plf515 Tracker

How to put SAS on Demand for Academics in an icon on my screen?

plf515 — Fri, 11 Jun 2021 21:16:28 GMT

After using SAS professionally for decades, I've retired. But I'm giving some presentations with SAS at regionals. So, I downloaded SODA Studio.

What I am wondering is if I can put a link to this on my screen, rather than having to go through the link in the e-mail that SAS sent to me.

Thanks

Peter

Re: What part of The Quad are you looking forward to visiting?

plf515 — Sat, 25 Jan 2020 12:49:37 GMT

I will certainly visit SAS Books to see if there's anything new to put on my "to be read" pile.

I don't think there's a schedule of SAS presentations yet, but I am sure there will be some I want to see - and if they are on your topic, they are always worthwhile.

The various vendors often have interesting things. I may see if any have work for me, as well.

Mostly, though, i look forward to the sort of casual interaction that happens at SAS conferences. For instance, I want to chat with SAS people who work in statistics and/or graphics . I have some comments, questions and suggestions (and praise).

Re: Structural Equation Modelling

plf515 — Wed, 22 Jan 2020 12:59:54 GMT

Does this usage note from SAS help?

http://support.sas.com/kb/22/529.html

In a CATX function, how can I use a paragraph break as a separator?

plf515 — Wed, 22 Jan 2020 12:42:15 GMT

Using SAS 9.4 on Windows

I have code that uses CATX to concatenate two variables with a comma in between:

stdy_sub = catx(', ', study, substance);

I would like the separator to be a paragraph break with some spaces after the break so that the result is something like:

studyname
substancename

Re: SAS GF 2020 Sooo...What sessions do you have in your agenda?

plf515 — Tue, 21 Jan 2020 19:37:41 GMT

Aside from my own HOW on "You are using PROC GLM too much (and what you should be using instead)" I will mostly be in analytics:statistics sessioins, with some data visualization sessions as well.

I also want to plan lots of time to hang out.

My tip to first timers: Plan LOTS of time to hang out and talk to people. You can read the papers at home - and if you have questions, most presenters will answer them via e-mail But you can't hang around with these people at home.

Yeah, yeah, we're a bunch of geeky introverts (well, a LOT of us are) but .... you'll be among your peers.

If someone is giving a talk you are interested in, by all means go the session. But find them before or after. Tell them you're interested.

Re: Forest plots in SAS

plf515 — Thu, 16 Jan 2020 14:40:59 GMT

Thanks! I had seen some of them but not all, despite googling.

Re: SAS Global Forum 2020 on GitHub!

plf515 — Thu, 16 Jan 2020 14:39:48 GMT

Hi Chris

I am willing to put my code on GitHub, but am completely new to it and have no idea how to do it.

You said that you could do it for us, which would be great. Should I send it to you? If so, at what e-mail?

Thanks!

Peter

Forest plots in SAS

plf515 — Tue, 07 Jan 2020 13:55:21 GMT

Hello

I am not sure if this is inappropriate here ... please let me know.

I am trying to make a rather complicated forest plot in SAS and having a lot of trouble -- more, I think, than is appropriate for a list like this.

I am interested in consulting with an expert on SGPLOT and probably the GTL to make a better forest plot program for SAS. This could be for pay or just to write some SAS papers (or maybe both).

Is this a good place to look for such a person and, if not, what would be a good place?

Thanks.

PROC QUANTREG question about estimated values, intercept and actual quantile

plf515 — Thu, 02 Jan 2020 20:35:37 GMT

Hi and HNY

I am using SAS 9.4 TS 1M5 on Windows.

If I run

title;
proc quantreg ci=sparsity/iid algorithm=interior(tolerance=5.e-4)  data=sashelp.bweight order=internal plots = none;
   model Weight = MomWtGain/quantile = (0.25, .5, .75);
run;

then I get some slightly puzzling results.

For each quantile (25, 50, 75) we have

a. The actual quantiles which are 3062.0 3402.0 and 3720.0

b, The predicted values at mean, which are 3068.02, 3394.93 and 3719.04

c. The intercepts which are 3062.0, 3389.46 and 3714.0

Sometimes they are exactly the same, sometimes slightly different.

Why is that?

Meta analysis

plf515 — Thu, 02 Jan 2020 15:02:32 GMT

SAS should write a PROC to do meta-analysis.

1. There's a shortage of good software to do this, so SAS would add to its market.

2. There are lots of articles on it, but they often contradict each other, so the expertise of the people at SAS Statistics would be valuable.

3. LOTS of people do meta-analysis.

4. There is a book on doing this in SAS but it is very dated.

5. There are some articles on doing it with PROC MIXED but they are scattered, not fully coherent, use old versions of SAS (I found good one that uses v 6). And no source that I found has a complete method. E.g. one that includes combining effects, forest plots, diagnostics and so on.

Equivalent of "make" statement in SAS PROC MIXED

plf515 — Tue, 31 Dec 2019 16:49:25 GMT

Hello

I am using SAS 9.4 1M5 on Windows

I am doing a meta-analysis (NOTE: SAS should write a PROC META) and found an old paper that uses this statement in PROC MIXED

proc mixed data = DATASET;
 class  study;
  model  EFFsize = /p s;
  random study/gdata = g s;
  repeated diag;
  make 'predicted' out = predv;
  make 'SolutionR' out = randv;
run;

where Effsize is the effect size and g is a diagonal data set.

However, the code is from version 6 and (unusually for SAS) it doesn't seem to be fully backward compatible.

