If data point is way off the trend line which of the following will not resolve the problem?
a. discard the data point
b. check for a mistake in recording the data
c. run the experiment again to verify the accuracy of the original data.

The thing that will not resolve the problem of your data is discarding the data point - A

The reason why this won't help is because this data point will still be present just not visible from your data. This is an issue because someone may think that if a data point isn't there that it also means everything is k with the data. 

The answer would be b. it is a learned trait