Running study from a network server
Running study from a network server
[edited from support email]
I am planning to run my MediaLab study off of a network server in our lab and will have up to seven people participating at any one time. Is there potential for problems with the collection of data from more than one computer entering the one server at the same time? I have measures using DirectRT as well.
In the same way that you can write data to a folder on another computer, you can also run an experiment from another computer. To do this, you must “map the drive” of the other computer which contains the experiment (see "Run" or "Network" in user's guide for more detail). Once you have done this, you can run the experiment as if it were located locally. From the “Run” menu, choose “Select and Run an Experiment” or “Select and Run a Single Questionnaire” and browse to the experiment located on the mapped drive (e.g., “h:\experiment1\myexperiment.exp”). This way multiple computers can access the same experiment from a single computer. As with writing data to another computer (see guide), it is recommended that you specify a “unique machine code” in Preferences since all the data from all machines will be writing to the data folder on this single computer and you will want to avoid data collisions. This option may not be suitable for experiments that require a lot of bandwidth (e.g., those that use a lot of video and sound) or for experiments using many computers on a slow network.
If you are using a responses.xls file, see also:
www.empirisoft.com/support/showthread.php?t=516
The computer that runs my DirectRT experiments is connected to a network. Does this affect the precision of the reaction-time measurements?
Some people have reported that they have been able to reduce minor variability in display timing by disconnecting from the network during DirectRT sessions. I would first check your log data files to see if you have any variability to deal with. If not, then it's not really an issue. If so, then try running a few sessions with the network cable pulled out and see what happens to that variability.
Hope that helps,
-Blair
