Re: News from the SPSS Community

[Eric Black]

Can somebody point me to documentation on this STATS LATENT CLASS extension?
 I tried to find something on the SPSS Community website
(www.ibm.com/developerworks/spssdevcentral) but every time I go there I am
hopelessly lost.
In particular, I would love to know whether I need AMOS (I use v21) and the
precise syntax guide for this extension.

=====================
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

[Anthony Babinec]

Eric,
A short answer is that this is an extension command that calls
the R package poLCA. So, no need for Amos.

Tony Babinec
[hidden email]


-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of
Eric Black
Sent: Thursday, February 28, 2013 11:45 AM
To: [hidden email]
Subject: Re: News from the SPSS Community

Can somebody point me to documentation on this STATS LATENT CLASS extension?
 I tried to find something on the SPSS Community website
(www.ibm.com/developerworks/spssdevcentral) but every time I go there I am
hopelessly lost.
In particular, I would love to know whether I need AMOS (I use v21) and the
precise syntax guide for this extension.

=====================
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

[Jon K Peck]

If you install the extension command, you will have a dialog box with help, and running the command with /HELP will display the syntax chart.

This command is not related to AMOS.  AMOS doesn't do this kind of model.  The command does require the R Essentials, which are also available through the site.  If you haven't installed them and their R prerequisite, be sure to download and read the installation instructions before installing the Essentials.

You can find the extension command from the SPSS Community site via
Downloads for SPSS Statistics
Extension Commands Collection

One thing that is confusing when you get to the list of commands, which spreads over several pages, is that alphabetical order (click on the Name sort heading) is case sensitive.

The SPSS Community site is at www.ibm.com/developerworks/spssdevcentral.

The rather unfriendly link direct to this particular extension command is
https://www.ibm.com/developerworks/mydeveloperworks/files/app?lang=en#/person/270002VCWN/file/fc36c7be-3e72-4f8f-b452-4413bff8b406

For  more on the statistical methods involved, see
http://www.jstatsoft.org/v42/i10/


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




From:        Eric Black <[hidden email]>
To:        [hidden email],
Date:        02/28/2013 10:48 AM
Subject:        Re: [SPSSX-L] News from the SPSS Community
Sent by:        "SPSSX(r) Discussion" <[hidden email]>

---

Can somebody point me to documentation on this STATS LATENT CLASS extension?
I tried to find something on the SPSS Community website
(www.ibm.com/developerworks/spssdevcentral) but every time I go there I am
hopelessly lost.
In particular, I would love to know whether I need AMOS (I use v21) and the
precise syntax guide for this extension.

=====================
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

[Eric Black]

Thanks Jon

One thing that is confusing to me:  When I try to install the RPlugin, I get
the message that "The IBM Statistics - Essentials for R 21 64bit requires a
folder in which a IBM SPSS Statistics 21.0 application is installed".  When
I point to "C:\Program Files\IBM\SPSS\Statistics\21", I do get an error
message (R 2.14 was not found in this location).

Yes, it wasn't - I have R 2.15 but that is not in the folder where SPSS is
installed.  Where exactly do I need to point the folder - to where SPSS is
installed or where R is installed?

=====================
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]

You need to install the version of R as specified in the installation instructions for the R Essentials.  For Statistics 21, that is R 2.14.x.  If you do that, the R Essentials installer will find it.

You can have multiple R versions installed simultaneously, but each version of Statistics requires a specific R version due to the technology required to do the integration.

HTH,


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




From:        Eric Black <[hidden email]>
To:        [hidden email],
Date:        02/28/2013 12:33 PM
Subject:        Re: [SPSSX-L] News from the SPSS Community
Sent by:        "SPSSX(r) Discussion" <[hidden email]>

---

Thanks Jon

One thing that is confusing to me:  When I try to install the RPlugin, I get
the message that "The IBM Statistics - Essentials for R 21 64bit requires a
folder in which a IBM SPSS Statistics 21.0 application is installed".  When
I point to "C:\Program Files\IBM\SPSS\Statistics\21", I do get an error
message (R 2.14 was not found in this location).

