SPSS warning when attempting mediation analysis using PROCESS macro

[DebHyden]

Hi, I've run a mediation analysis using PROCESS, but I only get partial results with the following warning ">MATRIX could not allocate memory for an object.  Reduce problem size, or
>release unused matrices using RELEASE statement.  Use DISPLAY statement to
>list all allocated objects.  .
>Execution of this command stops."

Does anyone know how I can rectify this problem? I'm not familiar with any of the solutions offered by the warning, and don't see how I can reduce the problem size, as I'm only using one predictor, one mediator and one outcome variable for the analysis.

Finally, is it possible that this particular warning is generated because my variable names are too long?

[Kirill Orlov]

MATRIX session in SPSS (which your macro uses) is dependent on RAM memory of your computer. If your data is large it may exceed your computer capacity. How many RAM do you have?

RELEASE statement clears unnecessary data from MATRIX memory and may help sometimes. But it needs for you or somebody to dig inside the syntax to find out places where it can be inserted. If the macro is compex and lengthy it might take time an expertise.

Can you contact with the macro's author (Hayes?) and report him the problem? Addressing the developer is the best way to go.

===================== 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 Peck]

For a start, you could insert some DISPLAY statements in the MATRIX code to see where the memory is going.  That might be useful to Prof Hayes as well.

On Tue, Jun 7, 2016 at 1:25 PM, Kirill Orlov <[hidden email]> wrote:

MATRIX session in SPSS (which your macro uses) is dependent on RAM memory of your computer. If your data is large it may exceed your computer capacity. How many RAM do you have?

RELEASE statement clears unnecessary data from MATRIX memory and may help sometimes. But it needs for you or somebody to dig inside the syntax to find out places where it can be inserted. If the macro is compex and lengthy it might take time an expertise.

Can you contact with the macro's author (Hayes?) and report him the problem? Addressing the developer is the best way to go.

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

[DebHyden]

Hi,

Firstly, thank you both for your suggestions. I seem to have solved the problem by shortening the variable names. Although Andrew Hayes (the developer) does warn against the use of long variable names, there is the option to allow them, providing they do not conflict.

Apparrently using the long name option really is not advisable.  I will however, contact him to let him know what happened so he may update the code if necessary.
As for PC RAM capability, I have been assured that it's a very high spec machine, with plenty of computing capability, so I doubt that's a problem.

p.s. It's not my own computer, but owned by a government agency.

[Kirill Orlov]

One should not use variable names longer than 8 characters whenever he uses MATRIX-END MATRIX session. It not an error but is hazardous. The developer of your macro ought to have warned in the documentation against using long variables names for the variables input to the macro.

09.06.2016 7:17, DebHyden пишет:

Hi,

Firstly, thank you both for your suggestions. I seem to have solved the
problem by shortening the variable names. Although Andrew Hayes (the
developer) does warn against the use of long variable names, there is the
option to allow them, providing they do not conflict.

Apparrently using the long name option really is not advisable.  I will
however, contact him to let him know what happened so he may update the code
if necessary.
As for PC RAM capability, I have been assured that it's a very high spec
machine, with plenty of computing capability, so I doubt that's a problem. 

p.s. It's not my own computer, but owned by a government agency.



--
View this message in context: http://spssx-discussion.1045642.n5.nabble.com/SPSS-warning-when-attempting-mediation-analysis-using-PROCESS-macro-tp5732331p5732356.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

[Jon Peck]

From the CSR...

A matrix variable name follows the same rules as those applicable to an ordinary IBM SPSS Statistics

variable name, except matrix variable names cannot exceed 8 bytes.

On Thu, Jun 9, 2016 at 7:06 AM, Kirill Orlov <[hidden email]> wrote:

One should not use variable names longer than 8 characters whenever he uses MATRIX-END MATRIX session. It not an error but is hazardous. The developer of your macro ought to have warned in the documentation against using long variables names for the variables input to the macro.

09.06.2016 7:17, DebHyden пишет:

Hi,

Firstly, thank you both for your suggestions. I seem to have solved the
problem by shortening the variable names. Although Andrew Hayes (the
developer) does warn against the use of long variable names, there is the
option to allow them, providing they do not conflict.

Apparrently using the long name option really is not advisable.  I will
however, contact him to let him know what happened so he may update the code
if necessary.
As for PC RAM capability, I have been assured that it's a very high spec
machine, with plenty of computing capability, so I doubt that's a problem. 

p.s. It's not my own computer, but owned by a government agency.



--
View this message in context: http://spssx-discussion.1045642.n5.nabble.com/SPSS-warning-when-attempting-mediation-analysis-using-PROCESS-macro-tp5732331p5732356.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

--

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