Saturday 25 June 2011

The National Children’s Study: a view from across the pond

The National Institutes of Health, the major government funding agency for medical research in the US, is in trouble. Although a threat of a serious funding cut was largely averted, funding for 2011 is nevertheless $260 million less than for 2010. It is predicted that only 1 in 6 grant applications will be funded this year, the lowest level yet. Such low funding levels are not only disappointing for researchers with good ideas, they also are inefficient, because scientists end up spending more time writing unsuccessful grant proposals than in doing research.

So it gets interesting to see what does get funded. NIH has a great website where you can find out how they spend their money. I came across it when doing a study on the amount of funding going to research on different neurodevelopmental disorders. I revisited the site recently to look in more detail at the kinds of study that get funded in this general area, and I had a bit of a surprise. There were huge sums of money going to something called the National Children’s Study. I added them up and they came to around $500 million.
More Googling revealed that the National Children’s Study is a longitudinal cohort study that plans to follow 100,000 children from the prenatal period to 21 years of age,  to investigate environmental and genetic influences on health and development. There’s no doubt that big longitudinal cohort studies have provided invaluable data in the past; they can provide samples that are large enough to detect small effects, and if the sample is representative of the population, they can be used to estimate prevalence of different conditions. They are essentially correlational, though, and therefore less conclusive than experimental studies for establishing cause and effect. 
The cost of such a large-scale enterprise is bound to be high, but just how high is acceptable? The National Children’s Study website indicated that the spend to date was actually higher than the figure I had from the NIH source - $608 million, with the bulk of the expenditure, $553 million, since 2007 when the Implementation phase of the project began. And it is clear that the future spend will be higher still, as the cost of following up children and assessing their development is taken into account. The total cost was estimated in 2004 to be $2.7 billion. 
The background to the study is clearly explained in an article by Nigel Paneth. During the Clinton presidency The Children's Act of 2000 was passed, instructing the NIH and other federal agencies to “. . . plan, develop, and implement a prospective cohort study, from birth to adulthood, to . . . incorporate behavioral, emotional, educational, and contextual consequences to enable a complete assessment of the physical, chemical, biologic, and psychosocial environmental influences on children's well-being . . . .” As Paneth drily pointed out, this directive was not accompanied by any funding during the 7 years of the Bush administration, so efforts were expended instead on planning and piloting the study, in anticipation of future funding. However, too much planning can be a bad thing, and “as a ship lying in port too long becomes encrusted with barnacles, so the National Children's Study became weighted down with a myriad of measures and instruments” (Paneth, 2010). But then, salvation. Under the Obama administration, the study returned to the political agenda, and $200 million per annum was allocated for 2009-2010. But problems remain. In an article entitled “Saving the National Children’s Study”, David Savitz and Roberta Ness argued that the study is in trouble. There has been debate about the best methods of recruitment of participants, and about the organizational structure of the project. “There are multiple advisory groups representing the lead federal agencies, the extramural research community; an executive committee consisting of selected Center investigators; and oversight by the National Children’s Study Program Office, the Director of NICHD, and the Director of NIH—yet how the major decisions affecting the future of the National Children’s Study will be made remains unclear.” (Savitz & Ness, 2010). Despite the huge sums of money allocated to the project, there is still uncertainty about methods. “To decide where the study should be done and who is capable of doing it, one has to know what the study will entail. The lack of a clear protocol has led to an extended, expensive, and not-fully informative pilot phase.” (ibid).
Of course you can’t know in advance exactly how longitudinal cohort data will be used: one of the benefits of previous cohort studies has been that they provide a resource to allow investigators to test their ideas against existing data, rather than having to collect a new dataset. For example, in the UK, the Avon Longitudinal Study of Parents and Children, which started in the 1990s, is now being used to investigate hypotheses about genetic and environmental influences on behaviour.  When the project started, it wasn't possible to anticipate the developments in genetics, but the prescience of the researchers in gathering DNA samples mean that the data are invaluable, especially since a wealth of environmental and psychological measures are also available. Nevertheless, even allowing for such uncertainty the vagueness of the Hypotheses section of the National Children’s Study website, is worrying:
Working hypotheses developed by the multiple teams of scientists have been summarized as an efficient and dynamic reference of the current questions to be addressed by the Study. The current list of hypotheses continues to evolve. It is expected that some hypotheses still being refined will be added and, over the long course of the Study, new questions will emerge and be added to the Study and some of the hypotheses included may become outdated and discarded. However, there is consensus among scientists planning the Study that as a group, the hypotheses alone can neither convey the true breadth of the Study nor completely guide the planning and design of the Study. To further define the full scope and topics of the Study, priority outcomes along with priority exposures are identified as health areas for the Study. The priority exposures and outcomes serve as an organizing framework for the Study hypotheses.” This sounds awfully like “We’re going to gather loads of data and hope that it pans out.”
Well, you might say, any nationally-based study of 100,000 families is bound to be complicated and expensive. But, as Savitz and Ness point out, there are examples of European studies that are comparable in size and scope, yet do not appear to have got so mired in difficulties. They draw attention to the Danish National Birth Cohort and the Mothers and Babies study in Norway, each of which is studying over 100,000 children enrolled in utero. I have not yet been able to find figures for the Danish study*, but the Norwegian study costed around $45 million over the 10 year recruitment period. Funding was announced this year for a new British Birth Cohort study that will follow 90,000 children from birth at a cost of £33.5 million. I’m sure that all these European cohort studies will ultimately cost far more than initially planned - such studies have a tendency to gobble up funds as the children grow older and the difficulties of maintaining a follow-up increase. The articles written on the Danish and Norwegian studies emphasise the logistic and ethical difficulties of doing studies of this kind, and the uncertainty about the future. But even so, the costs of the National Children’s Study seem set to exceed those of European studies by an order of magnitude.
I’m struck, as Savitz and Ness were, by the organizational differences between the European and US studies. Commenting on the Scandinavian cohort studies, they point out “These studies, each led by a very small group of investigators, took an early and consistent approach to their specifications and made tough decisions about what aspects to promote and what to curtail, based on a shared, explicit vision for the study.” This is very different from the huge committee-based structure in charge of the National Children’s Study. Committee-based decision-making even extends in the US study to writing up work for publication: “The Publications Committee will identify topics, set priorities, and facilitate the preparation of primary Study publications. The Committee will define the general scope and content of primary publications and assemble Writing Teams to prepare those publications. The Committee will define the charge of these Teams, monitor their progress, and review the manuscripts to ensure adherence to the charge given to the Teams, before the manuscript is submitted for publication in a peer-reviewed journal.” As an approach to scientific communication, this approach seems designed to kill stone dead any enthusiasm or creativity in the researchers. I think if NIH wants to get good science for its money, it would be better off taking an analogy of the scientist as an enthusiastic market gardener rather than factory farmer.
But a more fundamental question is whether the outcomes of this study will justify the enormous cost, or whether the money might be better spent elsewhere. A key question is how far we need another longitudinal cohort study, or whether the existing European studies will be able to answer many of the questions that the researchers are interested in. It’s not easy doing a cohort study anywhere, but it seems particularly tricky in a country that does not have a national health service. There are many areas funded by NIH where an individual investigator could do useful work with a grant of around half a million dollars. The funds spent to date on the National Children’s Study could have funded over 1200 such research projects. Viewed from this side of the pond it seems that, with the best of intentions, US funding for health research is simultaneously starving out the best of its scientists while feeding a bloated monster.

