*************************************************** * Projekt: Analyse af summer school øvelsesdata * Forfatter: SCA * Dato: 11. feb. 2021 * Revision: * 23. feb. 2021: Tilpasning til Mac * 26. feb. 2021: Øvelse med regression *************************************************** * Setup--------------------------------------------------------- clear* cd "C:\Users\au103612\Dropbox (Dept of Pol Science)\Professorat\Undervisning\Uddannelsespolitik, BA seminar\Øvelser\Sievertsen_ovelse" *cd "/Users/au103612/Dropbox (Dept of Pol Science)/Professorat/Undervisning/Uddannelsespolitik, BA seminar/Øvelser/Sievertsen_ovelse/" * ssc install balancetable global kontrolvar "female parental_schooling parental_lincome " global outcomes "test_score_raw test_score" * Data preparation-------------------------------------------------- * Indlæse csv/xls data og gemme som dta import delimited "school_data_1.csv" * import delimited "/Users/au103612/Dropbox (Dept of Pol Science)/Professorat/Undervisning/Uddannelsespolitik, BA seminar/Øvelser/Sievertsen_øvelse/school_data_1.csv", delimiter(";") clear destring parental_schooling test_year_5 test_year_6, ignore(NA) replace save "school_data_1.dta", replace import excel "school_data_3.xlsx", firstrow clear save "school_data_3.dta", replace * Flette datasæt use school_data_1 foreach number in 2 3 { merge 1:1 person_id using school_data_`number' drop _merge } * Reshape til langt format (samme elev optræder på flere linjer - en for hvert klassetrin/år) reshape long test_year_, i(person_id) j(year) rename test_year_ test_score_raw lab var letter "Invitation letter" * Standardize test score by year bysort year: egen test_score = std(test_score_raw) /*For Stata 15 el. tidligere, som ikke tillader by i fbm. egen... std: egen test_score_mean = mean(test_score_raw), by(year) egen test_score_sd = sd(test_score_raw), by(year) gen test_score = (test_score_raw-test_score_mean)/test_score_sd */ * Data analysis------------------------------------------------------ * Balance table balancetable letter female parental_schooling parental_lincome test_score using balance.tex if year==5, replace booktabs ctitles("No letter" "Letter" "Difference (2)-(1)") pvalues * Compare means of treatment groups (looper over de to test scores: raw og standardized) foreach out in $outcomes { balancetable letter `out' using treatment_`out'.tex if year==6, /// replace booktabs ctitles("No letter" "Letter" "Difference (2)-(1)") pvalues } * Regression ----------------------------------------------------- reg summercamp $kontrolvar predict summercamp_predicted, xb predict summercamp_res, res browse summercamp* * Regression af test-score på deltagelse i summer school - bruge esttab til output til Latex/Word eststo clear eststo: reg test_score summercamp eststo: reg test_score summercamp $kontrolvar eststo: reg test_score summercamp_res *Tabel til Latex esttab using regression.tex, se ar2 booktabs replace *Tabel til Word *esttab using regression.rtf, se replace * Regression af test-score på randomiseret letter eststo clear eststo: reg test_score letter eststo: reg test_score letter $kontrolvar *Tabel til Latex esttab using regression_letter.tex, se ar2 booktabs replace *Tabel til Word *esttab using regression_letter.rtf, se replace