PSYCH 100A Lecture Notes - Lecture 22: Null Hypothesis, Type I And Type Ii Errors, Orbiter Boom Sensor System
Document Summary
Suppose you run an experiment with 1000 subjects. Then you run an identical experiment with 100 subjects. Answer: tobs is likely to be smaller in the 100-subject experiment. Now, calculate the 99% confidence interval given that your sample sd is 15. Went from z to t so you had to look up the tcv bc he gave you a sample sd not population. Answer: the independent-samples-tobs is usually smaller than dependent-samples-tobs. Hypothesized sample mean, sample size, population means, population. No bc 4. 16 is within the accept null region. It need to be beyond 4. 16 to be significant: power. Power = the area of the distribution beyond 4. 16: calculating percentage of area. Z2 (other group mean) - (4. 16-4. 6)/1. 04 = -0. 42: 3 ingredients we need: Z1: will always be 1. 96, we just need to determine if it"s positive or negative. Z2: calculate new z-score based on other distribution"s mean. Increase n = narrower distributions = less overlap --> more power.