Yes, it wasn't - I have R 2.15 but that is not in the folder where SPSS is
installed.  Where exactly do I need to point the folder - to where SPSS is
installed or where R is installed?

=====================
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

[Ryan]

For the record, AMOS most certainly can fit a latent class model. 

Ryan 

On Feb 28, 2013, at 1:12 PM, Jon K Peck <[hidden email]> wrote:

If you install the extension command, you will have a dialog box with help, and running the command with /HELP will display the syntax chart.

This command is not related to AMOS.  AMOS doesn't do this kind of model.  The command does require the R Essentials, which are also available through the site.  If you haven't installed them and their R prerequisite, be sure to download and read the installation instructions before installing the Essentials.

You can find the extension command from the SPSS Community site via
Downloads for SPSS Statistics
Extension Commands Collection

One thing that is confusing when you get to the list of commands, which spreads over several pages, is that alphabetical order (click on the Name sort heading) is case sensitive.

The SPSS Community site is at www.ibm.com/developerworks/spssdevcentral.

The rather unfriendly link direct to this particular extension command is
https://www.ibm.com/developerworks/mydeveloperworks/files/app?lang=en#/person/270002VCWN/file/fc36c7be-3e72-4f8f-b452-4413bff8b406

For  more on the statistical methods involved, see
http://www.jstatsoft.org/v42/i10/


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




From:        Eric Black <[hidden email]>
To:        [hidden email],
Date:        02/28/2013 10:48 AM
Subject:        Re: [SPSSX-L] News from the SPSS Community
Sent by:        "SPSSX(r) Discussion" <[hidden email]>

---

Can somebody point me to documentation on this STATS LATENT CLASS extension?
I tried to find something on the SPSS Community website
(www.ibm.com/developerworks/spssdevcentral) but every time I go there I am
hopelessly lost.
In particular, I would love to know whether I need AMOS (I use v21) and the
precise syntax guide for this extension.

=====================
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

[Eric Black]

Great Jon, that worked.  Maybe I am the only one but I find those gems to be
well hidden on the IBM site.  I just wish there were a more accessible way
to find and install those great add-ons.

Om another note, how does memory allocation work with regard to the R-PlugIn.

When I run an LC model, I get the following message:

  Error: cannot allocate vector of size 1.1 Gb
  In addition: Warning messages:
  1: In y[[j]][cbind(c(1:N), ymat[, j])] <- 1 :
    Reached total allocation of 2047Mb: see help(memory.size)
  2: In y[[j]][cbind(c(1:N), ymat[, j])] <- 1 :
    Reached total allocation of 2047Mb: see help(memory.size)
  3: In y[[j]][cbind(c(1:N), ymat[, j])] <- 1 :
    Reached total allocation of 2047Mb: see help(memory.size)
  4: In y[[j]][cbind(c(1:N), ymat[, j])] <- 1 :
    Reached total allocation of 2047Mb: see help(memory.size)

I assume this is a memory issue with R and not with SPSS, right?

=====================
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

[David Marso]

"memory issue with R and not with SPSS, right? "
INDEED!!!!

"Maybe I am the only one but I find those gems to be
well hidden on the IBM site."
--
Nope, not alone.  IMNSHO:  The IBM site is a complete f'ing disaster!
After leaving that site upon fruitless searching for anything at all
I have the urge to self administer an industrial-bleach cranial enema.
I used to think the SPSS site sucked until I saw what the IBM peeps barfed out of their orifice!

OTOH: sometimes I feel a need to do penance so I surf around there simply for the masochistic fulfillment.
--

Eric Black wrote

Great Jon, that worked.  Maybe I am the only one but I find those gems to be
well hidden on the IBM site.  I just wish there were a more accessible way
to find and install those great add-ons.

Om another note, how does memory allocation work with regard to the R-PlugIn.

