Random Errors and Systematic Errors

Errors fall into two types, and each one is dealt with differently.Random errors make readings scatter unpredictably above and below the true value. They come from things that vary from one reading to the next, such as judging the exact bottom of a meniscus, slight changes in when the end point is judged, or small draughts affecting a balance. Random errors reduce precision. Because they are equally likely to be high or low, they can be reduced by repeating measurements and taking a mean.Systematic errors shift every reading by the same amount in the same direction. Examples:• a balance that is not zeroed (a zero error)• always reading a burette from the top of the meniscus• heat lost to the surroundings in an enthalpy experiment, making every temperature rise too small• a gas that is slightly soluble in the water it is collected overSystematic errors reduce accuracy but can leave the results precise. Repeating the experiment does not help, because every repeat has the same bias. Instead you must change the method or apparatus: calibrate or zero the instrument, read scales correctly, insulate the apparatus, or collect the gas in a syringe.A useful test when evaluating results: if repeats agree closely but the mean is far from the accepted value, look for a systematic error.