* Update on 2nd July 2011. I've now got a couple of estimates from researchers involved in the Danish study and both indicate a total cost since 1994 of less than $US 20 million. That is around 1/10th the cost of the National Children's Study just for one year, 2010. The Danish researchers make the point that the study benefits from the existence of national registers that minimize costs of case-finding, and that already contain much relevant information.

Kaiser, J. (2004). NIH Launches Controversial Long-Term Study of 100,000 U.S. Kids Science, 306 (5703), 1883-1883 DOI: 10.1126/science.306.5703.1883
Magnus, P., Irgens, L., Haug, K., Nystad, W., Skjaerven, R., Stoltenberg, C., & , . (2006). Cohort profile: The Norwegian Mother and Child Cohort Study (MoBa) International Journal of Epidemiology, 35 (5), 1146-1150 DOI: 10.1093/ije/dyl170
Olsen, J., Melbye, M., Olsen, S., Sorensen, T., Aaby, P., Nybo Andersen, A., Taxbol, D., Hansen, K., Juhl, M., Schow, T., Sorensen, H., Andresen, J., Mortensen, E., Wind Olesen, A., & Sondergaard, C. (2001). The Danish National Birth Cohort - its background, structure and aim Scandinavian Journal of Public Health, 29 (4), 300-307 DOI: 10.1177/14034948010290040201
Paneth, N. (2010). Saving the National Children╩╝s Study From Its Saviors Epidemiology, 21 (5), 602-604 DOI: 10.1097/EDE.0b013e3181ea5f89
Savitz, D., & Ness, R. (2010). Saving the National Children╩╝s Study Epidemiology, 21 (5), 598-601 DOI: 10.1097/EDE.0b013e3181e942cc

Tuesday 21 June 2011

Autism diagnosis and hyper-systemizing parents: Nottingham vs. Eindhoven

Family life in Eindhoven?

With an insensitivity one might have expected from a Sun headline writer, the New Scientist summarised the findings from a recent study thus: Childhood autism spikes in geek heartlands.  The bottom line is that autism is diagnosed more than twice as often in Eindhoven, a city in the Netherlands where many are employed in the IT industry, than in two other cities of comparable size, Haarlem and Utrecht. The explanation favoured by the researchers is that autism is characterised by a ‘systemizing’ style of thinking, which is part of normal human variation, and which is common in those in technical and mathematical occupations. Cities with a strong IT presence will attract an unusually high number of high-systemizer residents, and these people will be more likely to have offspring with autism. This is, then, a genetic explanation, and it’s given some plausibility by a large body of research demonstrating that parents of children with autism are more likely than other parents to show indicators of mild autistic-like characteristics, the so-called ‘broad autism phenotype’. (Although, on an admittedly quick search, the evidence that parents of children with autism tend to be high systemizers seems rather weak: mild symptoms of social and communicative problems seem a more pronounced feature of the broad autism phenotype than hyper-systemizing behaviours (Bishop et al., 2004; Wheelwright et al., 2010; Windham et al., 2010)).

As the researchers themselves note, the hyper-systemizing account is not the only possible explanation for their result. Crucially, the study relied on counting diagnoses from school records, rather than screening the population in a standard fashion. Although it’s not hard to recognise a case of classic Kanner autism, there’s far more disagreement about diagnoses for children with milder symptoms. As I argued on a Guardian blog, unless we have clear objective criteria for diagnosis, it’s hard to compare one prevalence rate with another. The different numbers could just reflect local expertise, policy or practice in diagnosing autism.

