Millisecond Forums

Thanks Dave. Will do.

You can do all of this using the approach I outlined. You just need to use variables, i.e., store the respecitve trial's cuttoff in a <values> entry, use that in /iscorrectresponse and correct latency accordingly (which boils down to a simple subtraction: latency measured from beginning of trial - trial's cutoff).

Dave,

Sorry, I think in application so this may not make sense. The application is for football lineman to practice responding to the snap of a football in a video file (there are audio files as well). Because each video/sound is a different length, I made each video it's own trial. I wanted to know the latency from the snap of the ball in the video to measure improvement over time. If I am thinking about it correctly, it would be impossible to measure improvement using the start of the videos because the videos are presented randomly and vary in length. I may have an average response of 4000 ms and a second person may have an average response of 3000 ms simply because the second person viewed shorter videos. In this sense, it would be impossible to judge improvement of response without measuring from the point  of the snap of the ball in the video. I'm not concerned about the negative latencies because those are mistakes (aka false start penalties). Hope this makes sense. 

You'll want to set /responsetime to 0, set the video's /playthrough to true, use /iscorrectresponse to check if latency is greater than or equal to the desired point in time (e.g, 3900). I don't see why you'd want to measure latency relative to said point. What latency should prior responses have then -- negaitve values? Finally please note that a <trial> can only collect a single response, not several.

Hi Dave,

My apologies, I think I did a poor job of relaying my question. I want to be able to measure response latency from the responsetime set point in the video. However, I want people to be able to respond prior to the that point and those responses to be recorded as incorrect responses so I believe responsetime may not be the attribute I want to use because it doesn't allow people to respond prior to the point I set the responsetime to (e.g., 2500). Is there a way to measure response latency from that point in the video, but also allow people to incorrectly respond prior to the responsetime point?

/responsetime specifies the point in time (starting from the beginning of the trial) at which Inquisit starts accepting responses. Any key presses prior to that will simply be ignored. I.e., this should work fine in your case and you should easily be able to test and verify this for yourself.