Please stop telling me how smart your toddler is

I was chatting to another mum with a child a few months older than mine when she started telling me how tough it was having a gifted daughter. ‘She gets bored with toys so quickly! And we keep having to buy things for children several years older than her, just to keep her stimulated!’

‘Mmm hmm?’ I am now playing a game with myself whereby I try and make the most noncommital noise possible without sounding like a hand blender.

‘We’ve been in touch with Mensa. They’re very excited about her! They say they’ll help us when she gets to school. She’s already having problems at nursery, the teachers just don’t understand how much stimulation she needs!’

The child in question was about 2 and a half. My game with myself worked all too well and so for the next few minutes she told me about this child in detail. Frankly, she sounded like she might be the new Messiah.

I was kind of astonished. This mum had always seemed to be quite a reasonable, sensitive person, but here she was telling me how amazing it was that her daughter made eye contact at three weeks and rolled over at seven. It was all quite rude to be honest, because there is only one possible response – (‘Oh really? Gosh.’) – and there was every chance that I was going to go home and lock my child in the cupboard with the shape sorter until he, too, could tell the difference between a square and a rectangle. Or at least down a bottle of wine and tell him bitterly that I wished I was Baby Messiah’s mum instead.

So the number one reason I don’t want you to tell me how smart your toddler is? It’s fucking rude. I got 11 A-grades at GCSE. Just thought you wanted to know. Oh wait, no you didn’t.

Reason number 2 I want you to shut up? (I don’t mean you, dear reader. I’m sure you would never do this.) You’re talking out of your arse. It’s pretty much impossible to work out how smart a child will be based on what they’re like before the age of three. I did a brief systematic review of the literature (methods below) and, basically, the age of reaching major childhood milestones is next to useless in predicting cognitive abilities later in life (except in particular groups – children born premature, or who suffered early brain damage), and while late walking or talking are important ways of identifying kids who might need extra help, there’s nothing to suggest that early walking or talking are a sign of being the next Einstein. One British study showed that children who learn to walk a month earlier than their peers will be, on average, 0.3 of an IQ point more intelligent, and 0.8% more of the children who spoke at 12 months rather than 13 stayed in education beyond the age of 16. Wow. Better start working on that Cambridge University application now. A Finnish study found that age of standing and walking ‘accounted for 7% of the variance in final school marks’ (which means that if all the children were running a race and the last was 100m behind when the first finished, then children who learnt to walk early started with a 7m headstart). Even careful laboratory tests in infancy could only predict 12% of the variance of 11-year old IQ. Another Finnish study showed that adults who had learned to walk later performed slightly worse on one out of a number of tests of memory and cognitive function. Studies that did find a strong association tended to include children with developmental delays, who tend to reach milestones later and have lower IQ – but this tells us nothing about whether reaching them early gives a *higher* IQ. One study explicitly demonstrated that any correlation is due to slower development in children with a subsequent IQ below 85.

Tl;dr – if your child is an early walker or talker, it doesn’t mean shit about their intelligence.

Reason number 3: the multiple comparisons problem. This is a really difficult problem pretty much all scientists have to contend with at some point. It basically means the more things you look at, the more likely one is going to be unusual and extreme JUST BY CHANCE. Say you are trying to work out if a pair of dice are weighted. If they are, they’re very likely to roll a double six. You throw them, and get a double six. This is unusual enough (it happens 1/36 of the time – less than the magic (but arbitrary) ‘5% level’ that’s considered statistically significant). You can have a pretty sure idea that the dice are loaded. But now, someone gives you 100 pairs of dice and asks you if *any* of them are loaded. You roll them all, and get 3 double sixes. But this time, you really can’t be sure that those three are loaded – if you take a 1 in 36 chance a hundred times, you would expect JUST BY CHANCE to get three rare events like this. So you can’t really say that they’re loaded at all.

