KneesUp
Guru
Your example isn't an unbiased sample if you select 9 from one side of the mean and 1 from the other, because in the general population there are the same number on each side.
Are most people thick sheep ?
- Thread starter clid61
- Start date
Page may contain affiliate links. Please see terms for details.
Even if you took 10 people at random off the street?
I know the size of the sample influences the distribution more evenly across the curve, but surely the chances of getting exactly 50% of the sample above, or below, the average must be pretty slim?
In the real world, might it not end up with, for example, slightly less than half the population being below average IQ?
I've no real expertise in statistical analysis - I just find it an interesting diversion.
classic33
Leg End Member
I keep a spare, does that count?
One of my favourite stats.
KneesUp
Guru
If you select lots of groups of 10 people at random (easier said than done) you will, on average, get 5 above average and 5 below average per group of 10 and it will get closer and close to a 50:50 split the more groups of 10 you randomly select - but if you just select 10 people, it's pretty unlikely it will be a representative sample of the population as a whole.
KneesUp
Guru
The number of legs is not a normal (bell-curve) distribution, hence the mean, median and mode are not the same, so you need to define what you mean by average [ / pedant ]
Yes I agree, it will get closer and closer to a 50:50 split; but will it ever get to an exact equilibrium? This could only ever happen if the sample size was an even number, and even then the probabilty of the split being exactly 50:50 would not be assured.
