About JOLSAS

JOLSAS · ‎08-22-2016

Hi ballardw, Thanks for the response! I did what you suggested, and it worked much better (though still not what I had hoped for, but I suppose I can settle with the much smaller sample size difference). For future reference, below is what I had in the end: proc rank data=have groups=2 out=want; var con1; ranks dummy; run; Thanks for the help!

JOLSAS · ‎07-18-2016

Yes It worked!!! Thanks a lot!

Rick_SAS · ‎05-27-2016

The best place to start is the documentation of the ESTIMATE statement. Since you are just getting started, you might want to read CONTRAST and ESTIMATE Statements Made Easy:The LSMESTIMATE Statement I also recommend Practicalities of Using ESTIMATE and CONTRAST Statements

JOLSAS · ‎05-26-2016

Thanks a lot!

JOLSAS · ‎04-21-2016

Reeza, sql does work. I identified all the id-date pairs that have multiple dates using the code below: proc sql; create table dup as select distinct id, date, count(date) as count from have group by id, date order by id, date, count; quit; Later on I mark those that have count>1 and merge with the original table. I was able to delete them. Thanks for help!

JOLSAS · ‎02-29-2016

Hi Tom, Thanks for the help! FiscalYear essentially puts all the months I want in a fiscal year under one value of FiscalYear. Ingenious! I ran your code, then added the following, and got what I want. In case anybody needs these in the future: proc sql; create table part2 as select firmID, std(return) as std_ret, FiscalYear, fyr from part1 group by firmID, FiscalYear, fyr; quit; data want; set part2; if fyr>=6 then fyear=FiscalYear+1; else fyear=FiscalYear; drop FiscalYear fyr; run; I defined fyear slightly different from FiscalYear because I need to merge the table want with other tables based on firmID and fyear. Problem solved! Thank you very much! JOL

JOLSAS · ‎12-13-2015

Hi Mohamed, Thanks a lot for the answer! I added year in and saw the individual estimates. Also when I included year as an independent variable, the estimates are not different with class or absorb. Thanks again!

JOLSAS · ‎11-01-2015

Thanks a lot! I ended up changing the format of date_as_numbers to real dates by doing the following: newdate = input(put(date,8.),yymmdd8.); format newdate yymmddn8. ; And datasets merged! Cheers!

MichelleHomes · ‎03-03-2014

Oops... Of course. A reset at the start of each id is needed. Good pick up Kurt. Cheers, Michelle

SteveDenham · ‎04-23-2013

While coding it all through IML is possible, this looks exactly like what PROC QUANTREG should give you. Take a look at the documentation here: http://support.sas.com/documentation/cdl/en/statug/63347/HTML/default/viewer.htm#qreg_toc.htm Steve Denham

Reeza · ‎04-24-2013

Yup, that was it, thanks

Reeza · ‎08-22-2011

Have you checked the output using a contrast statement?

JOLSAS · ‎08-12-2011

The shr is shares of options owned and the opt is value realized from option exercising. Yes, shr is a weight and I'm trying to compute the weighted average over the years. The reason that one company, in one year, could have multiple lines, is because one company, in one year, have multiple executives. One executive could have no share of options, while another could have some. Therefore, there are numbers and missing values. As a matter of fact, I cannot be 100% sure on this either. I think the best I can do is to try both sum/count and the average function, and see which produces the better result. In fact, I just asked my professor for his opinion.

DBailey · ‎08-12-2011

By identifying the most recent position, and the one immediately prior, you can calculate the changes this way as well: data work.detail_data; input gvkey execid year ceo cfo other; cards; 103 6 1992 0 0 1 103 6 1993 1 0 0 103 6 1994 1 0 0 259 6 1996 0 0 1 259 6 1997 0 0 1 259 6 1998 1 0 0 259 6 1999 1 0 0 259 6 2000 1 0 0 1990 6 2003 0 0 1 1990 6 2004 0 0 1 1990 6 2005 0 1 0 1990 6 2006 0 1 0 34 19 1994 0 0 1 34 19 1995 0 0 1 34 19 1996 0 0 1 34 19 1997 0 0 1 2503 19 1998 0 0 1 2503 19 1999 1 0 0 2503 19 2000 1 0 0 2503 19 2001 1 0 0 2503 19 2002 1 0 0 ; run; proc sql; create table work.most_recent as select t1.* from work.detail_data t1 where year=(select max(year) from work.detail_data where execid=t1.execid); create table work.immediately_prior as select t1.* from work.detail_data t1 where t1.year=(select max(year) from work.detail_data where execid=t1.execid and year < (select year from work.most_recent where execid=t1.execid)); select case when t1.ceo=1 then 'CEO' when t1.cfo=1 then 'CFO' else 'Other' end as From_Title, sum(case when t2.ceo=1 then 1 else 0 end) as To_CEO, sum(case when t2.cfo=1 then 1 else 0 end) as To_CFO, sum(case when t2.other=1 then 1 else 0 end) as To_Other from work.immediately_prior t1 inner join work.most_recent t2 on t1.execid=t2.execid group by calculated from_title; quit; which produces this output: From_Title To_CEO To_CFO To_Other CEO 1 0 0 CFO 0 1 0

JOLSAS · ‎08-11-2011

It worked perfectly!!!!!!!!!! Thank you sooooo much!!!! God, I would have never done this by myself.

Online Status	Offline
Date Last Visited	‎03-24-2018 02:24 AM

Re: Proc ttest: class continuous variable

Proc ttest: class continuous variable

Re: Calculating lead values

Calculating lead values

proc glm: estimate and contrast

Re: PROC GLM: hypothesis testing

Re: PROC GLM: hypothesis testing

PROC GLM: hypothesis testing

Re: Deleting duplicating dates

Deleting duplicating dates

Re: Change date format and informat

Re: Proc ttest: class continuous variable

Re: Calculating lead values

Re: proc glm: estimate and contrast

Re: PROC GLM: hypothesis testing

Re: Deleting duplicating dates

Re: Calculate standard deviation over fiscal year

Re: proc glm: class vs. absorb

Re: Change date format and informat

Re: RETAIN statement and iteration

Re: Need help with proc iml

Re: Get the empirical CDF

F statistic on part of the independent variables

calculate a new variable using variables * cases

resolution for missing lead function

Re: Grab items that fits this condition.