About JOLSAS

JOLSAS · ‎08-20-2011

Hi all, I have a regression on multiple independent variables. While the output automatically gives the F statistic on all the independent variables, I would like the F statistic on a subset of independent variables that I specify. An example would be: dependent = independent 1 + independent 2 + independent 3 + independent 4 + independent 5 I'm interested in the F statistic on the variables independent 1 - 3. Any ideas how to do this? Thanks.

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.

JOLSAS · ‎08-12-2011

I thought it was something like this! I tried using & but it gave me error. Thank you. It did give me what I want. Do you have any idea how to output the result in a new table? Is it something like proc sql; create table work.newtable as select * from ........? Thanks.

JOLSAS · ‎08-12-2011

Thanks. Including null values is what I want. I ran your code, and it gave me very good results. Now the thing is, a few different companies have these year-data. It;s like this: gvkey year shr opt 105 1992 2.4 6 105 1992 2.3 . 105 1992 . 86 105 1993 . . 105 1993 2.3 0 105 1993 3.5 8.9 258 1992 . 3.5 258 1992 3.5 0 258 1993 6 8.9 258 1994 . . 258 1994 0 2.3 If only group by year, the sas would return me average values for all the company in each year. What if I want average value of each company in each year? In the case above, the output would be like: gvkey year newvar 105 1992 105 1993 258 1992 258 1993 258 1994 Any ideas? Thank you. BTW, sometimes sql does lots of things with relatively simple codes! A good idea to use sql in this case!

JOLSAS · ‎08-12-2011

In the example, ( . * 6.5 + . * . + 2.3 * 0 + 2.6 * 0 + 3.5 * 8.9)/5 would equal to (. + . + 0 + 0 + 3.5 * 8.9)/5 = 3.5 * 8.9 / 5 . Does this make sense? Basically, if either of the two existing variables is missing, this makes the product of that line a missing value. But I think this wouldn't affect the entire equation. In the case of zeores, the product would be zero. But since I don't know for sure if the equation for each year contains how many missing values/zeroes/"good" values, I have to count all of them.

JOLSAS · ‎08-11-2011

I have a data set looking like this: year shr opt 1992 2.4 6 1992 2.3 . 1992 . 86 1993 . 6.5 1993 . . 1993 2.3 0 1993 2.6 0 1993 3.5 8.9 I want to output a new table looking like this: year new variable 1992 1993 where the new variable is the average of shr * opt in all the years. e.g. the new variable for 1992 equals (2.4 * 6 + 2.3 * . + . * 86)/3 and the new variable for 1993 equals ( . * 6.5 + . * . + 2.3 * 0 + 2.6 * 0 + 3.5 * 8.9)/5 I want to output the new variable for each year. Any idea how to do this? Thanks.

JOLSAS · ‎08-11-2011

Hello art, I am so sorry and embarrassed. Remember the selecting data you helped me with yesterday? While working on this table, I found a sentence in very small font, in a peculiar place, about the selecting data. What I described yesterday is to pick those years that executives change their firms. If they have multiple changes, then we pick all of them. Well, this new sentence told me that we should only pick up the last time the executive changes firms, which is to say, each executive would have only two different companies, being the last two companies he stayed. If he had previously stayed in other companies, we would delete the older items. I tried to go back and make revisions to your codes, but it's just too much for me... Can you think of a way of either revising the codes, or writing something new to delete those older company years? Wichever is convenient. In fact, with each executive only having two firms, I think I might be able to finish this table myself. Again, thank you so much!

JOLSAS · ‎08-11-2011

I looked at Chang's output /* on log ex current CEO CFO Other CEO 1 1 2 CFO 0 0 0 Other 2 1 3 */ I thought this MIGHT be the reason. Chang's codes seem to count the sum of how many cases of company change for each executive. For example, the data set has three cases of company changes, and one case of ceo -> ceo,one is ceo -> cfo, and the other is other -> ceo. The change of 1990 -> 34 doesn't count because the executive changed. I only care about executive changing their companies. But I don't care if I reach a new executive. I don't know if this makes sense... I guess what I'd want is a table like this: /* on log ex current CEO CFO Other Total CEO 1 1 0 2 CFO 0 0 0 0 Other 1 0 0 1 2 1 0 3 */ So basically I need to count the number of title changes as executives move from company to another. I hope this makes a little bit more sense. Anyway, thank you guys so much. I've been able to make some progress with my project, which I couldn't have done without your help.

JOLSAS · ‎08-11-2011

I had realized it and I fixed it. So I was looking at the result. I know I have 1143 different executives in total. so the number of ceo, cfo, and other should add up to 1143, and so should those of exceo,excfo, and exother. I used: proc freq data=three; tables ceo cfo other exceo excfo exother; run; but I have ceo, cfo , other adding up to 3114, and 1518 for the sum of the exes. I'm trying to figure out why...... Maybe I got something wrong.....

