Hi Ivm, I agree with your point when there are plenty of repeated observations in each subject cluster, either as discrete observations or via a frequency variable in the data. At least one other of our highly esteemed and luminary community members also agrees with you.  However where this is not always the case, to quote http://support.sas.com/documentation/cdl/en/statug/66859/HTML/default/viewer.htm#statug_glimmix_details13.htm " ... In this case, provided that the constant is large, the Laplace approximation to the marginal log likelihood is " I researched what is considered 'large' and found some surprising opinions - like 30(!) so I compared some of my own models, and found, to my relief, that 3-4 repeated measures did indeed indicate the laplace method was preferred. I'm now not so sure about precision of the LL approximation when there are fewer measures per subject cluster. Has anybody else noticed thought about this? Cheers. Damien ... View more

Hi FranAstorga. I only just got to helping out on this forum recently. Good on you for finding the explanation. I'm a fan of the laplace method in GLIMMIX as it has decreased estimation time and memory resource demands on big data sets dramatically. Another effect of this method is that in some situations it may introduce estimate bias..there is a note to this effect in the GLIMMIX documentation..from memory when number of repeat measure on subject are low. However if GLIMMIX fits a psuedo-likelhood with another option successfully as well to your data you could always use that as a check for bias and if none is evident then use the non-psuedo AIC (sounds better than it is I suspect!) with confidence where needed. Good luck with your GLIMMIX analysis. - Damien Mather ... View more

Nothing obviously wrong with the code except..when you get it to run so that you are happy..can you assume equal variances across strain, comp and comp*strain? If not, another adjustment choice might be more appropriate. Maybe this would get more helpful suggestions if you re-posted to the non-SAS-U-specific statistics discussion forum. The question is a bit too specialised for this forum I think.  Cheers. Damien ... View more

I can highly recommend Paul D. Allison's book: Survival Analysis using SAS: A Practical guide. As his title suggests, it is very practical! It covers a broad range of survival and hazard models. I found it a great reference when developing my own hazard models. see: SAS Press - Paul D. Allison Author Page Cheers. Damien ... View more