In particular "p" causes an error and the two MAKE statements cause warnings.

When I try to fix the error caused by p, the first MAKE statement stops working.

Can anyone translate the code to fit SAS 9.4?

Thanks

Advantages of higher or lower K1 in PROC ROBUSTREG

plf515 — Tue, 17 Dec 2019 17:50:21 GMT

In SAS PROC ROBUSTREG you can set K1, which affects the efficiency of the procedure. But I didn't see anything in the documentation about exactly what "efficiency" means nor about the advantages of changing K1 from its default value.

Any insights would be appreciated.

How to do meta-regression in SAS

plf515 — Wed, 11 Dec 2019 17:43:27 GMT

Hello

I am using SAS 9.4 on Windows.

I want to do meta-regression. I have the effect size (standardized differences of means), their standard errors, sample sizes and several covariates.

I'm familiar with PROC GLM and PROC MiXED and am aware that those are the main tools for meta-regression, but I wanted to get any tips or references that people had that are about meta-regression in particular, as opposed to more general regression problems

Thanks

Re: Discontinuous growth curve analysis - NLIN? GLIMMIX? Spline? MM?

plf515 — Wed, 30 Oct 2019 17:08:49 GMT

Thanks. That one uses a single "subject" and the only covariate is time (and it has a lot of times). I have about 70 subjects and only 5 times - so, I'm wondering if NLIN would be better than GLIMMIX

Discontinuous growth curve analysis - NLIN? GLIMMIX? Spline? MM?

plf515 — Wed, 30 Oct 2019 16:38:42 GMT

Using SAS 9.4 1M6 on Windows.

I have been hired to replace a statistician who left a project.

One of my tasks is to help my boss revise a paper that the earlier statistician worked on. I can't get into specifics (it hasn't been published yet) but, briefly, the dependent variable was measured at 5 time points - 3 at or before baseline and 2 after treatment. The previous statistician used what she called "discontinuous growth curve analysis" but cited no source and the results tables don't make it clear what exactly she did. My boss is not at all statistically inclined and, therefore, can't really help with this. The reviewers did not object to the previous analysis.

I'm thinking that this sounds like a model where there were multiple slopes fit to the results at different time points and that it would be expected that the baseline values would (on average) show no change while the slope from the 3rd baseline value to the first followup and from the first followup to the second would be expected to be nonzero.

I've not done this before in SAS. I'm not sure how to model it. I am considering:

PROC NLIN (seems right, but I have not used NLIN before)

PROC GLIMMIX with a spline effect, restricting it to one knot.
PROC GLIMMIX with an MM effect (I just found out about these).

Any thoughts are welcome

Peter

Output from a SPLINE effect with DETAILS - how to interpret?

plf515 — Wed, 11 Sep 2019 19:12:00 GMT

I am running SAS 9.4 TS 1M6 on Windows.

In GLIMMIX, I had the following code:

 effect splineaccent = spline(accent/naturalcubic details) ;

the results included:

Knots for Spline Effect splineaccentKnot Number Accent123

which, to me, indicates 3 knots at 25, 50 and 75 but also

Basis Details for Spline Effect splineaccentColumn Power Break Knot123

0
1
3	25

which I am not sure of. And the solutions for fixed effects showed 3 effects for spline. Are those 0-25, 25-50, 50-75 and an unshown (presumably 0) effect for 75 and up? Or is it the other way round (with the 0 effect for 0-25)? Or something else?

Re: How to use correlations in PROC CALIS for confirmatory factor analysis?

plf515 — Thu, 05 Sep 2019 21:24:24 GMT

I found the answer from SAS tech support. There is no need to modify the code more than what I did.

But they suggested perhaps treating the IVs as continuous rather than ordinal.

Peter

Re: How to use correlations in PROC CALIS for confirmatory factor analysis?

plf515 — Thu, 05 Sep 2019 21:13:14 GMT

Thanks Paige.

I'm not sure.

At least for me, the documentation for CALIS does not meet SAS' usual high standard for quality.

How to use correlations in PROC CALIS for confirmatory factor analysis?

plf515 — Thu, 05 Sep 2019 20:17:46 GMT

I am using SAS 9.4 1M6 on a Windows 10 machine.

I am doing a confirmatory factor analysis with ordinal variables (Likert type scales, some 1-5 and some 1-7).

I first tried doing this with PROC CALIS and an appropriate method. But method = wls gave an error that the weight matrix was singular or not positive definite (I think that my N of 202 is not enough) and method = mlm took a long time to run and produced output that was hard to interpret, so, I decided to work directly with the covariance matrix.

My old CALIS code was

proc calis data = new method = mlm;

factor
  F1 ---> q42 q52 q62 q72,
  F2 ---> q41 q51 q61 q71,
  F3 ---> q44 q54 q64 q74,
  F4 ---> q43 q53 q63 q73,
  F5 ---> q46 q56 q66 q76,
  F6 ---> q45 q55 q65 q75,
%.  ETC;


run;

I have now made a correlation data set using:

/*CORRELATIONS FOR INPUT TO CALIS*/
proc corr data = new spearman out = johncorr;
var q1-q80;
run;

but I am not quite sure how to modify my code to make it work. Do I just change the PROC CALIS statement to have data = johncorr? Or do I need to do more? Should I still include all the factor statements?

Re: Option for PROC TTEST with PAIRED data

plf515 — Fri, 30 Aug 2019 13:28:22 GMT

But then the graph is wrong, showing test2 on the left and test1 on the right.