Review Session # 2

9/13/01

4 pm

Bring NOT0102 file - we will be reviewing it

We have been saying repeatedly that the normal or Gaussian distribution is a THEORETICAL distribution. But we can approximate it if we take repeated random samples from a larger population of scores; it does not matter if the population variable is normally distributed or not. If it is normally distributed then the population variable in case is a random variable.

A random variable has a large number of causes that are mixed throughout the population.

Some examples:

height

IQ

Some nonrandom variables:

age

Regardless of the form of the population variable, if we take repeated random samples we will get a sampling distribution or means or proportions that is normal

CENTRAL LIMIT THEOREM TELLS US SO!

SAMPLING DISTRIBUTION

•Of sample means, sample proportions, whatever
•Take a bunch of samples
•Sample error makes for a normal distribution on a sampling distribution
•We then compare our sample results to the sampling distribution

Each sample mean does not exactly match the population mean because of sampling error

Review of CENTRAL LIMIT THEOREM

What is the role of the SEM?

•NORMAL DISTRIBUTION

Symmetric

•It is a Frequency dist. W/ small bars at the upper and lower ends
•Area “under the curve” = 1

GAUSSIAN distribution

IF YOU HAVE A NORMAL DISTRIBUTION OF SCORES you only need the MEAN and the SD to describe it

Z DISTRIBUTION OR STANDARD SCORE DISTRIBUTION

mean=0

sd=1

variance=1

Z = [(score - mean) / sd ]