respects?” She was gracious and welcoming, saying it happened sometimes, asked my name, and told me hers, thus confirming I had recognized him correctly. His sons smiled in welcome. So I turned to him and all I could say was: “Thank you for writing”. He smiled warmly and thanked me, and I left, very emotional.

Authors have great power over us readers. We let them drive the car of our imagination, at speeds and on journeys of their choosing, and let them hold the keys for ever.

Thank you, Vidiadhar Surajprasad.

Wednesday 10 August 2016

Detecting schizophrenia myths


The British Psychological Society publishes a Research Digest, and the latest one lists:

10 of The Most Widely Believed Myths in Psychology


Here is Myth No 3:

3. Violent offenders usually have a diagnosis of mental illness
When people with mental health problems commit violent crimes, the media takes a disproportionate interest. No wonder that surveys show that most of the public believe that people with mental illness are inherently violent. In fact, as Scott Lilienfeld and his colleagues explain in the 50 Great Myths of Popular Psychology, the evidence suggests that at least 90 per cent of people with mental illness do not commit violent acts, and the overwhelming majority of violent offenders are not mentally ill. Some patients with specific conditions (such as command-based hallucinations "telling them" to commit acts) are at increased risk, but actual acts of violence are rare.
A telling meta-analysis from 2011 concluded that 35,000 high-risk patients with a diagnosis of schizophrenia would need to be permanently watched or incarcerated to prevent one killing of a stranger by a patient.

In my last post I looked at the relevant item written by Scott Lilienfeld and colleagues in 50 Great Myths of Popular Psychology and found that the first few references, in my view, were not the endorsements that Lilienfeld and colleagues imagined they were, and that they had not dealt with the proper epidemiological papers in sufficient detail. (He disagrees, and will be writing a reply in due course). I thought that the rough estimate of male schizophrenic patients being five times more violent than members of the public was probably right.

Looking back at the BPS Research Digest I realized that I had not dealt with the last claim made in that publication, that “35,000 high risk patients with a diagnosis of schizophrenia would need to be permanently watched or incarcerated to prevent one killing of a stranger by a patient.” Here is the relevant paper:

The Predictive Value of Risk Categorization in Schizophrenia

Matthew M. Large, Christopher J. Ryan, Swaran P. Singh, Michael B. Paton, and Olav B. Nielssen,


DOI: 10.3109/10673229.2011.549770

You might imagine from the BPS Research Digest that this paper shows that you cannot detect which schizophrenic patients will be violent. Not really. The paper discusses whether a screening test applied to all schizophrenic patients would be able to detect a series of harms which vary considerably in frequency. The least frequent is the killing of a stranger in any one year, and that provides the astounding “35,000 patients to prevent one death” headline. Other harms are much more frequent, and more easy to detect. In defence of the above authors, their calculations are based on two assumptions: that the actuarial risks they use are the correct ones, and that the screening methods deployed are the best available. There are also new findings, available since they wrote their paper, to which I refer at the end of the post.

The base rates used by Large and colleagues are annual risk figures, not risks for longer periods. They derived them from from the published literature up till 2010.

As to screening instruments, Large et al mention Psychopathy Checklist–Revised (a semi-structured interview and questionnaire to elicit psychopathic personality); Historical Clinical Risk Management (specifically assessing violence risk from a historical, current and future perspective); Manchester Self-Harm Rule (assessing self harm and suicide risk), and the SAD PERSONS scale (suicide risk). So only the second scale is directly concerned with estimating the risk of violence to others. It is a well-validated scale with an extensive bibliography.

violence risk assessment

Is it any good at predicting violence? In a reassuring move, they answer this question using Receiver Operating Characteristics statistics, which show the overall accuracy of the measurement approach. I first came across this concept in the papers on attention written by John Morton, who thus taught me about ROCs and d prime long ago. Thanks, John.

A given area represents the probability that a randomly chosen person who scores positive on the dependent measure (in this study, is actually violent) will fall above any given cut-off on the predictor measure, and that an actually non-violent person will score below the cut-off (Mossman & Somoza, 1991). Thus, an area of .75 means that there is a 75% chance that an actually violent person would score above the cut-off for violence on the predictor, and an actually non-violent person would score below the cut-off. AUC values of 0.70 may be considered moderate to large, and .75 and above may be considered large.

The reported AUC values are in the .60s, .70s and some .80s according to which samples are being studied. In fact, the instrument is a good approach, and worthy of being improved further.