JOLSAS · ‎08-11-2011

Wow, I definitely appreciate you putting this much effort on this. Thank you very much! So I've run it, and after the line 6 "set one;", the log file says: ERROR: Variable ceo has been defined as both character and numeric. Any idea why is this? Thank you. ERROR: Variable cfo has been defined as both character and numeric.

JOLSAS · ‎08-11-2011

Dear all, I know there isn't a lead function in SAS and I happen to need this function here in this project I'm doing and this project is due in two weeks. I've googled a few solutions but I still cannot figure out how to do it in my example here. Time is urgent. So please help! I have a data set looking like this: gvkey execid year ceo cfo other 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 It's basically saying: the executive with ID (execid) 6 works in the company 103 (gvkey) from 1992 to 1994. S/he is a CEO in 1993 and 1994 (as shown by the "1" in the ceo column) and not a CFO or other executive (as shown by the "0"s in the other columns). S/he is an "other" executive in 1992 though. Then in 1996, this executive goes to the company 259 to work, being an "other" executive in the first two years and then a CEO in the last three years. Then he moved to the company 1990, and is an "other" executive for the first two years and a CFO for the next two years. And then the next section is for the executive #19. Now I'm interested in how many people are CEOs/CFOs/others in their "old" companies (as indicated by the title the last year he's there) and how many are CEOs/CFOs/others in their new companies (as indicated by the LAST year he's there)? This means that #6 moves from 103 as a CEO to 259 as a CEO but not as an other executive and then moves to 1990 as a CFO. Further, I need to build a table and fill in all the nine blanks (except the one under "prior title"): prior title new title CEO CFO other CEO CFO other So, how many people for all the nine possibilities of prior title -> new title? Using last.gvkey=1 and last.execid=0, I was able to identify those "old"titles. But to identify new titles and match these two, I need the lead function. Can someone help me with this please? Thank you so much!

JOLSAS · ‎08-11-2011

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

JOLSAS · ‎08-11-2011

Well this project I'm working on is a statistical one. But I guess you're right, this particular question is not a statistical one. Sorry!:smileysilly:

JOLSAS · ‎08-11-2011

Thank you so much for your reply! I'm still studying your program but I've run it and it did select those executives. I guess I didn't explain it clearly though. I would need all the years for the executives. So in the case of the executive #2, I would want the years in the firm 10553 from 1996 to 2000, and all the years in the firm 132502 from 1998 to 2001. This is to say, there are nine items for the executive #2 in the "want" data set (actually eight items if using the data set I provided, since #2 has three years in the second company, which satisfies the minimum required years condition ). Therefore, "want" would contain all the firm-years for each executive who satisfies the condition. Sorry I didn't explain clearly. Would you show me how to do this again? Thank you.

JOLSAS · ‎08-10-2011

Hi all, I'm quite new to SAS and I have this project that is due in two weeks. While I'm learning to program, what I need to do first is quite a big challenge. So I need your help! Basically I have a data set that has variables "execid", which is the IDs for company executives, "gvkey", which is the IDs for the companies. I also have a variable "year" indicating different years. The dataset looks like this: GVKEY EXECID YEAR 10553 2 1996 10553 2 1997 10553 2 1998 10553 2 1999 10553 2 2000 132502 2 1998 132502 2 1999 132502 2 2001 1078 3 1992 1078 3 1993 1013 6 1996 1013 6 1997 1013 6 1998 1013 6 1999 1013 6 2000 1013 6 2001 This says executive #2 works in the company 10553 from 1996 to 2000, he also works in the company 132502 from 1998 to 2001. Executive #3 works in the company 1078 from 1992 to 1993. Executive #6 works in the company 1013 from 1996 to 2001. Also, an executive may work in three or even four companies in the data period: 1992 to 2006. Now the task is to select those executives that "are employed for at least three years in each of at least two different firms". So in the case of the above example, executive 2 would qualify. #3 wouldn't qualify since he only works in 1078 for two years. #6 wouldn't qualify either since s/he only worked for one company. While I can do some simple defining variables, sorting, compute frequencies, etc. This is a little too much for me. My whole data set has over 100,000 items and Iexpect only a small part of them would qualify. Can someone please help me? I'm desperate. Many thanks!

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

F statistic on part of the independent variables

calculate a new variable using variables * cases

calculate a new variable using variables * cases

calculate a new variable using variables * cases

calculate a new variable using variables * cases

calculate a new variable using variables * cases

resolution for missing lead function

resolution for missing lead function

resolution for missing lead function

resolution for missing lead function

resolution for missing lead function

Re: Grab items that fits this condition.

Grab items that fits this condition.

Grab items that fits this condition.

Grab items that fits this condition.