Time Stamp inconsistency

[lisa]

On the computer program that I am still editing the data files
(--specifically --the log files) in DirectRT contain a time stamp for trials
where no response is necessary.
The trial length is 1500ms, which is what the time should be in this
column...but perhaps due to screen refreshing speeds, the amount of time
in the columns varies between 1517,1516,1518. We believe this time should be
consistent. I would like to know why is the screen refreshing at different
speeds?

Any help on this issue would be benificial.

Thank you

[jason_reed]

Hi Lisa,

You can use the MinMax column in the Style file to display the stimulus for whatever amount of time you want consistently for all of the trials where no responses are required. Let’s say that you want all the trials to appear for 1500ms and then automatically proceed to the next one. Let’s also say that you use a Style file with only one style in it and that it is label 1. Use values 0-1500! In the MinMax cell of the Style file for that single style line. Enter the value of 1 in the style column of the input file.
Let me know if this solves your problem.

Cheers.
Psychboiler

[rural.juror]

Hi. the variance that you see in the log file may be expected. In the DirectRT help, "Log File - More Data!" under Time n, it states:

"The actual display time for each stimulus screen. You might notice that the display time for the final stimulus is usually a little bit longer than the RT value (e.g., 1-15ms longer). This is because the time at which the subject responds is not the exact time at which the stimulus leaves the screen. Depending on how early into the screen refresh cycle the response occurs, the stimulus may be displayed for an additional number of milliseconds."

[jarvis]

One other note here Lisa--if DirecRT can not pull off exactly what you want on a given machine, you can often use the log file as a guide to tweak your input file. e.g., when you ask for 1500ms and you reliably get ~1516, try asking for 1490. Some machines may need this, some may not. This is definitely one of the purposes of the log file--i.e., to see how close DirectRT can get you to what you are asking. As far as the 1ms or 2ms variations, those are generally a function of rounding error. The times in the logfile can generallly be considered accurate to +/- 1ms. So sometimes you may see 1516, sometimes 1514. In such cases, 1515 is probably the best estimate. Hope that helps.