errors of observation - department of physics.docVIP

ERRORS OF OBSERVATION 1. INTRODUCTION The measurement of a physical quantity is always subject to errors. A statement of the result does not mean very much unless some indication is given to the accuracy of the measurement. Some basic knowledge of error analysis is essential for any beginner. It is customary and convenient to divide experimental errors into two types, systematic and random, according to the difference between their characters and the methods of treatment. (a) Systematic errors are constant errors which cause all results to be incorrect by roughly the same amount in the same direction. Usually they cannot be eliminated by averaging or treated by statistical method. Care must be taken in the interpretation of systematic error. (b) Random errors are minor experimental or accidental errors due to errors of judgement, or mildly fluctuating conditions which cannot be controlled. There are many causes of such fluctuations. Random errors may be estimated by statistical method. In most experiments, both random and systematic errors are present. A set of measurements with random errors only are shown in Fig. 1, another set with systematic plus random errors are shown in Fig. 2. Each mark in the figures indicates the result of a measurement. 5 2. SYSTEMATIC ERRORS Examples of systematic errors are those caused by (a) Incorrect calibration of an instrument : e.g. bias caused by friction or wear in its moving parts, and the failure to correct for the zero-reading. (b) Construction faults in the apparatus : e.g. off-centre pivoting of the pointer of a circular scale. (c) Non-constancy of experimental conditions: e.g. change of temperature and pressure which affects the value of the quantity measured and has not been properly taken into account. (d) Bias of the observer: e.g. tendency of some persons to press a stop watch a fraction of a second too early or to overestimate or underestimate the fraction of a scale division. (e) Incorrect use of the fo