|
|
|
Junior Member
 
Group: Forum Members
Last Login: 7/18/2007 6:31:28 PM
Posts: 11,
Visits: 112
|
|
| Hi - I'm building a protocol with 180 items, which I'd like presented in 9 runs of 20 items each. I am hoping to put the entire task into a single es file, with each run separated by a rest period that's built into the protocol. This allows me to have all of the items randomized to both condition and run on a subject-by-subject basis. I've figured out how to do it, and it's running smoothly. Hooray!  So here are the issues: - I'm concerned that my subjects may need a longer break than I've built in (right now it's up to 5 minutes between runs, though a new run can be started earlier with particular keypresses). Is there a way I can make the duration of the text object that marks the break indefinite? Or do I just put in a huge number so that it can wait up to, say, 20 minutes (like if one of my subjects with mobility impairments has to go to the bathroom)? - My other concern is that someone may just run out of steam in the middle and want to quit early. Of course, as per the rights of human subjects, I need to let them. If I build my protocol like this, with the whole thing in one file, is there a way to end the experiment mid-stream without losing the data already collected, so it's not a total loss? Thanks for any insight you may have! JoAnn.
|
|
|
|
|
Forum MVP
Group: Administrators
Last Login: Yesterday @ 9:26:59 AM
Posts: 576,
Visits: 1,254
|
|
| For your first part, you can specify "infinite" for the Duration this is either a duration of -1 or choose infinite from the drop down choices. Note that if you specify infinite, then you require a keypress to move on or it will stay there...well infinitely. You can break an experiment at any time by using Ctrl+Alt+Shift which will cause a prompt. The runtime is consistently generating a .txt file that you can use in E-Recovery to get the data up to the trial where the prompt happened. This is brute force and only use if you really need it (or during debugging). A smoother approach is to use GetUserBreakState function which will be non-zero when you press the Ctrl+Shift at the same time. Your experiement must explicitly look for this and then take action, which typically means calling List.Terminate on the trial/block lists so that the experiment gracefully ends. If there is not a sample on the web support site, then create a ticket and I know there are some examples to doing this. -Brandon
|
|
|
|
|
Junior Member
Group: Forum Members
Last Login: 7/18/2007 6:31:28 PM
Posts: 11,
Visits: 112
|
|
| Thanks, Brandon! I (clearly) didn't know about the "infinite" option, and hadn't seen it in the dropdown menu, so will try to figure it out. For ending early, I've been using ctrl-alt-shift to abort, but didn't realize that the .txt file is maintained, so if I can't figure out the more graceful option I at least know I can go that way. Appreciate your help! JoAnn.
|
|
|
|