One limitation of the published study is that the researchers are quoted as saying that their study was prompted by anecdotal reports that autism was abnormally common in Eindhoven. While it is worth checking out if the anecdote is accurate, this makes Eindhoven less than ideal for testing the hyper-systemizing hypothesis, as it potentially capitalises on a chance blip. What would be better would be a study with clear a priori predictions, based solely on levels of IT industries in different cities. In theory, it should be possible to do this using publicly-available data from the UK published by the Department of Children, Schools and Families. This dataset has the advantage of being comprehensive, unlike the school report data from the Eindhoven study, which relied on schools providing the data (- the response rate was 75% for Eindhoven but only 50% for Haarlem and 46% for Utrecht). A recent report by Lindsay (2011)  presented some data from this UK database on numbers of children with Special Educational Needs categorised as having Speech, Language and Communication Needs (SLCN) or Autistic Spectrum Disorder (ASD). The numbers of children with these labels varied massively from place to place, as shown in Table 1.

Table 1: Percentages of children with SEN diagnosed with SLCN or ASD; 
Data from local authorities with the highest or lowest % of either diagnosis

I have no idea whether the number of IT experts is higher in Nottingham than Leeds, but it’s noteworthy that places where you might expect high levels of hyper-systemizing, such as the university towns of Oxford and Cambridge, don’t feature among the places with very high rates of ASD diagnoses. Lindsay also points out that “the two neighbouring authorities of Nottingham and Nottinghamshire have almost identical prevalence rates for SLCN and ASD, despite one being a large city, and the other a shire county” (p. 143). Clearly, my intuitions are no substitute for real data, and it’s also important to note that the data in Table 1 are not population prevalence figures, but instead are proportions of children who have already been identified as having Special Educational Needs. One would need to use the frequencies of ASD in the population as a whole to test the hypothesis of a correlation between level of IT industry and rates of autism. The data are available, and this might seem like a nice project for someone to do, except for a major problem. As Lindsay emphasised, it is impossible to conclude from the UK data whether prevalence really do vary across the country, because the definitions of disorders are inconsistent from one area to another. It’s possible that criteria for autism are more standardised in the Netherlands than in the UK, but the UK data make me suspect that it’s just not possible to draw meaningful conclusions about prevalence from data based on educational records.


Bishop, D. V. M., Maybery, M., Maley, A., Wong, D., Hill, W., & Hallmayer, J. (2004). Using self-report to identify the broad phenotype in parents of children with autistic spectrum disorders: a study using the Autism-Spectrum Quotient. Journal of Child Psychology and Psychiatry, 45, 1431-1436.

Lindsay, G. (2011). The collection and analysis of data on children with speech, language and communication needs: The challenge to education and health services. Child Language Teaching & Therapy, 27(2), 135-150.

Roelfsema, M., Hoekstra, R., Allison, C., Wheelwright, S., Brayne, C., Matthews, F., & Baron-Cohen, S. (2011). Are Autism Spectrum Conditions More Prevalent in an Information-Technology Region? A School-Based Study of Three Regions in the Netherlands Journal of Autism and Developmental Disorders DOI: 10.1007/s10803-011-1302-1 
Wheelwright, S., Auyeung, B., Allison, C., & Baron-Cohen, S. (2010). Defining the broader, medium and narrow autism phenotype among parents using the Autism Spectrum Quotient (AQ). Molecular Autism, 1(1), 1-9. 
Windham, G. C., Fessel, K., & Grether, J. K. (2009). Autism spectrum disorders in relation to parental occupation in technical fields. Autism Research, 2(4), 183-191.

Sunday 19 June 2011

Should we ever fight lies with lies?

Photo from verdammelt's photostream
So here’s my moral dilemma. The pangolin is a threatened species. It is a peaceful creature which has evolved an impressive strategy of self-defence; it curls up into a scaly ball. The effectiveness of this strategy against predators can be seen here.  Unfortunately, this doesn’t work when your opponent has hands, and can just pick you up. Pangolin numbers are being depleted by poachers who sell them for Chinese medicine. Pangolin scales are known as Chuan Shan Jia and are reckoned to be effective for such purposes as “expelling wind-damp from the channels”. They are also recommended for treatment of cancer, see for instance, this website.

Although sterling work is being done by conservation agencies, time seems to be running out. So I wondered about taking a different tack. As Ben Goldacre has famously remarked, the Daily Mail has a mission to classify inanimate objects into those that give you cancer vs those that cause cancer, so how about getting pangolin scales on the former list? This idea has been germinating in my brain for a while, but there are two major obstacles. First, I have a reputation for truthfulness, which I’d be sorry to lose, and second, how would I start such a rumour?

Well, a recent post by Jo Brodie gave me an answer for part 2. Jo noted that if you upload photos to Flickr, they get picked up by people doing Google searches. Her photos of ads for weight-loss tea, accompanied by her comments that there’s no evidence for effectiveness, have actually appeared on websites advertising the stuff.

But Jo’s posts don’t involve any dishonesty. I want to spread a rumour about Chuan Shan Jia that is completely fictitious. I had some misgivings, but I did it. I opened a Flickr account and made a graph. I took a figure from a powerpoint presentation on cancer rates and doctored it to be about pangolin scales. You can see it here. [No you can't, link removed. See P.S. below!]. Whether the sort of people who use Chinese medicines will actually ever find this seems unlikely, but it seemed worth a shot. If anyone would be willing to translate the slide into Chinese, I'd be most grateful.

I had a moment wondering whether I’d get sued by purveyors of Chinese medicines, but I don't think I need to worry, since the use of endangered species in medicines is illegal.

So what about my reputation? Well I am hoping that I can preserve it by declaring here that you can rely on me to be honest in my science communications on all topics except use of medicines based on endangered species. If my pangolin hoax works, you can anticipate a Flickr posting on rhino horn next.

