The best method of reducing random error is
A: to change the instrument used for the measurement.
B: To take help of an experienced observer.
C: to repeat the experiment many times and then take the average of all results.
D: None of the above.

Hint: Performing a particular experiment many times and taking the measurements every time gives you an idea if your measurements are wrong or not. You can be wrong every time you take the measurements, however the readings will be different all the time.

Complete step by step answer:
Let us start with knowing, what is an error in measurement?
The simple answer to the question above is error literally means mistakes. Errors in measurement are the mistakes made in making measurements. These are not the mistakes made in using the measuring instrument or some other kind of mistake. It is the mistakes that are unintentional and are made in recording the values of measurements. These errors occur due minor deformities in the measuring instrument due to temperature. They are very small.
Errors are classified into two categories;
Systematic Errors: Systematic errors are errors in which the measured quantity is displaced from the true value by fixed quantity and in the same direction. This error might occur due to deformities in the instrument body, the markings are erased, wrong marking on the instrument, etc. Examples are zero errors, parallax errors, etc.
Random errors: Random errors are errors in which all measurements differ from their mean value and their magnitudes and directions are different. This error arises due to parallax when an observer reads the scale in an inconsistent direction, irregularities measured in the quantity as certain quantities do not follow a regular pattern, limitation of the equipment as certain equipment may be so sensitive that it can detect even the slightest variation on the signals, etc.
The systematic errors cannot be reduced from any measures. The only way to erase is to change the instrument used with a new one.
However, there is a way to reduce the random method. If the observer takes measurements again and again many times, the probability of measuring wrong values is very less. Still, the measured values are different in most of the cases. These values differ by a very small amount. So, if we take the average of all the values, we will get the true value of the measurement.
Therefore, option C is correct.

Note:
1) Although as mentioned that the systematic errors cannot be corrected, the zero error and parallax can be removed.
2) This is done by noting the zero error and then adding it to the value measured.
3) Similarly, parallax error can also be removed.