Rather than screening methods, I think that Operational Research is better. For example, evaluate those patients with prior problems of violence when they request re-admission for treatment (since not all those who apply can be admitted for treatment). That group will have a higher probability of offending than patients who comply with their medication. An illustration of the possibilities for a Bayesian approach can be derived from the very tables of frequencies given by the authors below.

The Large et al. paper is a careful piece of work, and I like their general approach. They know that actuarial methods are better than clinical judgment. They create a hypothetical screening instrument, but admit that the base rates for different harms are very different, and decide to use the same cut-offs for all harms, which may reduce the detection of violence to others. ROC curves would have been a welcome addition for comparative purposes. However, they say that they have set sensitivity and specificity to 80% each, exceeding what is reported in the general literature for future violence.

Here is the crucial table:

Positive predictive value in schizophrenia

As you can see, assaults are committed by 1 in 7 untreated schizophrenics and by 1 in 10 treated schizophrenic patients per year. Treatment is only moderately effective in this regard. These are very high rates compared to the general public. Even with a relatively weak predictor, you need to monitor only 2 or 3 patients to possibly prevent an assault. This would be a highly effective intervention, and should receive more publicity. Large et al. seem to be doubtful about the value of screening in general, but their own data show it has utility for assaults.

For violent crime, monitoring 26 patients is required for a possible prevention of a violent crime. Again, this is manageable given resources.

Homicide in untreated patients happens, according to this table, at the very high rate of 1 in 600 schizophrenics. That compares with homicide rates in the UK of 0.9 per 100,000 persons and in the US of 3.9 per 100,000 persons. (US 4 times as murderous as the UK). So, the rate of homicide in non-schizophrenics in the UK is 1 in 111,111 and in the US is 1 in 25,641 persons. Therefore, an untreated schizophrenic person, using the estimates given in this paper, is apparently about 42 times more likely to murder someone than a US citizen, and 185 times more likely to murder someone than a UK citizen. Can these figures be correct? If so, this is a very dangerous category of person. An instrument with a positive predictive value of 0.66%  (extremely low) requires that 151 persons be monitored. This would be onerous, but would prevent a murder. It is an indicator of the level of risk to the population when patients do not take their medication.

A homicide committed by a treated patient (1 in 10,000) means that treated schizophrenics are apparently 11 times more dangerous than UK citizens and almost 4 times more dangerous than US citizens. In the US it requires 2500 patients being monitored, a high number, and the best estimate of how difficult it would be to prevent one person being murdered, assuming most patients comply with treatment. Monitoring for most patients would probably involve no more than chasing up non-attenders at follow-ups, and doing some random checks on compliance with medication. This would be worth studying, particularly now that monitoring in diabetes is being trialled using mobile phone apps, with good results.

Here are a few reflections. Risk estimates vary considerably, but all are raised for schizophrenics, particularly in the early untreated phase. By implication, a schizophrenic patient who does not comply with medication falls into a high risk category. It seems very worth-while to screen for assaults, violent crime and homicide, particularly in untreated or medication-refusing schizophrenics.

Being alarmed about schizophrenia is understandable, and wishing them to have treatment and comply with treatment is also comprehensible. Screening schizophrenic patients in the UK is even more valuable than in the US, because of the massively increased risks compared to the lower UK norms.

Please check all these above figures for me. I have difficulty believing these risk estimates because they are so high. Clearly, detection rates are poor when the base rates are low, a familiar problem in signal detection of all sorts. The BPS Research Digest headline of the rarity of a schizophrenic committing a “stranger murder” is somewhat misleading, because we do not immediately have an accessible figure for “stranger murder” committed by the general public. One has to go through Bureaux of Justice Statistics 2008 to find that 25.5% of male murder victims and 11.9% of female murder victims are murdered by a stranger. Since 90% of offenders are male, I would guess that stranger murders are about 24% of the US total. Thus, in the US context, a US schizophrenic murders strangers at the rate of 1 in 140,000 and a normal US citizen murders strangers at the rate of (1 in 106,838) if I have done the sums correctly. Something wrong, I suspect. I looked up the original reference for this estimate, from Nielssen et al 2009, who is also co-author of the main paper being considered here.

Schizo stranger murder


Neilssen et al. say: A fixed effects model was used to calculate a pooled proportion of stranger homicides of 9.0% of all homicides by people with psychosis.