My husband is very disapproving and thinks it is never right to fight lies with lies. What do others think?

P.S. 20th June 2011
Well, the verdict in the comments is clear, and my husband is now feeling extremely smug. To go down this route is not just unethical but also foolish. So I won't. I had hoped that by declaring myself as a hoaxer at the same time as initiating a hoax, I could defend my integrity, but I see that this would be a dangerous assumption, and the unintended consequences could be serious.
Thanks to Jo, I've discovered that my inexperience with Flickr means that my hoax graph wasn't copyable anyway.
But I'm glad I posted this: very interesting and helpful to get the reactions, and Jo has given me ideas for pursuing my vendetta against pangolin medicines without any dishonesty. I also inspired a great set of tweets from @Artvanderley that demonstrate another approach: implant the idea there could be bad effects of pangolin medicine without claiming it: e.g. "40 years ago no one believed that ingesting pangolins was linked with cancer, excessive weight gain and genital atrophy". I suspect that some of you may think that even this may be sailing too close to the wind, but it's clearly far less heinous than inventing data.
I'm left feeling just a bit disappointed, as I have always been very well-behaved and law-abiding and was rather enjoying my brief dalliance with hoaxing. But yes, the issues are serious, and so I will climb wearily back up on the moral high ground.

Tuesday 14 June 2011

A gentle introduction to Twitter for the apprehensive academic

If I tell people I’m on Twitter, I tend to get one of three reactions:
a) Isn’t it all about what Lady Gaga had for breakfast?
b) How do you find the time?
c) You?!!! (Implication: Twitter is for hip juveniles rather than fossilised academics)

This is unfortunate, because Twitter is a valuable resource for academics. If you’re allowing  inaccurate stereotypes to deter you, you’re missing out.
First of all, you have to understand what Twitter is. It’s totally different from email, and more like a news broadcast. People all over the world are continually emitting tweets (very short messages) any of which can be viewed by anyone. You select what you want to attend to. There are two ways of doing this. The default method is to ‘follow’ particular people or organisations who tweet. Their tweets then appear in your timeline, which appears as a scrolling list when you open your Twitter page. The other method is to search for tweets that include a particular word: for instance, if you type ‘neuroscience’ into the search box at the top of the page, you’ll see all the tweets in the twitterverse that include that word, starting with the most recent.
If you want news about Lady Gaga, there’s plenty out there. But if you want information of a different kind, you can follow organisations such as the Royal Society, the Wellcome Trust, Guardian Science, the New York Times, Nature, etc. etc. Most scientific organisations, newspapers, and science journals are on Twitter, and by following them you have an up-to-date news stream about their activities.
It's OK to be a purely passive user of Twitter, just following people who interest you. In the circles I move in, a high proportion of tweets are messages pointing to a weblink, which may be a newspaper or journal article or a blog. This is where Twitter is such a useful resource for the academic: if you follow those who share your academic interests, they will point you to interesting stuff. When I first joined up I was impressed to find that within the first few days, I’d been directed to two new papers in my field that were very relevant to my work and that I hadn’t known about.
Many people remain as passive users, but you’ll get much more out of Twitter if you use it actively and emit your own tweets. Written an interesting paper? Starting up a blog? Twitter is a great way of informing people, but there’s a catch: you need to have followers, a topic I discuss more below.

How do I get started?
You can Google to find plenty of good guides to the mechanics of tweeting. See, for instance:
However, most of these are directed towards people who do have a keen interest in Lady Gaga’s breakfast, or who wish to use Twitter for business purposes. The suggestions here are to complement the ‘how to do it’ guides with advice geared towards academic users.
Signing up is dead easy: just follow the instructions at
You need a username. Keep it fairly short and avoid numbers or underlines: you want others to be able to remember it and type it easily. You can be anonymous if you wish, but I’d not recommend it: you are more likely to have interesting interactions with others if they know who you are. A brief description of what you do and what your interests are will help kindred spirits discover you. You get the chance to select your avatar, a little picture that appears alongside your tweets. It’s a good idea to have something other than the default picture of an egg - if you don’t want a photo of yourself, you can pick something symbolic, but aim for something to give yourself a distinctive presence. If you want, though, you can start with the egg and change it later.

How do I decide who to follow?
I started out by following my old friend and colleague Sophie Scott, or @sophiescott as she is known on Twitter. We have similar interests and a similar sense of humour, and so the first thing I did was to see who she was following. You can check out someone’s followers by clicking on their username at the top of a tweet. You’ll see their profile on the right hand side, with an indication of how many followers they have, and who is following them. Further clicking lets you see who these people are, and read their recent tweets. So it’s easy to get an idea of whether you’d like to see their tweets on a regular basis: if yes, a single click allows you to follow them.
The people I follow divide mostly into (a) organisations/public media, such as those mentioned above; (b) academics who work in areas that interest me; (c) journalists and bloggers. Although I have friendly relationships with many of those I follow, I don’t use Twitter as a means of keeping up with friends - it’s too public and the short message format is not useful for that.
I suggest you start out by just identifying a few people that look interesting to follow, and see whether you enjoy the Twitter experience. My recommendation would be to keep the number of people you follow restricted to no more than around 100. Many people follow far more than this, but I like my twitterstream to move at a reasonably sedate pace.
Getting fed up with tweets from someone you’re following? You can just unfollow them. They don’t get a message about this, so you can do it without embarrassment.

