Indian call centres & Irrelevant advert statistics
"Hello, my name is Tracy. How can I help you?"

Sorry "Tracy", but that's clearly not your real name. It's Manjula or something similar.

If you're going to pretend to be English, then cultivate an English accent, otherwise stop blatantly lying to me.

You don't live in Staines, you're in Calcutta (or whatever it's called these days).




"48% of women agtreed that product X really did improve their life (sample of 34 people)."

Sorry, but since when has 34 people been a statistically significant sample size?
( JugglerJAF It's curtains for you.Lacy gently wafting curtains, Mon 5 May 2008, 16:24, 5 replies)

re - the statistics
but 34 could be a statistically significant sample size. Depends on the experiment and the statistical test.

/pedantry
( crackhouseceilidhband Fuck off back to Mumsnet, Mon 5 May 2008, 17:19, closed)

Also sorry
but CHCB is right, the smallest statistically significant sample size is actually 30. But of course anyone with half a brain will look at those adverts and scoff, and I hate it too. If I had it my way I'd re-write the stats textbooks just so they couldn't do it.
(Petite chien is going to do science to it, Mon 5 May 2008, 18:46, closed)

48% is not good.
It still means 52% thought it didn't improve their life.
(paof2 Nude might not be rude, but rude could be nude..., Mon 5 May 2008, 20:30, closed)

and
for non-parametric tests you could probably get away with a sample size of 5 or 6...
( crackhouseceilidhband Fuck off back to Mumsnet, Mon 5 May 2008, 20:44, closed)

Betty Cooper
was the "name" of the last person I had to speak to when I called the Tiscali call centre. Why bother making up a British name, and why of all names choose Betty Cooper.
(hutchers, Mon 5 May 2008, 21:35, closed)