This statistic has been recently widely reported in the media, as well as by concerned charities (though one charity, I CAN, appears to have removed their press release from their website).
Can this really be true? Does it mean children's language is getting worse? Is it because of the 10% who were reported to have a TV in their room before the age of 1?
Well... probably not.
There are various serious problems with the Yougov poll. Firstly, although this wouldn't necessarily weight it in the direction of finding too many speech and language problems, it's an email poll. The researchers claim to have a 35-50% response rate and, although they don't say this, I'm assuming they told people before they started the poll that it was about children's language (well, they'd have to tell them it was about children under 7 as they only published responses from parents of children that age). So, from the start, parents who think their child might have language problems are biased to respond.
Secondly, the questionnaire asks: How old was your child when they said their first word?
The children being asked about are, at this point, aged up to 7 years. 71% of the children asked about were over 3. So the vast majority of these children said their first word more than a year ago.
It is very difficult to get parents to give accurate information about their child's concurrent language (at the same time you are asking). The best way to do it is to give them a checklist from which they can recognise words that their child understands or says. People are much better at recognition than recall. And they are much better at concurrent report than past report. But this survey uses recall (it asks the parents to tell them, without prompts, the first word their child said) and it uses past report.
Now, some researchers have successfully used past report. For example, a team working on children with autism and looking back at their milestones used prompts like "can you remember where you were staying, what you were doing, around your child's first birthday/first Christmas. Now, can you remember what kinds of play and toys your child liked then, and what they were doing. Can you remember if your chlid could do A, B, C at that time?". That study used home video of events like birthdays and Christmas to verify some of the things parents reported their children could do, but still couldn't verify them all.
Finally, the Yougov poll relies on parents' judgment about what constitutes "experiencing difficulty learning to talk or understanding speech". Parents reported that mainly they decided this by comparing their child to another child the same age or to their other children. Children's speech and language development varies hugely within a wide normal range. To know your child is in the bottom 5% of children learning to talk (usually the cut-off for clinical difficulties) you'd have to compare them to 19 other children who would all have to be better than your child. It's probably true that slightly more than 5% of children have difficulties learning to talk (probably because there are many reasons a child can have difficulties) but even so, parents do not usually have 19 other children to compare their child to (either "another", implying one, or siblings!).
Parents' judgments about whether their child is experiencing difficulties are hugely confounded by expectations and knowledge about typical language development. Parents expect boys to have difficulty learning to speak - and more parents reported their boys did have difficulty. 31% did report that they knew their child wasn't reaching milestones. Sadly, milestones reported in cheery helpful baby books and on cheery helpful baby websites are usually average. Surprise, surprise, half of children are below average. Being below average isn't "experiencing difficulty".
And parents don't all know very much about children's language development. Current theories suggest that providing a correct model when children make mistakes in language is the best way to help them. Thankfully most parents do know this. But 14% thought the best response was to correct them - research shows this rarely works. Even more worryingly, and flying in the face of evidence that shows children flourish when bilingual, and do better in a wide range of cognitive and language skills, most parents thought that when parents' first language wasn't English, they should still speak English to their child at home if they live in the UK. One wonders how many Welsh-speaking parents they asked. Thankfully far fewer of the parents who didn't speak English as a first language thought using English at home was a good strategy.
So we've seen that parents whose child has a problem are more likely to answer this poll, most parents were recalling their child's language a really long time ago, probably inaccurately, and a worryingly large proportion of parents failed to give an answer that would support their child's language development when asked about best practice in the home.
In fact, the actual ages reported for first words - apart from a few children whose parents said they have a disability - are almost all within the range of "don't panic". Just 4% of children spoke their first words, as recalled by parents, past 24 months (and a few of the 11% whose first words were 18-24 months may have been in the "let's have a look at it" category). And nearly 69% reported their child put two words together before 18 months (in fact, the middle of the age range for this is probably about 21 months - so, erm, talkative middle class kids whose parents have email, much?)
But one part of this survey rings a little true - some parents who thought their child had a speech problem felt unsupported. But again, perhaps given the figures this isn't a cause for concern. Only 23% thought their child didn't get any help - so, as 17% thought their child had a difficulty, 13% of parents thought their child had a difficulty but did receive some help. Since really only about 5-7% of children actually have a difficulty, this is pretty good - possibly another 8% of children are receiving help for a difficulty they don't have!
*Gernsbacher, M. A., Sauer, E. A., Geye, H. M., Schweigert, E. K., & Hill Goldsmith, H. (2008). Infant and toddler oral- and manual-motor skills predict later speech fluency in autism. Journal of Child Psychology & Psychiatry & Allied Disciplines, 49(1), 43-50.
(Note: if you fancy downloading the raw data, it's here - an excel file)