Active tweeting and attracting followers
You can have most fun with Twitter if you tweet yourself. For the beginner, there is a major problem: if you emit a tweet, the only people who will see it are your followers, and at the outset you have no followers. You may have something very amusing to say, or a really interesting paper just published, but it’s like standing at the top of a cliff and shouting into the wind. To get started, it helps to either be well-known, or to have tweeting friends. You can look for friends and colleagues by clicking on the ‘who to follow’ button, and if you find they have a Twitter presence, send them an email with your username to let them know you are there. With luck they’ll follow you, and tell others about your presence. It’s only worth doing this, though, if they are active Twitter users with followers: lots of people sign up but never use Twitter.
You may also drum up followers by following others. This is where it is important not to be too secretive: if I get a new follower, I’ll see their name and the brief bio that comes along with it, and if they look interesting, I may check out what they’ve been tweeting to see if I want to follow them. Twitter etiquette does not require that you follow someone just because they follow you, but following someone is a way of indicating your presence to them.
Another way to draw your tweeting to people’s attention is to use hashtags in your tweets. These act as keywords and are just words with the hash sign attached at the front, such as #neuroscience or #psychology. People who are searching on these topics will find your tweets and may decide to follow you.
If you are sending interesting tweets, the message will spread around the twittersphere and you will gradually get a following. You may wonder how on earth you are supposed to generate those interesting tweets that will persuade people to follow you. You don’t always have to. You can act as a transmitter for other people’s interesting tweets, by using the Retweet button below the tweet. This will just resend the tweet to your followers, preceded by RT and your username.  
You should not despair if at first you don’t have many followers. Although it’s true that a famous name will attract followers in droves, there are plenty of people who aren’t famous, but who have hundreds or even thousands of followers just because they give good value. And at the end of the day, you should not get too hung up on follower numbers. The charm of Twitter is that it lets you reach out to communicate with people all over the globe whom you might otherwise never encounter: a handful of like-minded people who appreciate your tweets is more important than a horde of followers who seldom read your messages.
What about spam?
Many newbies are worried that they will get followed by odd people. That certainly will happen. But the nice thing is that it has no impact on you. I attracted lots of provocatively dressed pouting followers when I started out. But they can follow me as much as they like; they won’t affect my stream of incoming tweets unless I follow them back. Various unsavoury characters will appear as followers for a day or two and then drop away. They hope that by following you you’ll take notice of them and buy whatever product they are purveying, but you just ignore them and they go away. Twitter discourages users who simply see it as a marketing opportunity, and is set up so you can readily report people for spam if it looks like they are doing that, but mostly the only bad thing that happens is that you have a fleeting moment of excitement at gaining a new follower, only to be disappointed to find they are someone who sees you as a potential client.
The one place where you may get more intrusive spam is if you press the @Mentions button at the top of the screen. [Update: in later versions of Twitter this is called @Connect]. Now, instead of your usual stream of tweets by followers, you will see just those tweets that mention you by username, and these will not necessarily be by your followers. In general, if you get people mentioning you in tweets, this creates a warm glow that others are interested in your tweets, but there are people who will try to exploit this, and so you may find tweets that mention your name in a tweet to lure you in to clicking a link to their website. In my experience, these are very rare, and when they occur they are usually easy to spot: if you click on their username, you’ll see they’ve sent the same message to many others. You should just report them for spam.
What should I tweet?
Quite simply, tweet to others things you think will interest them. Looking at tweets by others should give you an idea of what makes for a good tweet. Some very famous people are hopeless tweeters, because they just describe the mundane details of their life. What I actually want are either amusing observations, or useful information. Some people use Twitter to record their stream of consciousness. Unless you’re James Joyce, this is very dull for everyone else, and just makes you look egotistical. Come to think of it, James Joyce would have been a disaster on Twitter.
When you start tweeting, the 140 character limit seems impossible, but you learn by experience.  If you want to include a link to a website in a tweet, you will almost certainly need to shorten it. There are various programs for doing this, e.g.
As mentioned above, if I read a tweet by someone that I think will interest my followers, I'll retweet it. That’s a single click operation, one of the options given below each tweet. Retweeting is what makes Twitter such an effective communication medium: if an interesting message is retweeted by several people with many followers, who in turn retweet to their followers, it can rapidly spread all over the world.
If you’ve read this far, you’ll start to appreciate that tweets are a kind of currency; your status on Twitter is tied up with the extent to which you emit popular tweets. It is therefore as important to acknowledge the source of a good tweet as it is to reference an idea in a scientific article. If you use the retweet function, this happens automatically: the retweeted tweet will be marked RT and will show both your name and that of the originator. What’s a definite no-no is to copy someone else’s tweet and resend it without acknowledging the source. Etiquette is less clear on the extent to which one should send a tweet to thank others who promote one’s tweets. This may seem polite, but if over-used, it can descend into a rather irritating form of self-promotion -  in effect you are publicly drawing attention to the fact that others liked your tweet. A lot depends on how it’s done, but there’s a narrow line between being seen as courteous, and coming across as a self-congratulatory dick.
How much should you tweet? I’m more likely to unfollow someone who tweets too much rather than too little. Anyone who keeps repeating the same self-promotional message is quickly dropped from my list. If you have a blog or article you want to promote, it’s reasonable to plug it a few times on different days and at different times of day, to make sure the message gets out, but you’ll turn people off if you overdo it. If it’s interesting enough, your followers will do the work of promotion for you, by retweeting. On the other hand, there's not much point in following someone who tweets less than once a week, unless their tweets are really something special.
Remember, Twitter is totally public. I can search for someone’s name and then look at all their tweets. I would therefore strongly advise against tweeting anything at all that you would not want your friends and colleagues to see, or that could be deemed defamatory. There is a Delete option you can use on tweets that you come to regret, but by the time you select it, your tweet could have been sent all around the world.
How do I find the time?
It’s a big mistake to think you have to read every tweet that appears in your stream. I just turn to Twitter when I need distraction or entertainment. So you really don’t need to spend very long on Twitter unless you want to. The difficulty is that what’s happening on Twitter is often more interesting that what’s happening in other areas of life, and it can become quite addictive for that reason. I usually resolve not to look at Twitter during the working day, especially if I have a paper to write or an analysis to run. But sometimes the resolve faulters.
I hope that’s enough to persuade you to give Twitter a try. Happy tweeting!

