Need some clarification with statistics

**anjunabeats** · Nov2-09, 08:45 AM

Im currently doing an introductory stats course and have learnt:

- poisson distribution
- binomial distribution
- normal distribution
- t-student
- and chi-square

I think I know what most of them look like but am having a bit of a hard time distinguishing what distribution to use when faced with questions.

Heres my current understanding:

- poisson distribution - when pop mean is given over an interval?
- binomial distribution - when given sample size and p of success
- normal distribution - n >30 or if normal
- t-student - when population is normal distributed but variance parameter is unknown
- and chi-square - not too sure when to use.

Wondering if anyone can help me with my current problems, thanks in advance!

**zli034** · Nov3-09, 10:46 AM

the normal distrubuted population's random estimated variance is chi square distribute, when you ask what probability you can obtain a variance you need chi square.

**bpet** · Nov3-09, 04:12 PM

Originally Posted by anjunabeats

I think I know what most of them look like but am having a bit of a hard time distinguishing what distribution to use when faced with questions.

Firstly, is the goal here to pass the exam or to learn statistics?

**anjunabeats** · Nov3-09, 07:44 PM

Originally Posted by bpet

Firstly, is the goal here to pass the exam or to learn statistics?

To learn statistics, im already doing well enough to pass the exam. Im probably going to learn much more difficult concepts in the future so its best if i get the fundamentals right.

**bpet** · Nov4-09, 07:39 AM

Originally Posted by anjunabeats

...its best if i get the fundamentals right.

The key is understanding what assumptions are reasonable for the given data, and which are necessary in order to use a particular distribution in your repertoire. There aren't any fixed rules as to what applies when, but as you learn more distributions and see the various ways they arise you'll gain a more intuitive understanding of which candidates are most suitable for each situation.

Good luck with your studies!

**anjunabeats** · Nov5-09, 01:08 AM

Originally Posted by bpet

The key is understanding what assumptions are reasonable for the given data, and which are necessary in order to use a particular distribution in your repertoire. There aren't any fixed rules as to what applies when, but as you learn more distributions and see the various ways they arise you'll gain a more intuitive understanding of which candidates are most suitable for each situation.

Good luck with your studies!

thanks for pointing out the obvious.

**bpet** · Nov5-09, 05:02 AM

Originally Posted by anjunabeats

thanks for pointing out the obvious.

Glad you appreciate the point - many don't, which is why so much dodgy research gets published.

Regarding the original questions, it sounds like you're on the right track except maybe the Poisson distribution - not sure what you meant by "pop mean is given over an interval", did you mean to say "average occurrence rate is given over an interval"?

**anjunabeats** · Nov5-09, 06:57 AM

Originally Posted by bpet

Glad you appreciate the point - many don't, which is why so much dodgy research gets published.

Regarding the original questions, it sounds like you're on the right track except maybe the Poisson distribution - not sure what you meant by "pop mean is given over an interval", did you mean to say "average occurrence rate is given over an interval"?

Yeah i think i understand poisson distribution more clearly now, just finished the course today probably planning to learn more chunkier stats in the future.