The experiment should be conducted in a clear and systematic way to ensure the data is complete and of a high quality. In an experiment into the relationship between force, mass and acceleration a toy car of different masses runs down a ramp. The acceleration needs to be measured several times at each mass. The repeats would all need to be done in the same way and with care to ensure precise data. If you observe that a repeat is not similar to the others then it is good idea to repeat it.
It is also important to pay careful attention while the experiment is being carried out. It might be that the car starts to deviate from a straight line path; which if significant may mean that the method should be modified.
When using observations to collect data during an experiment it is important to be as accurate as possible. For example, a reaction commonly used to investigate rate is between sodium thiosulfate and hydrochloric acid. The reaction takes place in a conical flask standing on a piece of paper where a cross has been drawn. At first, it is easy to see the cross through the reaction mixture. However, during the reaction the solution becomes cloudy. The time taken for the cross to become completely obscured is used as a way of measuring the rate of reaction. Different people might make the call as to whether the cross has been completely obscured at slightly different points so it might be decided to use the same person to make all the observations.