STATISTICAL INFERENCE

SUMMARY

How To Choose The Appropriate Method For Statistical Inference

1. Are the data continuous or categorical:?
2. How many sets of measurements ?
   one - 1 variable measured on 1 sample
   two - 1 variable measured on 2 (independent or dependent) samples
   two - 2 variables measured on 1 sample

Scale	Sets of measurements
	1 sample	2 samples	1 sample
	1 variable	1 variable	2 variables
continuous	z or t	z or t	correlation or regression
categorical	binomial or c2	binomial or c2
	c2

1 sample/group - 1 variable - continuous scale

To make inferences about µ

use Z = ~ N(0,1) if s is known

OR

use ~ t_n-1 if s is estimated by s.

2 samples/groups - 1 variable - continuous scale

1 sample/group - 2 variables - both continuous

(i) Are both variables random (e.g. measurements or responses) or were the values of one variable chosen by the investigator?

(ii) Association or prediction ?

1 sample - 1 variable - categorical

2 samples/groups - 1 variable - categorical

1 sample/group - 2 variables - categorical