Thanks Steve. Nope, absolutely sure the data did not change. For a start, nothing was done to perm.yogurt3 between the proc glimmix steps. Secondly, step1 reported: Number of Observations Read 53027 Number of Observations Used 51441 Step2 reported: Number of Observations Read 53027 Number of Observations Used 51441 . We would generally consider the first set of covariance estimates a disappointment: CHOL(1,1) 0.9208 . CHOL(2,1) 0.9208 . CHOL(2,2) 0.9208 . CHOL(3,1) 0.9208 . CHOL(3,2) 0.9208 . CHOL(3,3) 0.9208 . CHOL(4,1) 0.9208 . CHOL(4,2) 0.9208 . CHOL(4,3) 0.9208 . CHOL(4,4) 0.9208 . CHOL(5,1) 0.9208 . CHOL(5,2) 0.9208 . CHOL(5,3) 0.9208 . CHOL(5,4) 0.9208 . CHOL(5,5) 0.9208 . CHOL(6,1) 0.9208 . CHOL(6,2) 0.9208 . CHOL(6,3) 0.9208 . CHOL(6,4) 0.9208 . CHOL(6,5) 0.9208 . CHOL(6,6) 0.9208 . CHOL(7,1) 0.9208 . CHOL(7,2) 0.9208 . CHOL(7,3) 0.9208 . CHOL(7,4) 0.9208 . CHOL(7,5) 0.9208 . CHOL(7,6) 0.9208 . CHOL(7,7) 0.9208 . CHOL(8,1) 0.9208 . CHOL(8,2) 0.9208 . CHOL(8,3) 0.9208 . CHOL(8,4) 0.9208 . CHOL(8,5) 0.9208 . CHOL(8,6) 0.9208 . CHOL(8,7) 0.9208 . CHOL(8,8) 0.9208 . HouseHoldID 0.1690 0.01440 PermMissAlts 0.9367 0.6027 Scale 4.2298 0.06749 Whereas the second step set are a lot more interesting (and credible!), and the transformations into the G matrix and thence into the GCORR matrix yield quite useful and actionable inferences for brand managers: CHOL(1,1) -0.00559 . CHOL(2,1) -0.00493 . CHOL(2,2) -0.00481 . CHOL(3,1) -0.00568 . CHOL(3,2) -0.00525 . CHOL(3,3) -0.00950 . CHOL(4,1) -0.00170 . CHOL(4,2) -0.00172 . CHOL(4,3) -0.00136 . CHOL(4,4) -0.00307 . CHOL(5,1) -0.00523 . CHOL(5,2) -0.00481 . CHOL(5,3) -0.00552 . CHOL(5,4) -0.00160 . CHOL(5,5) -0.00525 . CHOL(6,1) -0.00639 . CHOL(6,2) -0.00528 . CHOL(6,3) -0.00478 . CHOL(6,4) -0.00097 . CHOL(6,5) -0.00590 . CHOL(6,6) -0.01573 . CHOL(7,1) 0.007528 . CHOL(7,2) 0.006518 . CHOL(7,3) 0.006129 . CHOL(7,4) -0.00018 . CHOL(7,5) 0.007475 . CHOL(7,6) 0.009907 . CHOL(7,7) -0.02040 . CHOL(8,1) 0.001548 . CHOL(8,2) 0.001323 . CHOL(8,3) 0.002397 . CHOL(8,4) -0.00020 . CHOL(8,5) 0.001542 . CHOL(8,6) 0.002046 . CHOL(8,7) -0.00379 . CHOL(8,8) -0.00168 . HouseHoldID 9.676E-9 0.006443 PermMissAlts 0.02020 0.005175 Scale 1.6093 . I was thinking that it may be that the first step stalled at a stationary point NOT the global optimum, despite both steps specifying the same optimizing algorithm (double dogleg). I noted that the double dogleg uses the gradient to update an approximate hessian matrix. I wonder if one or more of the (apparent) RANDOM statement output options somehow influenced the gradient calculation (maybe just in terms of accuracy), thereby influencing both the approximate hessian update and the endpont? Thoughts? ... View more

Hi Steve. Richard DeVenezia's web page is an excellent resource too, and it has an example of the call module (Darryl's suggestion) that might be a fairly close template to what you want: it calls the win MD5 dll from sas. see http://www.devenezia.com/downloads/sas/sascbtbl/ Let us know if this apporach helps you. ... View more