The simplest way that scientists solve this problem is called ‘Bonferroni Correction’. If you’re only testing one thing, you might say that something that happens 5% of the time is unusual enough to be interesting. If you’re testing more than one thing, you just divide that percentage by the number of things you’re testing. So if you’re testing ten things, you need something you’d only expect 0.5% of the time to happen before it caught your attention.

How does this relate to toddlers? Just think of how much shit babies and toddlers do! They roll, sit, crawl, stand and walk. They smile, coo, babble, speak and string sentences together. They start to recognise friends and family, their toys, their neighbourhoods, animals, letters, numbers and shapes. They play with a hundred different sort of puzzle toys every week, learning to leave the pieces just where Mummy steps of bed with bare feet. Frankly, they do so much stuff, it would be astonishing if most kids weren’t ahead at some stuff and behind at others. Case in point: at nearly two, my son knows the difference between red, pink and orange, but still uses words one at a time. He can put individual red lentils in a tiny cup, but still prefers to hold my hand up and down steps. So when you’re boasting about how your daughter knows all the numbers up to ten, what I’m thinking is ‘Yeah yeah. But she doesn’t know the names of all the people who live In The Night Garden, and can she tell you where they sit in the Ninky-Nonk?’. And given there must be at least 20 scales we could measure our toddler’s achievement on (whisper that last bit – we don’t want to give the Education Secretary any ideas) I’d want your child to be the top 1 in 400 before I think she’s statistically significantly smart.

Reason number 4: Say your child is incredibly bright and advanced. What exactly do you think that means? How can they possibly live up to the dizzying height they reached when they learnt to say ‘bye-bye’ at ten months? Let’s say your kid is averagely special – say, top 5%? You’ll probably notice they are bright. You might even like to tell other parents just how bright they are. They’re likely to come top of their class and be the smartest person in pretty much every bar they go into. They’ll have an IQ of about 125, and with an IQ like that, they could be….a schoolteacher, or an accountant. Or maybe a doctor or lawyer. Which is really good. Good for them! I mean, these are great professions, really worthwhile. At least, that’s what you say to other parents, because those are best futures you can imagine for their only moderately gifted offspring. I mean, when you tell me that she’s the first kid in her whole nursery school to be able to write her name, you don’t add ‘…and maybe one day she might even become a pharmacist!’

What you’re imagining for your child is a little better. Now maybe you’re right, maybe they are in the top 0.1%, with an IQ above 150 – in which case, the future is looking pretty rosy. They’ve got a very good chance of getting a PhD or being a high-status professional. But, no matter how smart they are, what they’re unlikely to become, which I think is what you’re predicting for them, is an Ender Wiggan or a River Tam. Child prodigies are rare enough, and there’s scant evidence that a high IQ in childhood is a sure-fire predictor of adult genius. If your child has an IQ in the top 10%, 5%, 1% they will likely get a degree, maybe even a doctorate – but most likely, they will spend their lives being a solid middle-class professional. Earning a good wage, in a prestigious job, and with the chance to have a nicer life than much of the population. Great! But nothing so remarkable you need to tell other people about it.

Reason number 5: It’s super exciting watching your child grow up and learn new stuff. It’s kind of astonishing watching them transform from (let’s face it) jizz into a walking, talking mini-human in less the time than it’s taken me to get round to putting up that picture we got on our honeymoon. Trouble is, your child is almost certainly not a genius and if you can only enjoy watching them grow if you think they’re doing it better than other kids, you’re going to end up constantly disappointed by them, and they probably won’t like you much by the time they’re 12.

* I searched the Medline database on 4/2/2014 using the terms (‘developmental milestone *’ ‘developmental milestones *’ ‘milestone *’ ‘milestones *’, where * was ‘IQ’ or ‘intelligence’). Papers looking at children with known developmental problems or born prematurely, or those in areas where malnutrition in childhood is common, were excluded. Papers not in English or for which full text was not available online were excluded (because this is a blogpost, not a fucking paper. I am NOT walking across town to the library).