www.phidot.org

Hello everyone,
I am going through the worked examples for random effects models of Appendix D of the markbook. Following the example step by step, things seems to go well until I get the message 'Append this Model's output to database?' When I click 'yes', I get an Info message saying 'Could not import variance-covariance matrix to record'. The model is then added to the the output table but is expectedly way off in terms of deviance with empty or zero estimates.
Has someone encountered a similar behaviour, or am I missing something in the procedure ?
It is the last available version of mark, ran under the wine emulator (v 1.5.31) on a ubuntu raring (64 bits) laptop (other than the above problem, mark works pretty well under wine).
Thanks in advance for your help.
Olivier

OlivierD wrote:Hello everyone,
I am going through the worked examples for random effects models of Appendix D of the markbook. Following the example step by step, things seems to go well until I get the message 'Append this Model's output to database?' When I click 'yes', I get an Info message saying 'Could not import variance-covariance matrix to record'. The model is then added to the the output table but is expectedly way off in terms of deviance with empty or zero estimates.
Has someone encountered a similar behaviour, or am I missing something in the procedure ?
It is the last available version of mark, ran under the wine emulator (v 1.5.31) on a ubuntu raring (64 bits) laptop (other than the above problem, mark works pretty well under wine).
Thanks in advance for your help.
Olivier

So,

1\ try it on a real Windows machine. If it works there (as it should, since it works on the 3 machines I just tested it on), then

2\ the problem is with wine, and you're out of luck. There are any number of issues with wine and MARK, which is why I no longer recommend it.

OlivierD wrote:Hello everyone,
I am going through the worked examples for random effects models of Appendix D of the markbook. Following the example step by step, things seems to go well until I get the message 'Append this Model's output to database?' When I click 'yes', I get an Info message saying 'Could not import variance-covariance matrix to record'. The model is then added to the the output table but is expectedly way off in terms of deviance with empty or zero estimates.
Has someone encountered a similar behaviour, or am I missing something in the procedure ?
It is the last available version of mark, ran under the wine emulator (v 1.5.31) on a ubuntu raring (64 bits) laptop (other than the above problem, mark works pretty well under wine).
Thanks in advance for your help.
Olivier

Option now perhaps worth considering is using the native version of MARK for Linux, via the RMark front end.

See viewtopic.php?f=2&t=2688