@alante the questions follows one to another. You did one experiment with two trials. The result of your experiment is the average of the two trials. If you do more trials (we can said 100 trials), the result of your experiment will be the average of your 100 trials. In theory if you are careful enough in your technique the value is going to have a smaller experimental error if you repeat the trials many times than if you do just twice.
The you should average the two values of the two trials (107+113)/2=110
Since the experiment choose each increment of time every 3 seconds, you dont know whats happen between the 0 and 3 or between 3 and 6 you assume that the relationship is linear. This is the smallest time lapse that you are able to measure in this experiment. Maybe under other conditions in a new experiment you can measure whats happen a shortest times. (this is not your issue now).
So the question 3 is asking you at what time your initial sample is only 50% of the initial amount. You have only the measurements of time 0, 3, 6, 9, etc. You know that a
time 0 you have 100% and at
time 3s you have 55%, and at
time 6s you have 24.25%.
At what time you have 50% of your original sample? at 0, 3, 6? You can said that will be some time between 3 and 6 and you can add a second digit (uncertain) at your measurement and said 3.2s, or 4.3s, or anything between 3 and 6s. this value will be the half life of your atom.