Stats Firthlog warning citing R version

[galletas]

I just installed R Essentials 23 on my SPSS 23 for Windows. Also installed R 3.1.0 as it specifically requested in the plug-in installation manual.

Using these new capabilities, I ran a Firth Logistical Regression and at the end of the output got:
Warning messages:
1: package 'logistf' was built under R version 3.1.3
2: packake 'mice' was built under R version 3.1.3
3: package 'Rcpp' was built under R version 3.1.3

So, my interpretation is that I need to install R version 3.1.3 and re-run the regression? Do you agree?

[Andy W]

No, it is alright to assume that minor version differences are very unlikely to cause actual errors. (It is just a warning.) Presumably if an actual error occurred the software would warn you, but it is always caveat emptor.

[Jon K Peck]

It is not necessary to reinstall R.  That type of warning often occurs but is almost always harmless, as it is in this case.


Jon Peck (no "h") aka Kim
Senior Software Engineer, IBM
[hidden email]
phone: [phone]




From:        galletas <[hidden email]>
To:        [hidden email]
Date:        09/14/2015 08:08 AM
Subject:        [SPSSX-L] Stats Firthlog warning citing R version
Sent by:        "SPSSX(r) Discussion" <[hidden email]>

---

I just installed R Essentials 23 on my SPSS 23 for Windows. Also installed R
3.1.0 as it specifically requested in the plug-in installation manual.

Using these new capabilities, I ran a Firth Logistical Regression and at the
end of the output got:
Warning messages:
1: package 'logistf' was built under R version 3.1.3
2: packake 'mice' was built under R version 3.1.3
3: package 'Rcpp' was built under R version 3.1.3

So, my interpretation is that I need to install R version 3.1.3 and re-run
the regression? Do you agree?



--
View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Stats-Firthlog-warning-citing-R-version-tp5730601.html
Sent from the SPSSX Discussion mailing list archive at Nabble.com.

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD



===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD

[Rich Ulrich]

How does one keep track of, or respond to, such a /caveat emptor/?
Jon says this one is harmless.

Presumably, if there were a "problem,"  you might learn about it within
the next year or so if you kept informed on an R (or SPSS?) discussion list.

In practice:
 a) If the program crashes or does something really strange, you have a
starting point for tracking it down.
 b) If you are not a continuing user of the program (and reader of Lists),
you are not likely to find out about a subtle error or peculiarity.
 c) Strange problems that arise later are very rare, but not impossible.
 d) What you can and should do, is "Document it."

 I remember a recommendation I heard back in the mainframe era, that
one's published study always should name the software package and its
version/date in the report, for error-tracking in addition to other reasons
(such as - knowing default algorithms or specifications).  I believe that this
advice is regularly followed by people using limited-purpose, proprietary,
purchased software.  I think this software "warning"  should cause the
user to include the details -- "analyses performed by  XXX at version yyy,
library compiled under version zzz." 

--
Rich Ulrich

> Date: Mon, 14 Sep 2015 06:44:27 -0700
> From: [hidden email]
> Subject: Re: Stats Firthlog warning citing R version
> To: [hidden email]
>
> No, it is alright to assume that minor version differences are very unlikely
> to cause actual errors. (It is just a warning.) Presumably if an actual
> error occurred the software would warn you, but it is always caveat emptor.
>  ...

===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD

[Jon K Peck]

I say it is harmless, because R always generates this warning if the package has been built under even a minor update to R itself.  It is not a response to any particular behavior of the package or of R in general.  It is always a good idea, though, to note the exact version of the software that was used to generate any particular results.  We are add the package version information to the output from an extension that uses a package, but most of them do not currently display this information.

However, the package name is generally displayed, and you can get the exact version number by running code like this.

begin program r.
packageVersion("logistf")
end program.

using the name of the package as a string.




Jon Peck (no "h") aka Kim
Senior Software Engineer, IBM
[hidden email]
phone: [phone]




From:        Rich Ulrich <[hidden email]>
To:        [hidden email]
Date:        09/14/2015 10:56 AM
Subject:        Re: [SPSSX-L] Stats Firthlog warning citing R version
Sent by:        "SPSSX(r) Discussion" <[hidden email]>

---

How does one keep track of, or respond to, such a /caveat emptor/?
Jon says this one is harmless.

Presumably, if there were a "problem,"  you might learn about it within
the next year or so if you kept informed on an R (or SPSS?) discussion list.

In practice:
a) If the program crashes or does something really strange, you have a
starting point for tracking it down.
b) If you are not a continuing user of the program (and reader of Lists),
you are not likely to find out about a subtle error or peculiarity.
c) Strange problems that arise later are very rare, but not impossible.
d) What you can and should do, is "Document it."

I remember a recommendation I heard back in the mainframe era, that
one's published study always should name the software package and its
version/date in the report, for error-tracking in addition to other reasons
(such as - knowing default algorithms or specifications).  I believe that this
advice is regularly followed by people using limited-purpose, proprietary,
purchased software.  I think this software "warning"  should cause the
user to include the details -- "analyses performed by  XXX at version yyy,
library compiled under version zzz."  

--
Rich Ulrich

> Date: Mon, 14 Sep 2015 06:44:27 -0700
> From: [hidden email]
> Subject: Re: Stats Firthlog warning citing R version
> To: [hidden email]
>
> No, it is alright to assume that minor version differences are very unlikely
> to cause actual errors. (It is just a warning.) Presumably if an actual
> error occurred the software would warn you, but it is always caveat emptor.
>  ...

===================== To manage your subscription to SPSSX-L, send a message to LISTSERV@...(not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD

===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD