Unformatted text preview: 2/10/11 PADP 8120: Data Analysis and Sta4s4cal Modeling Sta$s$cal Inference 2 Significance Tests PRACTICE Spring 2011 Angela Fer4g, Ph.D. Step 2: We want to know if the mean educa4on level is sta4s4cally different from 12 years. Null hypothesis: educhd=12 Alterna4ve hypothesis: educhd12 (two-sided) Example Hypothesis Test Step 3: Open up day2.dta in stata and get the mean and standard devia4on of educhd Sum educhd: mean=12.956, sd=2.619, n=7857 Do a t-test: t = y - 0 , where se = s se Step 4: Look up p-value Step 5: Interpret n 1 2/10/11 Graphically H0 = 12 Sample mean = 12.9 Educa4on of Head H0 has a very small chance of being true. How to do it in Stata est educhd==12 --check that t-stat matches our calcula4on If you type reg educhd, there is a t-stat reported. This is tes4ng whether the mean is significantly different from 0. Calculate it by hand yourself. t= y-0 y = se se 2 2/10/11 Sample size ma]ers a lot Let's randomly choose 50 observa4ons bsample 50 ]est educhd==12 Is the sample's mean educa4on significantly different from 12 now? Graphically H0 = 12 Sample mean = 12.9 Educa4on of Head When n=50, H0 has a good chance of being true. 3 2/10/11 Now you do it Come up with your own hypotheses (using any PSID dataset you already have) Work together or alone, whatever you prefer Get the t-stat (compute by hand and then check with Stata) Get the corresponding p-value Interpret the result Everyone will report their result when we are done Next 4me Read A&F Ch. 7 Homework 4 due 4 ...
View Full Document
This note was uploaded on 01/18/2012 for the course PADP 8120 taught by Professor Fertig during the Summer '11 term at UGA.
- Summer '11