didactic said:
Its inaccurate because for every survey you have to take all the other factors. Where was the question asked? Are all those children City fans? Who is favorite player to their parents? You need a much broader number of children and also the demographic matters. The sample size matters too.
That doesn't make it inaccurate, it makes it a less than perfect survey, but it was a straw poll if anything. We're not talking a highly complex survey here with millions of pounds spent on it.
As for the specific questions you asked.
1 - In a classroom in or around Preston.
2 - Unlikely given it's an entire class.
3 - Why does that matter? The reason why someone likes a player (own choice/parental influence) isn't really a factor here, it's a simple "pick a player" question.
4 - Obviously, for an all inclusive, fully rounded survey you need a far larger group of participants from all sexes, creeds, nationalities etc. But, as previously mentioned, this wasn't an experiment, or really even a full survey, it was a straw poll.