When I run an LC model, I get the following message:

  Error: cannot allocate vector of size 1.1 Gb
  In addition: Warning messages:
  1: In y[[j]][cbind(c(1:N), ymat[, j])] <- 1 :
    Reached total allocation of 2047Mb: see help(memory.size)
  2: In y[[j]][cbind(c(1:N), ymat[, j])] <- 1 :
    Reached total allocation of 2047Mb: see help(memory.size)
  3: In y[[j]][cbind(c(1:N), ymat[, j])] <- 1 :
    Reached total allocation of 2047Mb: see help(memory.size)
  4: In y[[j]][cbind(c(1:N), ymat[, j])] <- 1 :
    Reached total allocation of 2047Mb: see help(memory.size)

I assume this is a memory issue with R and not with SPSS, right?

=====================
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]

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




From:        David Marso <[hidden email]>
To:        [hidden email],
Date:        03/01/2013 03:50 PM
Subject:        Re: [SPSSX-L] News from the SPSS Community
Sent by:        "SPSSX(r) Discussion" <[hidden email]>

---

"memory issue with R and not with SPSS, right? "
INDEED!!!!

"Maybe I am the only one but I find those gems to be
well hidden on the IBM site."
--
Nope, not alone.  IMNSHO:  The IBM site is a complete f'ing disaster!
After leaving that site upon fruitless searching for anything at all
I have the urge to self administer an industrial-bleach cranial enema.
>>>PLEASE DO.
I used to think the SPSS site sucked until I saw what the IBM peeps barfed
out of their orifice!
I fail to see why it is so hard.  If you start at the SPSS Community URL, you see an immediate link to
Downloads for SPSS Statistics.
On that page, you see lists of the plugin items, a list of collections that are divided into logical groups - Extension Commands, Utilities, Statistical Tools, etc and those collection items each have a brief description.

The books and articles section has a description of each item.

The forums are organized into logical areas.

The bookmark list aggregates important places that might be of interest.


I agree that the whole site is a little clunky, but the site is certainly not a disaster.


OTOH: sometimes I feel a need to do penance so I surf around there simply
for the masochistic fulfillment.
--


Eric Black wrote
> Great Jon, that worked.  Maybe I am the only one but I find those gems to
> be
> well hidden on the IBM site.  I just wish there were a more accessible way
> to find and install those great add-ons.
>
> Om another note, how does memory allocation work with regard to the
> R-PlugIn.
>
> When I run an LC model, I get the following message:
>
>   Error: cannot allocate vector of size 1.1 Gb
>   In addition: Warning messages:
>   1: In y[[j]][cbind(c(1:N), ymat[, j])] <- 1 :
>     Reached total allocation of 2047Mb: see help(memory.size)
>   2: In y[[j]][cbind(c(1:N), ymat[, j])] <- 1 :
>     Reached total allocation of 2047Mb: see help(memory.size)
>   3: In y[[j]][cbind(c(1:N), ymat[, j])] <- 1 :
>     Reached total allocation of 2047Mb: see help(memory.size)
>   4: In y[[j]][cbind(c(1:N), ymat[, j])] <- 1 :
>     Reached total allocation of 2047Mb: see help(memory.size)
>
> I assume this is a memory issue with R and not with SPSS, right?
>
> =====================
> To manage your subscription to SPSSX-L, send a message to

> LISTSERV@.UGA

>  (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





-----
Please reply to the list and not to my personal email.
Those desiring my consulting or training services please feel free to email me.
--
View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Re-News-from-the-SPSS-Community-tp5718302p5718337.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

[Eric Black]

Not sure if this is the right place to express the grievances, but to me
what is confusing is where should I look for a download?  Once you know
which path to go, it does make sense, but the issue I have is that those who
are new to the site don't know that.  I assume, just as me, they would try
the search function and that simply leads nowhere.  In this case, when I
type in "Stats Latent Class", I get results but shouldn't the actual
download be somewhere in the top of that?

