For readings to provide evidence so that a concliusion can be drawn, they need to be reliable and valid. Reliable means that we can repeat the experiment and take similar readings, and valid means that the conclusion must be based on the readings.
The readings may not be the same each time the experiment is done, but they should be similar. If the time for an apple to fall a fixed distance is taken many times, the measured times will vary slightly, but should be very similar.
We can improve the reliability of our results by taking an average. Reliability is not the same as precision however. Reliability can be improved by taking more readings but accuracy can be improved by reducing the error associated with each reading – by using a ruler with a truer scale. If the ruler is marked in millimetres, we can align the ruler with a standard ruler and the two should match closely. If we use such a ruler to measure a length rather than a piece of string, the accuracy of our measurement is likely to be greater.
Precision is not the same as accuracy or reliability. Precision is a measure of the smallest possible error. If a ruler is marked in millimetres, the precision of each measurement is 1 millimetre, but measurements made with the ruler will only be reliable if many measurements of a particular length give very similar results and accuracy can only be improved