* Income inequality between men and women in Germany (GSOEP-data) * --------------------------------------------------------------- version 9 set more off capture log close log using an2, replace * Data: GSOEP 1997, Sample use data1, clear // -> Note 1 drop ymove - np9507 * Descriptive statistic of income summarize income mvdecode income, mv(0=.a) // -> Note 2 sort gender by gender: summarize income summarize income, detail * Employment by gender label define emplb 1 "full time" 2 "part time" 3 "retraining" /// 4 "irregular" 5 "not working" 6 "military service" 7 "unemployed", modify label values emp emplb tabulate emp gender, colum nofreq // -> Note 3 * Preparation for the regression analysis generate men = gender == 1 // gender generate fulltime = emp == 1 // fulltime vs. part-time replace fulltime = . if emp > 2 // -> Note 4 * Regression analysis I regress income men fulltime * Preparation regression analysis II generate age = 1997 - ybirth // age egen miss = rowmiss(income fulltime men age) summarize age if miss==0 // -> Note 5 generate age_z = age - r(mean) if miss == 0 gen agemen = age_z * men // age x gender * Regression analysis II regress income men fulltime age_z agemen log close exit Description ----------- This is an analysis of income inequality between men and women in Germany. Hypothesis see Kohler/Kreuter (2003, Chapter 1 and 2). The higher amount of part time working women is not a sufficient explanation for the inequality in average income between men and women. And in addition, even though there is a higher income inequality among older people, younger women are still affected. Notes: ------ 1) SOEP - Education-Sample and random selection of one person per houshold (Created with crdata1.do). 2) Respondents with income equal to zero are excluded from further analysis 3) Women are more often part time employed than men. It is reasonable to control for employment status. 4) This command excludes all respondents that are not employed. 5) Centering the age variable -- see Aiken/West (1991)