P.S. 18th June 2011
The Reply Option
I’ve been asked about the Reply option that appears when you hover over a tweet.
You can use this even if you aren’t following someone. Your message will automatically be prefaced by the username of the person you reply to. It won’t appear in their regular stream of tweets, but they will see it if they look under @Mentions. This can be a way of getting yourself involved in a discussion, even if you don’t have followers. And if the discussion develops well, the person you reply to may decide to follow you.
Twitter won’t show your Reply to your followers unless they are also following the person you reply to. Complicated, huh? The reason is that otherwise your followers will be treated to just one side of a personal conversation. But if you think your reply may be of interest to a broader audience, you can make it visible to all by just putting a full stop in front of the @ at the start of the Reply. (In fact, any character will do, but . is traditional. The rule is that provided tweets don’t start with @ your followers will see them).

P.P.S.28th October 2011
I've written a short piece about blogging for academics that may interest readers of this post.

26.3.12. PPPS. A final warning! There is one bad thing that can happen to you on Twitter, that won’t happen to you provided you are forewarned. These are what I call “drink me” tweets, after the tempting bottles that Alice was confronted with in Wonderland. They say things like “Someone is writing bad things about you”, or “Here’s a funny photo of your”, or “Find out your Twitter ranking”, and point you to a link. Usually they come from someone you have never heard of (and so will only be seen on the @Connect timeline). NEVER CLICK ON THE LINK. These are attempts to take over your Twitter account (they may actually ask you to click to confirm this is OK. It isn’t). Once your account is compromised, the fun begins. Your account will be used to send messages to all your followers, typically advertising something, and often encouraging them to respond in a way that will compromise their account. Deeply embarrassing. For the same reason, if you get a message from someone you ARE following which looks weird - e.g. “My Twitter ranking is 5.5, find out yours!”, don’t go near it.
There is useful advice on what to do if your account is compromised here

And a note on Favourites. I didn’t mention Favourites in my original post, as it was pretty self-evident how they worked. However, the good folks at Twitter then updated the interface, and it’s no longer obvious. If you want to keep a Tweet - perhaps to read later, or to refer back to - select the option to Favourite it. To see your Favourites, you need to go to My Profile.
On other Twitter platforms, such as Tweetdeck, you can have a column showing Favourites without additional clicking.

7th Nov 2012. PPPPS. Email and Twitter
Unfortunately, whenever Twitter produces an upgrade, it seems to make the experience worse. One relatively recent change has been the irritating expansion of Email Notifications. The last thing I want is Twitter stuff intruding into my email, but when you first sign up to Twitter (or upgrade to latest version) these notifications are On by default. It's easy to switch them off though.
From top menu bar select Me, then Edit your Profile.
Then from menu bar on left, Email Notifications will allow you to select or deselect specific kinds of email that Twitter can send you.

Saturday 11 June 2011

Brain scans show that.....

I was set off today by a report that “fMRI scans prove music is more emotionally stimulating if you listen with your eyes closed”. What’s wrong with that? Well, if I want to know if music is more stimulating with the eyes closed, the sensible experiment would be to ask people to report their emotions while listening to music with eyes open or shut. A brain scan gives you a 3D representation of how regional blood flow changes over time when a person performs a particular task or experiences a specific mental state: it does not measure emotional stimulation.
It seems every week we have another claim that brain scans have shown something about our cognitive or emotional characteristics. To see how accurate this perception was,  I tried Googling the phrase “brain scans show” or “brain scans prove”. Within minutes I had a long list of hits, most of which made unsupportable claims.
Neuroscientists have for many years been interested in documenting how brain activity in different brain regions correlates with mental activity or mental states. For some mental states, we know enough to say that when we do cognitive operation X, or experience emotion Y, brain region Z is usually activated. But what we can’t say is that if we see activation in Z, this is equivalent to cognitive operation X or emotion Y. For a start, most  mental activities involve a whole network of brain regions acting together, and there can also be substantial variation from one person to another in brain structure and function. But more seriously, there is a logical problem at the heart of this argument. Consider not the brain but the skin. When I am anxious, I tend to sweat. If you observe that I am sweating, you may suspect that I am anxious. But I could be sweating for other reasons: I may have been exercising, the room may be hot, or I may have been eating chillis or taking medication. You could test your theory that I am anxious by asking me. Suppose I deny being anxious: would you believe me? Maybe not: perhaps I am really anxious but reluctant to admit it. But you’d be foolish to regard sweating as a more reliable indicator of my emotion than my verbal report unless there was good reason to believe I might be lying. You’d be even more foolish to assume that I’m not anxious if I’m not sweating. Sweating is a correlate of anxiety (and an unreliable one at that): it is not the same thing as anxiety.
But it seems that in the media, there is a belief that activity from a brain scan provides a direct measure of  specific mental states; not only that, the brain scan is seen as a more reliable indicator than traditional behavioural measures. You may find it instructive to consider the claims made in the links below, and then consider how they’d look if you took away the phrase ‘brain scans show”. In general, you can see that the brain scan adds nothing: you could test each of these notions with behavioural data. It’s also worth considering how you’d react if a behavioural study and a brain scan study obtained opposing results. A prudent response would be to trust the behaviour. I've got nothing against fMRI studies as a method for studying brain-behaviour relationships: my criticism is for those who interpret the brain activity as if it was the behaviour. As Neuroskeptic put it in a post last year, “Brain scans prove that the brain does stuff”.This was the one Google result that cheered me up. Here are the others:

