Hi all,   I have a simple mission yet with a complication I have not found a way to solve. I want to do a two-sample t-test by a continuous variable.   Say I have variables con1 and con2, both are continuous. I want to see if in group 1 (con1>=median of con1 across whole sample) vs. in group 2 (con1<median of con1 across whole  sample), values of con2 are different. Below is what I have right now:   proc univariate data=have; var con1; run; /* manually get con1 median, say the median is 1.523*/ data a1; set have;      if con1>= 1.523 then dummy=1; if con1< 1.523 then dummy=0; if con1=. then dummy=.; run; proc ttest data=a1; class dummy; var con2; run;   Essentially I manually get the median of con1, and make a categorical variable based on the median ("dummy" in the above example) . The problem with this is: due to rounding, though the median really is 1.522823104, SAS output would show median of 1.523. And this makes the sample sizes of group1 and group2 vastly different. I want to make sure that group1 and group2 have the same size (or different by 1 if whole sample size is an odd number). Is there an easy way of doing this?   I appreciate any help!   JOL     ... View more

Dear all,   I am trying to calculate next year value based on a panel data like below:   ID                Year                Value 1                 1993                0.29 1                 1994                0.65 1                 1995                0.79 2                 1993                1.23 2                 1994                1.56 3                 2001                0.98 3                 2002                0.73   What I would want is to calculate lead(value). For example:   ID                Year                Value                Value_nxt 1                 1993                0.29                   0.65 1                 1994                0.65                   0.79 1                 1995                0.79 ...................   Below is my code. Somehow it's not working right. Can anyone please tell me where went wrong and how I can fix this? Much appreciated!   proc sort data=have; by ID year; run; data want; set have; by ID year;      if eof1=0 then set have (firstobs=2 rename=(value=value_nxt) ) end=eof1;      if last.ID then value_nxt=.;  run;   JOL ... View more

Hi all,   I have such a proc glm test:   proc glm data=dataset1;      class f1 f2 f3 ;      model car = a b a*b f1 f2 f3 / solution; run; quit;   Normally I would get individual t value and p value for each of the independent variables. But I also want to test this:   b1+b3=0. b1 being the coefficient for variable a, and b3 being the coefficient for a*b   I've been doing some googling, and it seems that I need to use estimate or contrast statement? I didn't get the help I needed from googling. Anyone please help? Appreciate it! ... View more

Hi all,   I've been trying to deal with an issue like this:   Dataset:   ROW    ID      DATE 1           1       20010112 2           1       20010212 3           2       20000430 4           2       20000514 5           2       20000514 6           2       20010112 7           3       20090807   In the above dataset, ID 2 has two rows with the same date 20000514. I want to remove both of thses two rows. I tried using lag function, which will remove the second of the two rows (row 5) but not the first row (row 4). I suspect I should simulate a lead function? But somehow it doesn't work. Below are my codes:   /*This is me trying to mark row 5 for deleting later*/ proc sort data=have; by id date; run; data want1; set have;      by id date;      lagdate=lag(date);      if lagdate=date then removal=1; else removal=0; run;   /*This is me trying to mark row 4 for deleting later, but somehow it still marks row 5. */ proc sort data=want1; by id descending date; run; data want; set want1;      by id descending date;      leaddate=lag(date);      if leaddate=date then removal2=1; else removal2=0; run;   Can anyone help me solve this issue please? Thanks in advance!   J ... View more

Hi all,   I have a dataset about firm stock monthly returns that looks like the following:   firmID              date              return            fyr 1                    20100131      0.0003           9 1                    20100228      0.0001           9 ....................... 1                    20101231     0.0012            9 1                    20110131      -0.0015          9 ......................... 2                    20110430     -0.1201          12 2                    20110531      -0.0008         12 2                    20110630      0.0154          12 ........................... 2                    20151231     0.0087           12 3                    19960531     0.0789           5 3                    19960630     0.1023           5 ............................... 3                     20141231    -0.1234          5   I have over one-thousand firms, and these firms may have data up to 30 years. I have one return for each pair of firm-date, and these returns are monthly. fyr is the fiscal year-end. I want to calculate standard deviation for monthly return over the past fiscal year. In other words:   For firm 1, I want to calculate: 2010: standard deviation of (2009 Oct. return, 2009 Nov. return, 2009 Dec. return, 2010 Jan. return, ....., 2010 Sept. return) 2011: standard deviation of (2010 Oct. return, 2010 Nov. return, 2010 Dec. return, 2011 Jan. return, ....., 2011 Sept. return) ......   For firm 2, I want to calculate: 2011: standard deviation of (2011 Jan. return, 2011 Feb. return, 2011 Mar. return, ....., 2011 Dec. return) 2012 2013 ..........     For firm 3, I want to calculate: 1996: standard deviation of (1995 Jun. return, 1995 Jul. return, ...., 1995 Dec. return, 1996, jan. return,  ....., 1996 Apr. return) 1997 1998 ..............   So the output will look something like this:   firmID         year          std_ret 1               2010           0.20394 1               2011           1.23324 ............ 2              2011            0.234234 2              2012            1.234 ............. 2              2015            3.34732 3              1996            2.34234 .............. 3               2014            1.2345   I might have missing values on return. So I might not have the full 12-month data on each pair of firm-year. I want to restrict SAS to calculate only if I have at least 7 months out of 12 months. If there is not sufficient data, then SAS should drop this firm-year.   This is not a rolling standard deviation as I only use each line once. However for different firms, the calculation will be different as I start and end at a different month.   Please help! This seems difficult for me. Any suggestion is greatly appreicated. Thank you!   ... View more

Hi all,   I've recently begun using proc glm to run fixed effects regressions. Say if I want to run a regression of variable A over independent variables B and C, while controlling for year fixed effects, google tells me there are two ways to do this:   proc glm; class year; model A=B C / solution noint; run; quit;   or   proc glm; absorb year; model A=B C / solution noint; run; quit;   Google also tells me that the only difference between the above two is with absorb, estimates of year will not be shown, while with class, they will be shown.   So I tried both, and I don't see individual year fixed effects estimates in either. Also, perhaps more importantly to me, the estimates in B and C are vastly different with the above two procedures. Please help! Which one should I use? And what is the reason for difference estimates on B and C? Thanks a lot! ... View more