A fixed effects model was therefore used to calculate a pooled rate of stranger homicides committed by offenders with psychosis of 1 in 14.3 million people per year (95% CI=1 in 18.9 million per year to 1 in 11.5 million per year, Z=8.278,P<0.001). If it is assumed that 0.5% of the population have schizophrenia, the annual risk of a stranger homicide by a person with schizophrenia can be estimated to be about 1 in 70 000 patients per annum. If the prevalence of schizophrenia-related psychosis is assumed to be 1%, the estimated risk of stranger homicide is lower, about 1 in 140 000 patients per annum.

This explains the estimate used in the 2011 detection paper. However, if events have been under-recorded in the published literature, then the rates could be somewhat higher. I see the rates as quite variable, typical of rare events. However (see Sariaslan et al. paper below) the meth0ds do not give us certainty that all instances have been captured. At this stage I will need to do further work to understand the “stranger murder” calculations.

Here is a more recent finding about the general dangerousness of schizophrenia in a full population sample of 3,232,010 from Sariaslan, Larsson and Fazel (2015):

schizophrenia and violent crime in swedenA Sariaslan1, H Larsson2 and S Fazel1

We observed that nearly one in four (23%) schizophrenia patients had ever been convicted of a violent crime, whereas the equivalent prevalence was 11% in patients diagnosed with bipolar disorder and 3% in controls.

So, I revise my earlier view that there is a fivefold increase in violence in schizophrenia. In this 3 million person population sample, schizophrenics are 7.7 times more violent, bipolars 3.7 times more violent. Some mental disorders lead to violence. It would be good to get more details on the types of violent criminal event, but from the public’s point of view, the picture is clear enough.


Time to do some further work on detection, and arrange services on the understanding that some prevention of violence is possible, and worth doing.

Sunday 7 August 2016

Myths dispelled, new myths propagated


It is a familiar journalistic trick to make a list of popular myths and then dispel them. One example is “50 Great Myths of Popular Psychology: Shattering widespread misconceptions about human behaviour” by Lilienfeld, Lynn, Ruscio and Beyerstein 2010. The authors are university psychologists of professorial rank, with active research careers, and their identification of myths and their attempted refutations are often entertaining. I am glad that some weak arguments have been exposed, and some popular beliefs revealed to have little research foundation, or to at least have foundations which can be questioned.

However, my view is that the “myths refuted” style of thinking is a prescription for poor science. In the late Earl “Buz” Hunt’s marvellous phrase, it leads to a type of argument which is “lawyerly” rather than “scholarly” because once you have identified something as a “myth” then you are honour-bound to destroy it, even if part of it is true.

So, I am going to look at one of Lilienfeld et al’s arguments from the point of view of argumentation and fact. Since the items are short, I will give their substantial arguments in full.

Myth 43 Most mentally ill are violent.

First, note the absolute nature of the myth: what is being dispelled is that a majority of mentally ill patients are violent. The myth is presented in exaggerated form, as a straw man, thus obscuring an important issue. What people fear is that mentally ill people are more violent than mentally well people, and on that basis it may be prudent to avoid them.

Second, the relevant question is “are mentally ill people more violent than mentally well people, and if so by how much?”

A third relevant question is whether the most-focussed-upon disorder within mental illness, schizophrenia, makes people more violent, and if so by how much? 

Prospective studies of mentally ill people will give the answer. An associated question, though not strictly relevant to the main issue, is whether violence can be predicted by concentrating on a particular sub-set of the mentally ill, or particular behavioural indicators. Here are the key arguments from Lilienfeld et al.

Yet commonplace public beliefs about mental illness and violence don’t square with the bulk of the research evidence (Applebaum, 2004; Teplin, 1985). Admittedly, most studies point to a modestly heightened risk of violence among people with severe mental illnesses, such as schizophrenia and bipolar disorder, once called manic depression (Monahan, 1992). Yet even this elevated risk appears limited to only a relatively small subset of people with these illnesses. For example, in most studies, people with paranoid delusions (such as the false belief of being pursued by the Central Intelligence Agency) and substance abuse disorders (Harris & Lurigio, 2007; Steadman et al., 1998; Swanson et al., 1996), but not other mentally ill people, are at heightened risk for violence. Indeed, in some recent studies, severely mentally ill patients without substance abuse disorders showed no higher risk for violence than other individuals (Elbogen & Johnson, 2009). Nevertheless, psychiatric patients who take their medication regularly aren’t at elevated risk for violence compared with members of the general population (Steadman et al., 1998). There’s also some evidence that patients with “command hallucinations”— hearing voices instructing a person to commit an act like a murder— are at heightened risk for violence. (Junginger & McGuire, 2001; McNiel, Eisner, & Binder, 2000).

