**Matched pairs***t*test- We measure the
**same**unit twice - e.g. "before" and "after" for the same person

- We measure the
**2-sample***t*test- We have 2
**seperate**populations and hence 2 independent samples - e.g. men vs. women

- We have 2

**Matched pairs***t*test- Create two
**numeric**variables representing the pair - Variable view:
- Data view:

- Create two

**Two-sample***t*test- Create one
**numeric**variable containing all the data (both samples), and one**string**variable labeling the group (e.g. "A" and "B") - Variable view:
- Data view:

- Create one

- Always use SPSS to solve
*t*test problems - In 2-sample comparison of means output,
**ALWAYS**use the bottom row (equal variances**NOT**assumed) - In matched pairs, we only have one $\mu$ ($\mu_{diff}$) in the null hypothesis, e.g. $H_0:\mu_{diff}=0$. State what "diff" stands for.
- In 2-sample
*t*test, we have two $\mu$'s, e.g., $\mu_A=\mu_B$. State what "A" and "B" stand for. - http://statr.me/courses.html