format of timestamps

sdewitt

Hello,

I recently updated a very out of data version of opensesame, to a slightly less out of date version that would still run my task, new version is 0.27.2. In this version, when I look at the time columns (e.g. time_instructions), i am given a very large value that I can't seem to get to make sense in the context of seconds or milliseconds. for example, I have time_instructions listed as 15802071. Which is an event that happens right after the task loads. The task was screen recorded and lasted, as an example 17 minutes and 34 seconds for one person, and the last timestamp in the data output reads as 16734343. If someone could help me discern how to convert this values that are put out by default in the excel log, I would greatly appreciate it!

Best,

Sam

eduard

Hi Sam,

These is most likely millisecond since the start of the experiment. But best try it out. For example, have an experiment with a single two items with a fixed duration (sketchpad of 1 second or so), and then check the output whether you can detect how the 1s difference is coded.

Hope this helps,

Eduard

sdewitt

Thank you so much, Eduard, for your response. With testing, it took some time to deduce what was going on. I was able to ascertain that these large values are in fact milliseconds. However, for this version of opensesame, as opposed to my previous version, it seems, rather oddly, that the milliseconds are not relative to the start of the task file being inspected, rather they are relative to the first task file generated since the last time I opened the opensesame program that houses my task. This means that, unless I have the timestamp for the first file I ran since last opening opensesame, I do not have a start time. And indeed, I do not have the original file I ran before I ran all of my participants, as it was overwritten since it was not a usable file (i.e. it was a test).

Are you familiar with this issue in version 0.27.2? It was not the case in my previous version. Every task file seemed to have time stamps relative to its individual start time.

As a side note/question, in that previous version where timestamps seemed to be relative to start time per file, my first screen was always instructions, and the timestamp for that page was never 0, meaning some time had already elapsed. My default assumption is that, the start time is when I enter a participant number and click save (thus generating the .csv file), and thus the timestamp for when that .csv was created is the true start time, is this a valid assumption?

I appreciate all the help!

eduard

Hi,

Are you familiar with this issue in version 0.27.2? It was not the case in my previous version. Every task file seemed to have time stamps relative to its individual start time.

No I am not, but as you say, the 0.27 times are long gone :)

true start time, is this a valid assumption?

Almost, the timer is started when the experiment object is created. I can't say for sure when this happened, but probably slightly after the participant number was provided.

This means that, unless I have the timestamp for the first file I ran since last opening opensesame, I do not have a start time.

You don't really need it, though, right? As long as you know the scale, you can remove the earliest timestamp from all values and you have the experiment time. For RTs, you don't even need to do that. Just compare the timestamps between which you need the interval. All that is independent from the overall offset.

Eduard