We did an experiment in which each trial the participants had 3 secs to respond to each stimuli. Unfortunately, participants failed to respond to some of trials in 3 secs. There are two alternate ways to treat these trials:
1. Delete these trials.
2. Treat these trials as a wrong responses.
I wonder which one is better and why?
Dear Liang,
neither of the two options is a good one. It is better to classify them as "response out of time". The main 2 question however are:
1. why do you have set the response time to 3 sec (is there any literature or rational for that)?
2. Why were the subjects not able to respond in time?
Without knowing details about this experiment, the distribution curve of time-to-respond might matter. As a consequence, it might be eventually better to lift the 3 sec limitation and simply describe the distribution of the response time as a result.
If this might not be relevant (you need however justify, why it might not be relevant for the interpretation of the results (i.e. it might be = to the rational for which you have chosen the 3 sec limit), then all subjects not having provided any response in time, should be excluded from the analysis, because the input requested was not provided in time. By doing it, the correct description however would be "response out of time" or "no response within given time period". Simply deleting the data would alterate the experimental context, and suggest that all partecipants have answered in time (what is not true).
Best regards
Manolo
Thank you for you answer. It has been a long time since I last came here, and I am sorry for this late response!
- Badges
- Science topic