Still, the best estimates suggest that 90% of more of people with serious mental illnesses, including schizophrenia, never commit violent acts (Hodgins et al., 1996). Moreover, severe mental illness probably accounts for only about 3-5% of all violent crimes (Monahan, 1996; Walsh, Buchanan, & Fahy, 2001). In fact, people with schizophrenia and other severe mental disorders are far more likely to be victims than perpetrators of violence (Teplin, McClelland, Abram, & Weiner, 2005), probably because their weakened mental capacity renders them vulnerable to attacks by others. Furthermore, most major mental disorders, including major depression and anxiety disorders (such as phobias and obsessive-compulsive disorder), aren’t associated with a heightened risk of physical aggression.

Notice the form of the argument: a modestly heightened risk of violence among people with severe mental illnesses, such as schizophrenia and bipolar disorder. Fine, but it would have been better to state the risk compared to the normal population risk. This way the reader has to chase up the references. Start with Applebaum 2004. In fact this is a letter by J.S. Appelbaum. On examination it has no actual data to contribute. Quoting it as a reference is feeble, and immediately calls into question the notion that these authors are in a position to challenge “myths”.

Next, Teplin (1985). This is better, in that it contains data, and arguments about the link between mental illness and criminality. Teplin rejects the obvious approach, which is to take a set of people who have been given in-patient treatment for mental illness, and then see what their subsequent criminal records are, compared with the general population. I will call this the epidemiological paradigm: “study psychiatric patients and see whether they commit crimes later”. Instead she plumps for what is in my view a more complex and error-prone approach, which is to look at the behaviour of those who come into contact with the Police at criminal incidents, as part of a larger study on how the Police dealt with mentally ill people. Psychologists went with the police to disturbances, and rated citizen behaviour for signs of :

severe mental disorder, e.g., confusion/disorientation; withdrawal/unresponsiveness; paranoid, inappropriate, or bizarre speech and/or behavior; and self-destructive behaviors

As you can see, this is back to front. Instead of following a defined population, people known to have had mental illness, and then seeing how many of them commit criminal acts, this study looks at people who are interacting with the police (criminals, victims, bystanders) and tries to determine how many of them are in the midst of an episode of severe mental disturbance. The defined base population has been lost, and things have to be estimated from the population of “people of interest to the Police”, a far weaker method.

Here are the results, as presented in the paper:


Mental disorder and criminality

Describing this table, Teplin says this of those with severe mental disorder:

they were somewhat more likely, 35.3 versus 23.4% for non-mentally-disordered persons, to be suspects.

Somewhat more likely? 50% more likely. However, this is to misunderstand the table, which has nothing to say about how many psychiatric patients go on to commit crimes. It does not tell us whether the 85 severely disturbed persons are what you would expect among 2122 citizens. At 4% of the population this is quite a high rate of severe mental illness. For example, lifetime prevalence of  all psychotic disorders was judged to be 3.06% in 2007 in US populations. So, there may be a 33% over-representation of disturbed persons in this police sample. Frankly, the method is too weak for any firm conclusions, though even this weak method shows mentally ill persons to be over-represented as suspects of crimes, by 50%.

Teplin goes on to say:

mentally disordered citizens made up less than 5% of the persons who were involved with the police.

She gives the median rate of severe disorder as 1.7% so her observed rate is 3 times higher! She gets out of this by saying that the study area included some neighbourhoods with more mental patients, ignoring that that is what you would expect from psychotic downward social drift. Her task was to find out whether mentally ill people committed more crimes than average, not to find out where they lived.

Teplin keeps making the irrelevant point (irrelevant to the issue of whether mentally ill people commit more crimes) that most interactions of mentally ill people with the police are not about violent crimes. This obscures the finding of interest. Here is her second table, about the suspects themselves:



Mental disorder and type of crime

On her own figures, severely mentally disordered persons are 3 times more likely to have been the perpetrators of violent personal crime. However, since we do not have proper base rates on a defined population of mentally ill persons all we can say is that they appear to be higher, and it could be a chance finding. This study does not confirm that severely mentally disordered persons are no different from the normal population, though the sparse evidence points the other way.

