I'm doing a statistical analysis of an experiment I'm conducting, and right now I'm trying to define a proper methods for data mining. Please take a look at my problem:
I'm doing a user research recording times of completion of a task. As an output I have 10 times of completion from each participant. As you can imagine the time of completion of the first try is the highest and further on while user gets more and more familiar with the interface it exponentially decays. So the graphical representation of the obtained data for each user looks more or less like this expo1.png
I'm trying to get the efficiency time of the tested user interface, which means the average time user needs to complete the task. I assume (i) that the average here is not representative at all, because you have a huge deviation from the average on the first attempt, as user is not familiar with the UI, thus has close to 0 efficiency, (ii) more to the end of the exponential line you go more representative it gets in terms of actual working efficiency. Thus I think that proper weighing of the data could reveal an actual average working efficiency for a certain participant. I tried to indicate it on the image right here expo2.png
The problem is my Hardcore Math XP level is hardly a 0 and I'm not familiar with the concept of weight distribution at all. Could you please assist me in solving this problem.
If you need any clarification please let me know, I'll update the question.
Thank you in advance good sirs!