Hi Toto. I have summarised your problem description and provided some code that may or may not be of more help to you as a template for what you need to do. To be fair to us SAS community helpers I am guessing that you could have given us all some more information about your problem. To be fair to you as a student I am guessing you are probably stressing out about your research task.  Am I right so far? This is likely still not all of what you need but hopefully it will stimulate some more clarification and/or be close enough so you can figure out the rest. Please let us all know how you get on and if you need more suggestions or help. Cheers. Damien /* I work on creating an omission sample with some restrictions in order to replicate a paper for my research. The paper I am trying to replicate mentions that "Using CRSP data on dividend payment history, a potential dividend omission is identified when a firm has not paid a dividend within 1 quarter, 6 months or 1 year from the previous payment if the firm pays quarterly, semi-annual or annual dividends respectively." In this identification, we allow for "late" payments and define a 3-year period to consist of 1128 days or approximately 31 days in a month. For example if I had the following obseravtions and variables how do I add the restriction I mentioned above (based on the restrction this is not an omission event as the last declaration is in the 3 -year period) */ data Have01(label='Sample data provided by student'); informat DclrDt ddmmyy10.; format DclrDt date9.; input PermNo DclrDt; cards; 10006 3/8/1962 10006 15/3/1962 10006 2/11/1962 10006 2/8/1963 10006 6/4/1965 10006 6/6/1968 10007 4/5/1975 10007 15/8/1975 10007 2/11/1975 10007 4/1/1976 10007 6/4/1982 10007 7/7/1982 ; run; /* Let's say that we have the above sample , as you can see in company with permno 10007 there are quartely distributions until 4/1/1976, then the company stopped paying dividend until 6/4/1982. This is by defintion an omission event (diff> 1128 days). For company 10006, we have that it stops paying dividend on 6/6/1968, and it never distribute dividend again. So if I have these cases in my sample I have to identify both of them and not only case 10006, which is easy as I keep the last observation. So my code for identifying my sample ought to have 2 cases. a. final distribution if firm never pays dividend again ---no more data---keep the last distribution b. previous to last observation -last observation >1128 then i have also an omission and keep the previous to last abservation. ("last" does not mean that firm stop paying dividend for ever. ..for example a company had a long history of distributions, it stops for 5 years and then started again -in this case i need the last observation before it stops paying. I care about the diff in days between 2 declarations and which declaration is the last one or the previous before the last declaration if the days between the last and the previous to last is over 1128 days How can i find how often each company pays dividends (how many days between declarations) and compare that number with the criteria of late payment (3-year period)? */ proc sort data=Have01; by PermNo DclrDt; run; data Have02 (label='Sample data sorted by Firm ID and divident payment date, and time from previous dividend in QTRs  and days') ; length Lag1DclrDt 8 LagDclrDays QTRs FlagLateDclr 3; format Lag1DclrDt Date9.; set Have01; by PermNo DclrDt; Lag1DclrDt=lag1(DclrDt); if First.PermNo then Lag1DclrDt=.; QTRs=intck('QTR',Lag1DclrDt,DclrDt); LagDclrDays=DclrDt-Lag1DclrDt; FlagLateDclr=(LagDclrDays gt 1128); run; proc summary data=Have02 nway; class PermNo; Var QTRs DclrDt; output   out=Have03   (    drop=_type_ _freq_    label='Useful dividend stats by firm: Usual dividend interval in QTRs, first and last dividend date'   )   median(QTRs) = MedianDclrQtrs   min(DclrDt) = FirstDcltDt   max(DclrDt) = LastDclrDt ; run; proc sql; create table Have04   (label='Sample dividend data joined to useful firm stats: usual divident interval in QTRs, late dividend and  omission flags') as   select    a.*,    b.*,    case    when     a.QTRs gt b.MedianDclrQTRs and     a.FlagLateDclr ne 1    then     1    else     0    end as FlagOmission   from    Have02 as a,    Have03 as b   where    a.PermNo eq b.PermNo ; quit; run; data Have05; set Have04(where=(FlagLateDclr eq 1 or FlagOmission eq 1)); run; ... View more

Hi Maggi. Sorry I'm answering your questions in reverse time order. Please see my previous response to your subsequent question. IF you still really need to understand the process, the SAS Institute supplies an autocall macro %DS2CSV(data=yourlibrary.yourSASdataset, runmode=b, csvfile=yourpath/yourfile.csv); If you can follow SAS Macro code, have a look at the the source code for that macro. It will be ds2csv.sas and it should be in one of your sas autocall locations for your UNIX SAS installation. on my installation it is in the SAS install files in the /core/sasmacro folder Cheers. Damien ... View more

Hi Maggi. Easiest for you I think would be for you to use the PCFILES engine of the libname statement. see http://support.sas.com/documentation/cdl/en/acpcref/63181/HTML/default/viewer.htm#n12tx6b9ilne0kn198ddk83po0lc.htm Let me know if this helps. PS I miss my SAS/UNIX combination sometimes - SAS filename statement with the pipe option pointing to an underlying unix pipe was something I found very helpful. Cheers. Damien ... View more

Start wtih the number of measurement items, double that. (two parameters per measurement item, a factor loading, and an error variance). E.g. if you have 7 latent factors with 3 measurement items each you have 21 measurement items each with 2 parameters = 42. Add the 'triangular number' ((n x (n-1))/2) associated wtih the number (n) of latent factors (and/or manifest structural items if you have a non-standard SEM). E.g. if you have 7 latent factors, (7 x 6)/2 = 21. This is the number of covariances amongst your 7 latent factors you will need to specify for the initial measurement model. subtotal = 42 + 21 = 63. Multiply this by 10 = 630. This is your recommended sample size. ... View more