Brain scans show how a disorder leads individuals to perceive themselves as ugly
Brain Scans Prove Branding Strength

Brain scans prove ‘that loving feeling’ truly is addictive
Acupuncture reduces pain perception, brain scans show

Brain Scans Show Women Feel More Pain Than Men

Sunday 5 June 2011

Review of "How Genes Influence Behavior "

by Jonathan Flint, Ralph J. Greenspan and Kenneth S. Kendler

Oxford University Press, 2010

One would expect an introductory textbook by these three eminent scientists to be good, and it is. It covers behavioural and molecular methods, using examples from the authors’ own work in psychiatric genetics. The field is put in historical context, with the false turnings explored, as well as the major breakthroughs. The authors take us on a journey through family, adoption and twin studies, on to  the history of molecular psychiatric genetics, starting with the early, and usually unreplicated, linkage studies, through genome wide association studies and animal models using rodents and drosophila. Concepts are explained with a minimum of technical and statistical detail, making the book accessible to a wide readership. What sets this book apart from others on this topic are the personal and often witty accounts of doing research. Kendler’s anecdote of a family study of schizophrenia in Ireland is particularly engaging, including an encounter with an elderly grandmother who expressed indignation that he should be driving around all over the place interviewing people to find out if mental illness ran in families. “’Why everybody knows that! Take the O’Donnells for example. They are as mad as can be and it goes back generations.” (p. 12). One also gets a sense of the hard graft involved in doing research in this area: Kendler estimated that his study took 18 person years just to gather the data, and evidently required persistence and a willingness to work heroically antisocial hours by the research team.  
Chapter 5, on genetic association analysis, is a sobering account of the ‘quagmire’ resulting from thousands of pounds worth of funding directed to finding genes that might explain psychiatric conditions. I’d recommend this chapter to those looking for (nongenetic) associations between individual differences in behaviour and brain markers: I see close analogies between the problems of molecular genetics and those found in neuroscience, where the sheer quantity of data makes it easy for spurious associations to be mistaken for real effects. So far neuroscientists have barely confronted this issue, though it’s pleasing to see Kriegeskorte et al (2009) making a start.

Anyone contemplating a research career in molecular genetics would do well to read the paragraph on p 34-35 debunking popular assumptions about life in the laboratory: “A combination of intelligence and technology applied to taxing but interesting biological problems leads to ground-breaking discoveries that could cure disease or change our understanding of the universe.” Well, no. The work is dull, repetitive, and usually unsuccessful. And if after years of work you do find something interesting, your laboratory head will ask: “Why is this finding wrong?”. In fact, I suspect only really good laboratory heads do that - the bad ones rush gleefully into print, which is why the field is littered with nonreplicable findings. But as Flint et al point out, you need to ask it because it’s embarrassing to publishing something that is wrong, your peers will do their best to find flaws in the work, and, most importantly, “it’s only by publishing findings that are robust to any possible criticism that we’ll make any progress”. A refreshingly old-fashioned take on the scientific process that is all-too-often forgotten in the current climate where we’re all encouraged to publish as much as we possibly can, and where a dramatic but non-replicable result may get you two papers in a high-impact journal: one for the original finding, and the other for the failure to replicate.

Kriegeskorte N, Simmons WK, Bellgowan PS, & Baker CI (2009). Circular analysis in systems neuroscience: the dangers of double dipping. Nature neuroscience, 12 (5), 535-40 PMID: 19396166

Friday 3 June 2011

Is poor parenting really to blame for children's school problems?

It’s tough being a parent of a child with developmental problems. You coo over your baby, anticipating how they’ll grow into a cheerful and contented individual, only to find that the older they get, the more the problems become apparent. Your child may be late starting to talk, and then communicate with difficulty, may struggle with reading, be clumsy, hopeless at maths, or poor at making friends. In many cases, several of these problems occur together. Gradually, as life throws more challenges at him or her, your charming toddler turns into an anxious and unhappy child, reluctant to go to school, fearing failure at every turn, and attracting ridicule from other children.

What you really don’t need at this point is to be told that it’s all your fault. Yet there is a band of self-appointed ‘experts’ who seem determined to keep this message alive, and to add guilt to the other emotions that parents experience. I know of mothers who think that their child’s problems are due to their having an occasional drink during pregnancy, going out to work when their child was young, or allowing them to eat food with e-numbers. And the media absolutely lap it up. Berating feckless and inadequate parents is a popular sport that attracts readers who can wallow in a sense of complacency that their child isn’t like that.