Then:  Why do I need to get certain things (like Essentials and Plugin for
R) from IBM Passport and others directly from the DevCentral?  I can see how
these are separate sections for IBM but isn't that division irrelevant to
the (paying!) user?

Next: there are two sections on the downloads site:
1. Python modules for use in writing Python programs
2. Statistical Tools for use with IBM SPSS Statistics. May be standard
syntax, Python programs, R programs, or other material.

As a consumer/user, I don't see the distinction between Python modules and
Python programs.

I want to reiterate:  I think there are great extensions here and I am very
appreciative of the work that you and IBM put in it but I feel like for a
licensed product that is not really cheap, it is very, very cumbersome to
find those things, especially if I compare it with Stata or even with R.
Installing extensions/programs in those two is a breeze IMHO and I wish the
same were true for SPSS.

=====================
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

[Tom1234]

Essentially IBM should be given a slap for their heinous crimes against the principles of HCI and user accessibility; the layout and methods that abomination of a site uses simply baffle and confound all those who enter the labyrinth they call their website.
Angrymarine aside here are my thoughts on what you've touched upon:

Eric Black wrote

search function and that simply leads nowhere.  In this case, when I
type in "Stats Latent Class", I get results but shouldn't the actual
download be somewhere in the top of that?

You're quite right Eric, the natural instinct of any user attempting to gain information about spss would use the search bar, and upon searching a series of nonsensical articles about IBM's world domination are displayed (slight exaggeration but it's in the spirit of it) and maybe a few items relating (sometimes very vaguely) to the item you had searched for. Unlike most sites there does not seem to be a strong link between the download links for SPSS's additional packages/plugins and the search bar.

Eric Black wrote

Why do I need to get certain things (like Essentials and Plugin for
R) from IBM Passport and others directly from the DevCentral

Another good point, surely a single, all-purpose SPSS download centre would be more useful, listing all available extensions for the version you've selected/searched for.

Eric Black wrote

As a consumer/user, I don't see the distinction between Python modules and
Python programs.

According to Python's website, 'A module is a file containing Python definitions and statements'. A program usually uses several modules and internal scripts to perform a task.

I personally think SPSS should have a built in extension finder/installer, much like the one that Netbeans and many other programs employ, where all plugins are uploaded to a central location, and each time you start SPSS (on a new start screen) it would notify you if any of your current extensions require updating. There would also be an 'Extensions library' dialogue, where you could manage installed plugins, search for new plugins and install them without having to leave the application and wade through dozens of pages before finally, exhausted, reaching your destination.

IMHO this would make for a better SPSS and much happier and less confused users.

[Jon K Peck]

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




From:        Eric Black <[hidden email]>
To:        [hidden email],
Date:        03/04/2013 06:21 AM
Subject:        Re: [SPSSX-L] News from the SPSS Community
Sent by:        "SPSSX(r) Discussion" <[hidden email]>

---

Not sure if this is the right place to express the grievances, but to me
what is confusing is where should I look for a download?  Once you know
which path to go, it does make sense, but the issue I have is that those who
are new to the site don't know that.  I assume, just as me, they would try
the search function and that simply leads nowhere.  In this case, when I
type in "Stats Latent Class", I get results but shouldn't the actual
download be somewhere in the top of that?
>>>We are limited to what we can do with the standard developerWorks system, and we are constrained by IBM Legal requirements.  developerWorks is due for a major upgrade shortly, so that may allow some improvements.  However, if you want to search for files to download,  you need to change the droplist at the top to "Files".  Then you can search for items.  Since this search seems to be based mainly on file names, you would search for Stats_Latent_Class in this case, which will show that item.  Looking in the Collections works better if you don't know the name.

Then:  Why do I need to get certain things (like Essentials and Plugin for
R) from IBM Passport and others directly from the DevCentral?  I can see how
these are separate sections for IBM but isn't that division irrelevant to
the (paying!) user?
>>>Items are in different places because of legal requirements.  The license and terms vary for different items.  Everything on developerWorks is governed by the dW Terms & Conditions.  We can't put the R materials on dW, because they are governed by the GPL, and proprietary items can't go there because the license is different.  I wish everything was unified, but we do try to at least put links to the right places.

