NOTES, ABBA DS 123, Chaffe-Stengel, 6/6/2012

HYPOTHESES: Between H0 and HA, the entire number line should be covered. The textbook does not follow this protocol. Always form the hypotheses with the population parameter, not the sample statistic.

H0: Always assume H0 is true. The EQUAL SIGN always goes in H0. Use [pic].

HA: Because the alternative hypothesis contains the reason for doing the study in the first place, it is often easier to form HA first. The EQUAL SIGN NEVER goes in HA. Use [pic].

REJECTION REGION:

( = Show graph with critical State decision rule
[If t, give df.] boundary/ies & labels

In a one-tailed hypothesis test, α appears in one tail of the distribution. In fact, α appears in the tail that the alternative hypothesis designates. The Type I error, α, is the amount of area left in the tail beyond the critical z or t bound, or the amount of area in the rejection region, as shown in Figures 1 and 2 below.

| | |
| | |
| | |
| | |
| | |
| | |
| |