The latest in this line of commentators is Sally Goddard Blythe, who is described as Director of the Institute for Neuro-Physiological Psychology and child development expert. In her new book ‘The Genius of Natural Childhood’ she claims that many children are failing to develop vital physical and communication skills needed to start school, because of a lack of interaction between parents and children during the early years.
According to the Daily Telegraph “growing numbers of parents are turning to electronic toys and automatic baby walkers or rocking chairs to keep toddlers occupied. But this often impairs children’s natural development, meaning many are wrongly labelled as suffering behavioural problems when they start compulsory education”. The book notes that singing lullabies and using music gets a baby ready for language at a young age, while reading fairy tales can teach children moral behaviour and empathy.  Finally, “many early developmental problems could be overcome with old fashioned one-on-one interaction between parents and children.” The account in the Daily Mail was even more stark: Poor parenting means half of five-year-olds are unfit to start school shouted the headline, with the text continuing, “Children are failing to develop vital physical and communication skills after being robbed of interaction with mothers and fathers during the early years.” Interestingly, in an interview yesterday on BBC Radio 5 Live, Blythe seemed to change tack and back off from blaming parental communication, focusing more on apparently alarming rates of motor skill deficits in primary school children. Nevertheless, links to the Mail and Telegraph articles feature prominently on her blog.
So what’s the basis of Blythe’s claims? Let’s look first of all at her credentials, as given on her website. Director of an Institute sounds impressive, but less so when it turns out to be a private self-funding institution. She has letters after her name, including MSc and FRSA. I couldn’t readily find what subject the MSc is in, but FRSA is Fellow of the Royal Society of Arts. It’s not a particular mark of distinction: you can apply to become a fellow on their website,  provided you are willing to pay £150 per year and can give evidence that you “share or demonstrate a commitment to positive social change in professional, civic or personal life”. She also describes herself as a “freelance consultant in neuro-developmental education”. That fancy-sounding language gives no validation - anyone can set themselves up as a consultant. It only carries force if it is a title awarded by an organisation with some credentials, such as a university or a hospital.
Never mind. Maybe she has good data to back up the statements about parenting. Well, I've not yet had a chance to look at the book, so perhaps I will be pleasantly surprised, but I doubt it. I read widely on this topic and I know of no credible scientific support for the claims. I searched Harzing’s Publish or Perish for articles by Sally Goddard Blythe. Since 1998 she has had three papers in peer-reviewed journals, with a total of 41 citations between them. None of the papers involves research on effects of parenting. Two books were also listed, but since these aren’t peer reviewed, they wouldn’t normally be regarded as evidence of scholarly contribution, unless supported by significant peer-reviewed work. There was also a piece in an journal with the bold title Medical Veritas (The Journal of Medical Truth). I’d not come across this one before, and it’s not featured on Web of Knowledge, but suffice it to say that a high proportion of articles are concerned with vaccination, and Dr Andrew Wakefield is an editor. This paper also had no data on parenting.
One reason why this story hit such a nerve with me is that I’ve researched causes of children's language and communication difficulties for many years, and, contrary to popular opinion, parental behaviour does not seem to play an important role. There are converging lines of evidence that can be considered. The weakest kind is correlational: if you find that aspects of parenting correlate with the child’s language development, you can’t be sure of a causal link; there could be a third variable that simultaneously influences both the parenting and the child’s language. Nevertheless, this approach does allow you to assess the plausibility of potential causal routes. A recent epidemiological study looked at factors associated with late language emergence (LLE) in toddlers in Australia, and concluded “Risk for LLE at 24 months was not associated with particular strata of parental educational levels, socioeconomic resources, parental mental health, parenting practices, or family functioning. Significant predictors included familial history of LLE, male gender, and early neurobiological growth.” This study concluded that neurobiological factors and genes were important in determining which children had language difficulties. A more powerful design is the twin study, because it allows some dissociation of genetic and environmental influences. If environmental factors are important, then you’d expect two twins growing up together to resemble one another. However, twins share genes as well as environments. More specifically, we can distinguish identical twins, who are genetically the same, and non-identical twins, who are as similar as regular brothers or sisters. One of the first things to strike me when studying twin pairs where at least one twin had language difficulties was how different two twins from the same family could be - provided they were genetically non-identical. This impression was borne out by statistical analyses of twin data, which repeatedly find that the similarity of language abilities between two twins growing up together is strongly influenced by their genetic relationship (see Bishop, 2006). A further piece of evidence is obtained by turning the question on its head: if children require a certain amount of talking and singing to develop normal language, what happens to children raised in unusual environments where they have little exposure to such stimulation? Some hearing children of deaf parents grow up in circumstances where they are exposed to little intelligible speech except from a TV. When such children were first studied, everyone predicted they’d have serious language difficulties. Instead, the studies demonstrated just how robust language development usually is: most hearing children of congenitally deaf parents learn to talk quite normally (Schiff-Myers, 1998). I would not want to claim that parental behaviour is unimportant. Interactions between genes and environments make their effects difficult to disentangle (see earlier blogpost). I'd cheerfully go along with encouraging parents to sing to their children and read books. But it's simply not realistic to attribute children's school difficulties to lack of interaction with parents.
Mrs Goddard Blythe is entitled to her views. My concern is with the blurring of the distinction between opinion and evidence. When a view about effects of parenting is widely promulgated on national media, and is expressed by someone who is described as a consultant in neuro-developmental education and Director of an Institute, the natural assumption is made that (a) they are speaking from a position of authority, and (b) they have some hard evidence. In this case, neither appears to be true. 

Bishop, D. V. M. (2006). What causes specific language impairment in children? Current Directions in Psychological Science, 15, 217-221.

Schiff-Myers, N. (1988). Hearing children of deaf parents. In D. Bishop & K. Mogford (Eds.), Language development in exceptional circumstances (pp. 47-61). Edinburgh: Churchill Livingstone.

Zubrick, Stephen R., Taylor, Catherine L., Rice, Mabel L., & Slegers, David W. (2007). Late language emergence at 24 months: An epidemiological study of prevalence, predictors, and covariates Journal of Speech, Language and Hearing Research DOI: 10.1044/1092-4388(2007/106)