Next: there are two sections on the downloads site:
1. Python modules for use in writing Python programs
2. Statistical Tools for use with IBM SPSS Statistics. May be standard
syntax, Python programs, R programs, or other material.

As a consumer/user, I don't see the distinction between Python modules and
Python programs.
>>>Some items are actually cross listed when they fit into more than one category.  The Python modules in that collection are modules with useful utilities that are not part of the core plugin.  For example, there is an extendedTransforms.py module that you might want to use when you are writing a Python program, but it is not distributed with the plugin/Essentials.  These items could be from IBM or contributed by third parties.  We can update these items off-cycle, so new features or bug fixes don't have to wait for the next fixpack or major release.

I want to reiterate:  I think there are great extensions here and I am very
appreciative of the work that you and IBM put in it but I feel like for a
licensed product that is not really cheap, it is very, very cumbersome to
find those things, especially if I compare it with Stata or even with R.
Installing extensions/programs in those two is a breeze IMHO and I wish the
same were true for SPSS.
>>>So do I, and we are hoping for future improvements.

=====================
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

[Albert-Jan Roskam]

>Next: there are two sections on the downloads site:

>1. Python modules for use in writing Python programs
>2. Statistical Tools for use with IBM SPSS Statistics. May be standard
>syntax, Python programs, R programs, or other material.
>
>As a consumer/user, I don't see the distinction between Python modules
and
>Python programs.
>>>>Some items are actually cross listed when
they fit into more than one category.  The Python modules in that
collection are modules with useful utilities that are not part of the core
plugin.  For example, there is an extendedTransforms.py module that
you might want to use when you are writing a Python program, but it is
not distributed with the plugin/Essentials.  These items could be
from IBM or contributed by third parties.  We can update these items
off-cycle, so new features or bug fixes don't have to wait for the next
fixpack or major release.

I never tried this, but is it possible to use pip to install or update Python packages hosted on developer works? I checked http://www.pip-installer.org/en/latest/usage.html#description and it seems to be possible.
Another thing that would be useful for many users (but not me :-( is an auto-update feature hot/interim fixes and fixpacks.

Regards,
Albert-Jan

=====================
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]

You would need the exact url from dW.  For spe files, this is a binary format, and I doubt that pip could be persuaded to work with it, but one could certainly write a utility to do this, since spe files are a zip format, and installation is just a matter of placing files in the right place.

If you go to a specific file, you can request automatic notification when the item is updated.


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




From:        Albert-Jan Roskam <[hidden email]>
To:        Jon K Peck/Chicago/IBM@IBMUS, "[hidden email]" <[hidden email]>,
Date:        03/04/2013 11:52 AM
Subject:        Re: [SPSSX-L] News from the SPSS Community

---

>Next: there are two sections on the downloads site:

>1. Python modules for use in writing Python programs
>2. Statistical Tools for use with IBM SPSS Statistics. May be standard
>syntax, Python programs, R programs, or other material.
>
>As a consumer/user, I don't see the distinction between Python modules
and
>Python programs.
>>>>Some items are actually cross listed when
they fit into more than one category.  The Python modules in that
collection are modules with useful utilities that are not part of the core
plugin.  For example, there is an extendedTransforms.py module that
you might want to use when you are writing a Python program, but it is
not distributed with the plugin/Essentials.  These items could be
from IBM or contributed by third parties.  We can update these items
off-cycle, so new features or bug fixes don't have to wait for the next
fixpack or major release.

I never tried this, but is it possible to use pip to install or update Python packages hosted on developer works? I checked http://www.pip-installer.org/en/latest/usage.html#description and it seems to be possible.
Another thing that would be useful for many users (but not me :-( is an auto-update feature hot/interim fixes and fixpacks.

Regards,
Albert-Jan