{[ promptMessage ]}

Bookmark it

{[ promptMessage ]}

day6prs

# day6prs - PADP 8120 Data Analysis and Sta4s4cal Modeling...

This preview shows page 1. Sign up to view the full content.

This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: 2/18/11 PADP 8120: Data Analysis and Sta4s4cal Modeling Comparing Groups PRACTICE Spring 2011 Angela Fer4g, Ph.D. Let's do hand calcula4ons first We want to know whether the average family income of female headed households is significantly different from that of male-headed households Get Xbar, s, and n of family income for both groups bysort femalehd: sum faminc Calculate z-score of the difference and get the p-value z= X women - X men 2 swomen s2 + men n women n men 1 2/18/11 Now with Stata Ttest faminc, by(femalehd) unequal Now a propor4ons example We want to know whether the propor4on of households receiving food stamps is different by female headed status Get propor4on, s, and n of food stamps for both groups Calculate z-score of the difference and get the p- value z= ^ ^ 2 - 1 1 1 ^ ^ (1- ) + n1 n 2 Bysort femalehd: sum fstamps 2 2/18/11 Now with Stata prtest fstamps, by(femalehd) Do one of each on your own 3 2/18/11 Read A&F Ch. 8 Make dra\ tables (using excel) of your descrip4ve sta4s4cs and make the appendix Table 1: sta4s4cs on outcomes, key independent and control variables Make sure there is one sample size for all variables so drop anyone missing variables; provide the n Every variable should have a meaningful sta4s4c mean, median; if categorical, break into relevant binary variables and provide propor4ons, etc. Provide meaningful labels (Family income, not faminc) For next 4me Table 2: Compare sta4s4cs across meaningful groups If studying effect of religiosity on dona4ons, break people up at roughly the median on the religiosity scale, and compare all sta4s4cs for the low religiosity vs. the high religiosity people, with significance tests Do comparisons for various groups (high dona4ons vs. low dona4ons, for example) if appropriate. Appendix: Detailed descrip4on of all variables. 4 ...
View Full Document

{[ snackBarMessage ]}