Hello Team, I have a weighted count from a sample, but it returns a different report when I run the proc means operation on the same sample. 'Sub' is a binary-valued variable (1,0) and has no missing values but values that are less than the weighted count. Could someone please clarify this to me? Thank you! /*weighted count*/ proc sql; create table sample as SELECT Year,group, Sum(weight) AS freq FROM biz.workers(where=(age between 25 and 64 and year>=2014)) GROUP BY year,group; run; /*Statistics for Sub*/ proc means data=biz.workers(where=(age between 25 and 64 and year>=2014)) n mean stderr lclm uclm alpha=0.05 vardef=df Maxdec=4; class year group; var sub /weight=weight; run; WEIGHTED COUNT OF POPULATION YEAR A B C Total 2015          798,535              62,753                  76,191                 937,479 2016          802,110              66,391                  81,967                 950,468 2017          824,222              60,694                  90,189                 975,105 2018          829,473              74,191                  94,218                 997,882 2019          838,824              77,843                  99,673              1,016,340   Analysis Variable : Sub Census year Group N Obs N Mean Std Error Lower 95% CL for Mean Upper 95% CL for Mean 2015 A           401,948              401,948 0.001 0.000 0.001 0.001 2015 B             35,036                35,036 0.000 0.000 0.000 0.001 2015 C          101,315              101,315 0.000 0.000 0.000 0.001 2016 A          409,568              409,568 0.001 0.000 0.001 0.001 2016 B            35,755                35,755 0.000 0.000 0.000 0.001 2016 C          106,454              106,454 0.000 0.000 0.000 0.001 2017 A          419,671              419,671 0.001 0.000 0.001 0.001 2017 B            35,641                35,641 0.000 0.000 0.000 0.001 2017 C          114,356              114,356 0.000 0.000 0.000 0.001 2018 A          425,452              425,452 0.001 0.000 0.001 0.001 2018 B            36,094                36,094 0.001 0.000 0.001 0.001 2018 C          118,881              118,881 0.000 0.000 0.000 0.001 2019 A          437,484              437,484 0.001 0.000 0.001 0.001 2019 B            36,043                36,043 0.000 0.000 0.000 0.000 2019 C          124,615              124,615 0.000 0.000 0.000 0.000 ... View more

Hello, I'd like to see how a program's pre- and post-implementation trends compare. I've having difficulties correctly creating the positive and negative values. Please see sample data below (Trend should be 0, if the program was implemented in 2012, +1 for every year after, -1 for every year before) data biz_trend; set biz_data; by business adoption_year; retain Trend; if first.business and first.adoption_year then Trend=0; else Trend=Trend+1; run;   Year Business Products Adoption_year 2010 Company A     2011 Company A     2012 Company A convenience  2012 2013 Company A     2014 Company A shopping 2014 2014 Company A raw materials 2014 2015 Company A     2016 Company A     2017 Company A major equipment 2017 2010 Company B     2011 Company B     2012 Company B     2013 Company B     2014 Company B     2015 Company B     2016 Company B raw materials 2016 2017 Company B convenience  2017 2010 Company C     2011 Company C     2012 Company C     2013 Company C     2014 Company C convenience  2014 2015 Company C     2016 Company C major equipment 2016 2017 Company C convenience  2017   What I want Year Business Adoption_year Trend 2010 Company A   -2 2011 Company A   -1 2012 Company A 2012 0 2013 Company A   1 2014 Company A 2014 2 2014 Company A 2014 3 2015 Company A   4 2016 Company A   5 2017 Company A 2017 6 2010 Company B   -6 2011 Company B   -5 2012 Company B   -4 2013 Company B   -3 2014 Company B   -2 2015 Company B   -1 2016 Company B 2016 0 2017 Company B 2017 1 2010 Company C   -4 2011 Company C   -3 2012 Company C   -2 2013 Company C   -1 2014 Company C 2014 0 2015 Company C   1 2016 Company C 2016 2 2017 Company C 2017 3 ... View more

data want; set have; Mili_Year_1=first_year_mili; Mili_Year_2=first_year_mili +1; Mili_Year_n=Last_year_mili; run; Hello, I have a wide-form dataset that I'd like to convert to long-form, but the years are inconsistent. I want to create new variables that correspond to the period for each regime and does not extend past the last year by regime. What I have on the left is what I have, and what I want on the right is what I want. I want to do this without listing all the variables. Thank you very much! country first_year_mili last_year_mili total_years_mili first_year_civ last_year_civ total_years_civ first_year_int last_year_int total_years_int   country Year Type A 2000 2004 5 2006 2010 5 2005 2005 1     A 2000 mili B 2005 2005 1 2007 2012 6 2006 2007 2     A 2001 mili C  2002 2007 6 2010 2012 3 2009 2009 1     A 2002 mili                         A 2003 mili                         A 2004 mili                         A 2005 int                         A 2006 civ                         A 2007 civ                         A 2008 civ                         A 2009 civ                         A 2010 civ                         B 2005 mili                         B 2006 int                         B 2007 int                         B 2007 civ                         B 2008 civ                         B 2009 civ                         B 2010 civ                         B 2011 civ                         B 2012 civ                         C  2002 mili                         C  2003 mili                         C  2004 mili                         C  2005 mili                         C  2006 mili                         C  2007 mili                         C  2008                           C  2009 int                         C  2010 civ                         C  2011 civ                         C  2012 civ                               ... View more

Hello Experts,   I have a dataset (see attached) with information for students in an examination. I would like to calculate the weighted average of all levels for each unit by category and year E.g for Unit 1090, English, 2010 weight(level 1) =((Attended(Level 1)/Attended(1)+.............Attended(Level B2))>>>>>>for all levels weighted_mean=sum(weight(Level 1)*mean_score(Level 1)..............weight(Level B2)*mean_score(Level B2))>>>>>>>for all levels I tried to create macros for each variable required to compute the weighted mean but it takes forever to run. What's the best way to do this? can I create all the macros in a single batch of codes? and how do I put the macros together to calculate the weighted mean.   %macro unit_loop(start,end); proc sql; %do unit=&start. %TO &end.; create table exam.weighted&unit. as select *from exam.scores where unit_code=&unit.; %END quit; %mend; %unit_loop(start=1090,end=401401) Thank you,   ... View more