In an attempt to make their case, Lilienfeld et al say:

Still, the best estimates suggest that 90% or more of people with serious mental illnesses, including schizophrenia, never commit violent acts (Hodgins et al., 1996).

That means that 10% do commit violent acts. What is the rate for those without serious mental illness? In the US in 2012 total violent crime was 387 per 100,000 which is a rate of 0.39%. On that reasoning the violence rate for the seriously mentally ill is 10/0.39 = 25 times higher. I think that this is probably an error, even though it is presented as a reassurance. A comment as silly as that was reportedly made by a former president of the Royal College of Psychiatry, though certainly not the present one.

A final comment from Teplin:

In an exhaustive review of the pertinent research literature, Monahan and Steadman concluded that if a number of socio-demographic factors known to be related to crime are taken into account e.g., race, age, and prior criminality, the relationship between mental disorder and criminality substantially diminishes.

I think that sums up the problem with these arguments: someone with mental disorder and a criminal record is given statistical “time off” for previous crimes, even though they may have been caused by mental disorder. “Taking into account” becomes a form of washing them away from the causal nexus. In my view, this is very silly, and Monahan, Steadman and Teplin must know that, or at least acknowledge that “taking into account” covariates can be a misleading procedure. Before “correcting” any finding, the assumptions behind the correction should be made explicit.

Lilienfeld et al continue: Moreover, severe mental illness probably accounts for only about 3-5% of all violent crimes (Monahan, 1996; Walsh, Buchanan, & Fahy, 2001). In fact, people with schizophrenia and other severe mental disorders are far more likely to be victims than perpetrators of violence (Teplin, McClelland, Abram, & Weiner, 2005), crimes (Monahan, 1996; Walsh, Buchanan, & Fahy, 2001). In fact, people with schizophrenia and other severe mental disorders are far more likely to be victims than perpetrators of violence (Teplin, McClelland, Abram, & Weiner, 2005),

So, do the other studies follow the epidemiological paradigm?The other papers mentioned by Lilienfeld et al. are interesting in that they identify the patients who are most likely to be violent: patients with command delusions who have not taken their medication. How violent are they? The rates are not given. 

Are there other relevant papers Lilienfeld et al. could have mentioned, particularly those done the right way round: identifying the mentally ill and then checking whether they are involved in violent acts?

For a fuller review see:


For more recent research, admittedly after the publication of the myths book, but making the same point:


So, the myth is not a myth: schizophrenic patients are roughly 5 times more violent than the general public. It is up to you whether you want to avoid them on that basis.

The rest of the Lilienfeld “myths” book has many myth-busting chapters with which I either agree or want to agree. However, I haven’t checked the references, so I cannot be sure whether these myths have been dispelled or propagated.

Monday 1 August 2016

The g nexus: Detterman detail


It has been very hard getting intelligence research results out to the public. All researchers, whatever their subject, tend encounter the same problem: they get excited about findings which the general public don’t find interesting, or don’t understand. Researchers complain that their subject is ignored, misrepresented and on the odd occasion that it is noticed at all, the treatment is superficial, the selection of supposed experts absurd, the main points mangled

Intelligence research has experienced all of the above, and more. Researchers get thrown out of the university posts, harassed with legal challenges and cumbersome investigations, and systematically avoided by publishers and research funding bodies whenever the results are considered unwelcome.

One researcher who was given the full “ignoral” was Chris Brand. Before being consigned to the outer darkness he wrote a chapter in 1987 entitled “The importance of general intelligence” and Arthur Jensen quoted his summary in Table 9.3 (page 300) of his book “The g factor” in 1998. Brand had also written a book entitled “The g factor” in 1996, which was to have been published by Wiley, who then chose to “de-publish” it.

The list of positive and negative correlations with intelligence that Doug Detterman showed in his lecture is his updated version of Brand’s 1987 original, and as intelligence research flourishes it will continue to be updated. Any slide has to be a summary, but some of the entries were a bit too terse. Doug has provided some very brief explanations of the items readers found confusing.

“Abnormal is normal” means that the same genes that cause normal processes are the same ones involved in what we call "abnormal" conditions.  In other words, abnormality is simply the extremes of normality.

“World conservatism” refers to a score on a test of world conservatism and can be read as conservative in the common political sense of the word.

“Drop out rates” are, of course, negatively correlated with intelligence.  If  I said that, then what I should have said is that educational completion levels are positively correlated with intelligence.

For the background detail, Doug has provided a chapter he is working on.