Fencepost
Admired Member
- Joined
- Dec 30, 2006
- Posts
- 429
- Media
- 3
- Likes
- 792
- Points
- 248
- Location
- Los Angeles, CA
- Sexuality
- 100% Straight, 0% Gay
- Gender
- Male
I think you throw yourself into the discussion.
You have to go back to your class. You didn't understand it at all...
The level of confidence is NOT the probability in a normal distribution.
Usually researchers use a level of confidence of 95% ( in experiments in labs even 99% ).
This determines how big the sample size must be of a certain population to calculate the margin of error ( the confidence interval ).
With a sample size of 300 you have a small margin of error for a big population.
That will change only a bit if you increase it to 1000.
The most important factor is the selection of the sample of the population that you want to investigate. (representative)
Sir, I am sure you are right- I am truly a rank amateur when it comes to sadistics. I will stand facing the back corner, wearing my dunce cap. I do, however, contend that we do not have a true, normal distribution- it is skewed just a bit to the right. I am not smart enough to determine how this skewing affects SDs. Can you illuminate the audience?