STATS 250- Midterm Exam Guide - Comprehensive Notes for the exam ( 14 pages long!)

• School: U of M
• Department: Statistics
• Course: STATS 250
• Professor: Brenda Gunderson

Document Summary

As n increases, the bell curve gets skinnier (smaller variance) Confidence interval: provides a range of plausible values for the parameter. If null value is in the interval can"t reject null hypothesis. If the null value is outside of the interval reject null hypotheis. If there is wide/narrow interval (dependent on level of significance, type of interval), and the null value is close to one of the interval values, you can"t tell. Look up new t* value and compute new margin. As significance level increases, the margin gets smaller. Hypothesis testing: tests if the null is true. If n is small, convert question to count and use binomial. If n is big, convert to count and use normal approximation, or use related normal approximation for sample proportion. Random sample: study design, if gathered over time make a time plot and look for stability. Stand deviation would stay the same but standard error would change.